UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Establishing the functional role of ORMDL3 in innate immunity Hsu, Karolynn Jo 2012

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

Item Metadata

Download

Media
24-ubc_2012_fall_hsu_karolynn.pdf [ 3.89MB ]
Metadata
JSON: 24-1.0072762.json
JSON-LD: 24-1.0072762-ld.json
RDF/XML (Pretty): 24-1.0072762-rdf.xml
RDF/JSON: 24-1.0072762-rdf.json
Turtle: 24-1.0072762-turtle.txt
N-Triples: 24-1.0072762-rdf-ntriples.txt
Original Record: 24-1.0072762-source.json
Full Text
24-1.0072762-fulltext.txt
Citation
24-1.0072762.ris

Full Text

 ESTABLISHING THE FUNCTIONAL ROLE OF ORMDL3 IN INNATE IMMUNITY   by   Karolynn Jo Hsu    B.Sc., The University of British Columbia, 2010    A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   MASTER OF SCIENCE   in   The Faculty of Graduate Studies   (Microbiology and Immunology)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   April 2012    © Karolynn Jo Hsu, 2012 ii  ABSTRACT  Asthma and allergic diseases are rapidly becoming the most common chronic diseases in the developed world. Nearly 1 in 3 Canadians suffer from some form of allergy and more than 300 million individuals in the developed world suffer from asthma. These complex disorders are caused by the interaction of various genetic and environmental factors. Genome-wide association studies have been widely used to identify genes associated with asthma susceptibility. The gene, ORMDL3, was shown to be associated with early-onset asthma. Asthmatic patients have elevated expression levels of this gene. The gene encodes a transmembrane protein localized in the endoplasmic reticulum (ER) that may be involved in ER stress and inflammation. Its functional role in asthma pathogenesis, however, has yet to be elucidated. In this research, we investigated the functional role of ORMDL3 in innate immunity. Experimentally, ORMDL3 expression levels were manipulated in vitro in airway cells using overexpression plasmid and siRNA technologies. The effects of ORMDL3 expression levels on inflammatory responses were then explored. After manipulation of ORMDL3 expression levels, cells were stimulated with various immune response-inducing factors. Supernatants collected after stimulation were analyzed and no differences in pro- inflammatory cytokine production were observed. These results suggest that variation in ORMDL3 expression levels does not affect innate immune production of IL-6 and IL-8 in airway cells. ORMDL3 knockdown also did not affect expression of other immune-related genes.   iii  TABLE OF CONTENTS ABSTRACT ............................................................................................................................. ii	
   TABLE OF CONTENTS ........................................................................................................ iii	
   LIST OF TABLES .................................................................................................................... v	
   LIST OF FIGURES ................................................................................................................. vi	
   LIST OF ABBREVIATIONS ................................................................................................ vii	
   ACKNOWLEDGEMENTS ..................................................................................................... ix	
   1     INTRODUCTION ............................................................................................................. 1	
   1.1 Innate immunity .............................................................................................................. 1	
   1.2 Asthma and immunity ..................................................................................................... 2	
   1.3 Susceptibility genes in asthma ........................................................................................ 5	
   1.4 ORMDL3 ......................................................................................................................... 8	
   1.5 ER stress and UPR activation ......................................................................................... 9	
   1.6 Hypothesis and objectives ............................................................................................. 10	
   2     METHODS AND MATERIALS .................................................................................... 14	
   2.1 Cell culture .................................................................................................................... 15	
   2.2 Cloning ORMDL3 cDNA into pEGFP-N1 vector ........................................................ 15	
   2.2.1 Cloning protocol ..................................................................................................... 16	
   2.3 Cell transfection ............................................................................................................ 17	
   2.3.1 Transfection protocol ............................................................................................. 17	
   2.4 Microscopy ................................................................................................................... 18	
   2.5 Cell stimulation and immune response quantification .................................................. 18	
   2.5.1 ELISA protocol ...................................................................................................... 18	
   2.6 ER stress and UPR activation ....................................................................................... 19	
   2.7 RNA isolation and reverse transcription ....................................................................... 20	
   2.8 Quantification of mRNA expression ............................................................................ 20	
   2.9 Immunoblot analysis ..................................................................................................... 20	
   2.9.1 Classical immunoblot method ................................................................................ 21	
   2.9.2 SNAP i.d.® immunoblot method ............................................................................ 21	
   2.10 PCR array .................................................................................................................... 22	
   2.11 Gene expression validation ......................................................................................... 23	
   iv  2.12 Statistical analysis ....................................................................................................... 23	
   3     RESULTS ........................................................................................................................ 28	
   3.1 ORMDL3 expression in airway epithelial cells ............................................................. 28	
   3.2 Transfection optimization ............................................................................................. 29	
   3.3 Gene knockdown .......................................................................................................... 30	
   3.3.1 ORMDL3 knockdown validation ........................................................................... 30	
   3.3.2 Knockdown effect on the UPR ............................................................................... 31	
   3.3.3 Knockdown, cell stimulation, and cytokine production ......................................... 32	
   3.4 Gene overexpression ..................................................................................................... 33	
   3.4.1 Transient expression of ORMDL3-eGFP .............................................................. 33	
   3.4.2 Overexpression, cell stimulation, and cytokine production ................................... 34	
   3.5 Co-transfection .............................................................................................................. 34	
   3.5.1 Plasmid and siRNA co-transfection ....................................................................... 34	
   3.5.2 Co-transfection, cell stimulation, and cytokine production ................................... 35	
   3.5.3 Effect on the UPR .................................................................................................. 36	
   3.5.4 Repetition in A549 cells ......................................................................................... 36	
   3.5.5 PCR array ............................................................................................................... 36	
   4     DISCUSSION .................................................................................................................. 61	
   REFERENCES ....................................................................................................................... 66	
     v  LIST OF TABLES Table 2.1   Volumes for PCR, digestion, and ligation reactions ............................................ 24	
   Table 2.2   Primer sequences for sequencing ......................................................................... 24	
   Table 2.3   Small interfering RNA sequences ........................................................................ 25	
   Table 2.4   Volumes per qPCR reaction (well) ...................................................................... 25	
   Table 2.5   Primer sequences for qPCR ................................................................................. 26	
   Table 3.1   Percent identity of ORMDL1 and ORMDL2 compared to ORMDL3 .................. 38	
   Table 3.2   Genes differentially regulated in cells with low ORMDL3 expression ............... 38	
      vi  LIST OF FIGURES Figure 1.1   World prevalence of asthma. .............................................................................. 12	
   Figure 1.2   ER stress-induced inflammatory responses ........................................................ 13	
   Figure 2.1   Experimental approach ....................................................................................... 27	
   Figure 3.1   Gene expression profile of GSDMA, GSDMB, and ORMDL3 in A549 and 1HAE cell lines .................................................................................................................................. 39	
   Figure 3.2   Gene expression profile of ORMDL1, ORMDL2, and ORMDL3 in various cell lines ......................................................................................................................................... 40	
   Figure 3.3   MAPK1 gene knockdown in A549 cells. ........................................................... 41	
   Figure 3.4   Transfection optimization with siRNA in 1HAE cells ....................................... 42	
   Figure 3.5   Transfection optimization with pEGFP-ORMDL3 in 1HAE cells .................... 43	
   Figure 3.6   Specificity of ORMDL3 siRNA. ........................................................................ 44	
   Figure 3.7   Unfolded protein response activation ................................................................. 45	
   Figure 3.8   Expression of UPR gene markers in unstimulated A549 cells with ORMDL3 knockdown .............................................................................................................................. 46	
   Figure 3.9   ORMDL3 gene knockdown in 1HAE cells ........................................................ 47	
   Figure 3.10   Cytokine production in siRNA-transfected 1HAE cells .................................. 48	
   Figure. 3.11   Schematic depiction of pEGFP-ORMDL3 plasmid ........................................ 49	
   Figure 3.12   Expression of eGFP or ORMDL3-eGFP in 1HAE cells .................................. 50	
   Figure 3.13   Relative ORMDL3 expression in plasmid-transfected 1HAE cells .................. 51	
   Figure 3.14   Cytokine production in plasmid-transfected 1HAE cells ................................. 52	
   Figure 3.15   Visualization of 1HAE cells co-transfected with plasmid and siRNA using fluorescent microscopy ........................................................................................................... 53	
   Figure 3.16   ORMDL3 expression in plasmid and siRNA co-transfected 1HAE cells ........ 54	
   Figure 3.17   Cytokine production in plasmid and siRNA co-transfected 1HAE cells ......... 55	
   Figure 3.18   Expression of UPR gene markers in stimulated 1HAE cells with ORMDL3 knockdown .............................................................................................................................. 56	
   Figure 3.19   Cytokine production in plasmid and siRNA co-transfected A549 cells .......... 57	
   Figure 3.20   PCR array ......................................................................................................... 58	
   Figure 3.21   Validation of differentially expressed genes .................................................... 59	
   Figure 3.22   CCL2 production in plasmid and siRNA co-transfected 1HAE cells .............. 60	
   vii  LIST OF ABBREVIATIONS Abbreviation Name PRR pathogen recognition receptors PAMP pathogen associated molecular pattern TLR toll-like receptor NLR NOD-like receptor IBD inflammatory bowel disease TH1/2 T helper type 1/2 IgE Immunoglobulin E IL- interleukin TSLP thymic stromal lymphopoietin GWAS genome-wide association study SNP single nucleotide polymorphism ORM1/2 orosomucoid 1/2 ORMDL1/2/3 orosomucoid 1-like 1/2/3 GSDMA/B gasdermin A/B ZPBP2 zona pellucida binding protein 2 ER endoplasmic reticulum UPR unfolded protein response SERCA sarco/endoplamic reticulum Ca2+ ATPase JNK c-Jun N-terminal kinase AP-1 activator protein 1 NFκB nuclear factor-kappa B IKK I kappa B kinase ROS reactive oxygen species NO nitric oxide A549 adenocarcinomic human alveolar basal epithelial cells 1HAEo-  (1HAE) normal human airway epithelial cells Thp-1 human acute monocytic leukemia cells Caco-2 human epithelial colorectal adenocarcinoma cells cDNA complementary deoxyribonucleic acid (DNA) dNTP deoxyribonucleotide triphosphate siRNA small interfering ribonucleic acid (RNA) mRNA messenger RNA qPCR quantitative polymerase chain reaction (PCR) eGFP enhanced green fluorescent protein (GFP) CMV Cytomegalovirus FCS fetal calf serum BSA bovine serum albumin  viii              Abbreviation Name TNFα tumor necrosis factor α LPS lipopolysaccharide ELISA enzyme-linked immunosorbent assay PBST phosphate-buffered saline (PBS) Tween-20 TBST tris-buffered saline (TBS) Tween-20 SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis PERK PKR-like eukaryotic initiation factor 2a kinase IRE1 inositol requiring enzyme 1 ATF6 activating transcription factor-6 XBP-1 X-box binding protein 1 CHOP C/EBP homologous protein eIF2α eukaryotic initiation factor 2α GAPDH glyceraldehyde 3-phosphate dehydrogenase PPIA peptidylprolyl isomerase A CCL2/5 chemokine (C-C motif) ligand 2/5 VEGFA vascular endothelial growth factor A ADRB2 adrenergic beta-2 receptor IL12A interleukin 12 subunit A IL13RA1 interleukin 13 receptor alpha 1 CSF2 colony stimulating factor 2 BCL6 B-cell lymphoma 6 ix  ACKNOWLEDGEMENTS  First and foremost, I would like to thank my supervisor, Dr. Stuart Turvey, for the opportunity to pursue my studies in his lab. I am grateful for your guidance and encouragement. I would like to acknowledge my committee members, Drs. Ken Harder and Andrew Sandford. Our discussions were constructive and helpful throughout my research project. I had the opportunity to work with many great people in the Turvey lab. Thank you all for your help and support with ideas and experiments. Lastly, I am grateful for all the love and support from family and friends. Thank you! Research was funded by the Child & Family Research Institute and AllerGen. 1  1     INTRODUCTION  1.1 INNATE IMMUNITY The innate immune system plays a critical role in the host defense mechanism against invading pathogens. In humans, it serves as the first line of host immune defense and it is complemented by subsequent activation of the adaptive immune system. The human innate immune system consists of anatomical and physiological barriers (epithelia, mucociliary clearance, low stomach pH), humoral components (cytokines, anti-microbial peptides, complement proteins), cellular components (dendritic cells, macrophages, granulocytes), and a limited number of invariant receptors [1, 2]. Physical barriers serve as the interface between internal milieu and external environment. When this is compromised, cellular and humoral components of the innate immune response respond to clear the pathogen. Within minutes of infectious insult, pathogen recognition receptors (PRRs) on hematopoietic and non-hematopoietic cells recognize specific conserved components of pathogens. Recognition of these conserved components, named pathogen-associated molecular patterns (PAMPs), triggers the release of chemokines and cytokines that initiate an inflammatory response [3, 4]. Unlike the adaptive immune system, innate immunity responds quickly and no previous exposure to the pathogen is required to mount an immune response [1]. Its importance lies in its rapid initial response to clear infection. The airway epithelium is a critical component in host defense of the respiratory tract. It functions as both a physical barrier to inhaled pollutants, allergens, and microorganisms, and a contributor to innate immune defense. Its role in the innate immune response is significant for the prevention of infection, as the human lung is exposed to many airborne pathogens. Airway epithelial cells sense antigen exposure through various PRRs such as 2  Toll-like receptors (TLRs) and nucleotide binding oligomerization domain (NOD)-like receptors (NLRs). Upon recognition of a pathogen, the cells respond by releasing anti- microbial peptides and cytokines [4, 5]. This localized inflammatory reaction serves to recruit phagocytes that remove microorganisms, as well as activate cells that may initiate the adaptive immune response.   1.2 ASTHMA AND IMMUNITY In addition to its critical preventative role against infection, the innate immune system also contributes to the development of a variety of autoimmune and inflammatory disorders. These disorders include rheumatoid arthritis, systemic lupus erythematosus, inflammatory bowel disease (IBD), asthma and atopy [1, 3]. Asthma and allergic diseases are rapidly becoming the most common chronic diseases in the developed world (Figure 1.1). Nearly 1 in 3 Canadians suffer from some form of allergy, and more than 300 million individuals in the developed world suffer from asthma [6-8]. Medical expenditures for asthma, particularly medications, cost an estimated 12.7 billion dollars annually in the US in 1998 [9]. Current asthma therapy treats the symptoms of the disease. Two types of medication are commonly used to treat patients with asthma: medication to treat immediate symptoms and medication to control inflammation. Inhaled salbutamol and formoterol are examples of β2-agonists that are effective for immediate symptom relief. These drugs act as bronchodilators that relax airway smooth muscles [10]. These β2-agonists, however, have little anti-inflammatory effect. In order to prevent or reduce inflammation in the lung that leads to symptoms, patients also take anti-inflammatory medication such as corticosteroids or 3  leukotriene antagonists. Corticosteroids suppress inflammation by mechanisms which lead to inhibition of eosinophils or increased transcription of anti-inflammatory genes [10, 11]. Leukotriene modifiers, in contrast, inhibit bronchoconstriction by potentially inhibiting survival of activated inflammatory cells [12]. Although these therapies have been very successful at controlling symptoms in asthmatics, they are ineffective in up to 25% of patients [6]. Understanding the pathogenesis of asthma and allergic diseases is therefore crucial in the development of new treatments that address the initial triggers, rather than final symptoms, of these diseases. Asthma and allergic diseases are complex disorders caused by the interaction of various genetic and environmental factors [13-15]. Although the causative link between the environment and the initiation of asthma is poorly understood, it is well established that environmental factors can lead to the exacerbation of asthma symptoms. Environmental risk factors include cigarette smoke, pollution, viral infections, and the presence of mold and dust in the home (home cleanliness) [16].  Genetic contribution to the development of asthma is estimated to vary between 36-79%, though there is no obvious monogenic pattern of inheritance [17]. Because the implications of environment and genetic factors can be quite varied, the resulting asthma phenotypes are also very heterogeneous. Phenotypically, asthma is characterized by chronic inflammation and hyper-reactivity of the airways, mucus overproduction, and reversible bronchoconstriction [14]. More importantly, asthma patients frequently display airway tissue remodeling, such as epithelial metaplasia, smooth muscle proliferation, and enhanced matrix deposition [13, 18, 19]. These changes contribute to the chronicity of the disease. Asthma is also often associated with atopy, although recent research suggests that elevated immunoglobulin E (IgE) levels does not cause asthma as 4  previously believed [20]. Rather, the elevated levels of IgE are more likely to be a secondary effect of asthma. The symptoms of asthma are largely driven by dysregulated T helper type 2 (TH2) responses [21, 22]. Cytokines IL-13, IL-9, IL-5, IL-4, produced by TH2 cells in the lung lead to airway hyper-reactivity, recruitment of mast cells, eosinophils, and regulation of IgE synthesis [21]. Overproduction of such cytokines is characteristic of asthma [19, 22]. However, the reasons for these TH2-driven immune responses are not clear. The hygiene hypothesis is one hypothesis that was developed to explain the increase in prevalence of asthma, allergy, and autoimmunity in industrialized countries [23]. This hypothesis postulates that microbial exposure in early life modulates a shift from a TH2-dominant phenotype at birth to a TH1 phenotype [24]. Thus, lower exposure to microbes and decreased infection rates associated with smaller family size, germ-free environments, and increased immunizations and antibiotic use, contribute to the development of TH2 immune responses [22-24]. Innate immune responses are also involved in asthma pathogenesis. Airway epithelia are central in host defense and immune regulation. These cells are among the first to encounter pathogens and play an important role in shaping downstream immune responses. Interleukin-6 (IL-6) and interleukin-8 (IL-8) (alias CXCL8) are two pro-inflammatory cytokines produced by airway cells that are important in the innate defense function. Both IL-6 and IL-8 are more highly produced in asthmatics than in healthy individuals [25-29]. Interleukin-6 has both pro- and anti-inflammatory effects. However, during acute inflammatory reactions, it has a pro-inflammatory role and is key in stimulating production of acute-phase proteins such as C-reactive protein and serum amyloid A [30]. It is also 5  known to induce the adaptive immune response by stimulating T-cell activation, B-cell differentiation, and antibody production [29, 30]. In contrast, IL-8 acts as a chemoattractant, recruiting neutrophils and other hematopoetic cells to sites of inflammation [29]. Other cytokines and chemokines produced by the airway epithelia include thymic stromal lymphopoietin (TSLP), IL-25, and IL-33 [21]. These cytokines can in turn cause TH2-like responses by inducing mast cells, eosinophils, basophils, natural helper cells, and natural killer T cells to produce TH2 cytokines [21]. Any dysregulation of the innate immune response can result in hypersensitivity to environmental factors, leading to asthma symptoms. Even so, the underlying mechanisms causing these clinical phenotypes still remain unclear. Recent advancements have enabled researchers to identify candidate genes that influence the risk of asthma. This information is now being used to better understand the mechanistic complexity of this disease.   1.3 SUSCEPTIBILITY GENES IN ASTHMA Complex disorders are influenced by multiple genes and therefore do not follow simple Mendelian patterns of inheritance. Two approaches, association studies and linkage studies, have been used to identify genes that may be involved in asthma pathogenesis [31]. Association studies compare non-related individuals in a population, whereas linkage studies compare individuals within a family [31]. Both approaches involve scanning DNA markers in an individual’s genome and looking for patterns among affected versus healthy individuals. Association studies can be divided into genome-wide association studies (GWAS) or candidate-gene association studies. Using a case-control setup, GWAS scan single nucleotide polymorphisms (SNPs) across the entire genome. In contrast, candidate- 6  gene association studies are often more focused, targeting specific genes in the genome. Despite their differences, all of these approaches aim to identify genetic variations between patients with a phenotype and non-affected individuals in order to link particular genes to the disease. Association studies are important for linking genetic variants to disease; however, this approach is limited in its ability to determine which polymorphisms are causal. Functional studies are therefore required to dissect how and why a gene contributes to a phenotype. More importantly, in a complex disease such as asthma, many genes can influence and contribute to the disease. The goal is to identify key genes involved in asthma pathogenesis that can be used as therapeutic targets or diagnostic biomarkers. Several genome-wide and candidate-gene association studies have been performed to date with respect to asthma susceptibility, and researchers have been able to identify many genes associated with disease pathogenesis. Genetic polymorphisms on chromosome 17 have identified the locus, chromosome 17q21, to be associated with childhood asthma susceptibility [20, 32-39]. A list of the studies linking the locus to asthma was recently summarized in Immunological Reviews [40]. Of particular interest is the gene orosomucoid 1 (Saccharomyces cerevisiae)-like 3 (ORMDL3), which has been associated with both asthma and inflammatory bowel disease (IBD) [36, 41]. Association of ORMDL3 in both asthma and IBD is of interest because the lung and gut are composed of similar mucosal surface cells [42]. These tissues are exposed to many potentially harmful antigens and are thus tightly regulated by the mucosal immune system. This unique system is responsible for maintaining a delicate equilibrium between antigen responsiveness and tolerance and is therefore responsible for preventing hyper-reactivity [42]. Inappropriate immune responses to foreign components or commensal bacteria can lead to inflammation characteristic of asthma and 7  IBD. The similarities between asthma and IBD, the lung and the gut, suggest that ORMDL3 may be involved in dysregulation of the immune system. Other genes within the 17q21 locus include gasdermin A (GSDMA, alias GSDM1), gasdermin B (GSDMB, alias GSDML) and zona pellucida binding protein 2 (ZPBP2) [38]. The gene, GSDMA, is a tumor suppressor and regulates apoptosis in epithelium of the upper gastrointestinal tract [43]. Elevated expression of GSDMB is found in cancer cells, which suggests that it is involved in tumor development [43-45]. In contrast, the ZPBP2 gene is expressed in the acrosomal matrix of spermatozoa and plays a role in fertilization [46]. Because ZPBP2 does not play a functional role in immunity or in airway cells, we did not pursue investigating this gene. Expression of GSDMB and ORMDL3 are correlated and may be cis-regulated [36, 38]. However, because the polymorphisms in this locus are within intronic or regulatory regions, genes ORMDL3, GSDMA and GSDMB are all potential candidate genes involved in asthma pathogenesis. In this investigation, we focused on the gene ORMDL3. Moffatt and colleagues first reported that multiple SNPs on chromosome 17q21 linked ORMDL3 to the risk of childhood asthma [36]. This association has since been reproduced in multiple independent studies across different populations [20, 33, 34, 37, 39, 47-49]. Although evidence points to ORMDL3 being a predominant risk factor for early- onset asthma, others have also shown an association with adult-onset asthma [50, 51]. Furthermore, the polymorphisms may be involved in regulation of mRNA expression of 17q21 locus genes, including ORMDL3 [36, 52]. Analysis of four polymorphisms (rs7216389, rs4795405, rs8079416, rs3859192) in cord blood mononuclear cells showed that children homozygous for all four risk alleles had higher ORMDL3 expression upon 8  stimulation and also elevated levels of IL-17 secretion in both unstimulated and stimulated cells [52]. This correlation further supports the idea that ORMDL3 is involved in immunity.  1.4 ORMDL3 The ORMDL3 gene is a member of a family of highly conserved endoplasmic reticulum (ER)-localized transmembrane proteins. Homologs of these genes are present in various organisms including yeast, microsporidia, plants, Drosophila, and urochordates [53]. The human ORMDL family consists of three genes (ORMDL1, ORMDL2, ORMDL3) [53]. The functions of the ORMDL proteins are currently unknown, but a recent study suggested that ORMDL3 protein is involved in ER-mediated Ca2+ homeostasis and activation of the unfolded protein response (UPR) – ORMDL3 may inhibit sarco/endoplasmic reticulum Ca2+ ATPase (SERCA) activity [41, 54]. Inhibition of SERCA expression or function in airway smooth muscle contributes to airway remodeling in asthmatics [55]. Another study showed that human ORMDL3 partially rescued the phenotype of yeast orm1 orm2 knock-out mutants [53, 56]. The normal function of the yeast ORM1/2 proteins is to negatively regulate ER- membrane sphingolipid synthesis [56, 57]. The observation that ORMDL3 can partially compensate for the loss of ORM1/2 function suggests that ORMDL3 is involved in sphingolipid homeostasis. Disruption of ER sphingolipid homeostasis may also activate the UPR [56]. Activation of the UPR and subsequent inflammatory responses caused by changes in ORMDL3 expression may explain the role of ORMDL3 in asthma pathogenesis.      9  1.5 ER STRESS AND UPR ACTIVATION The ER is important for many cellular functions, including protein synthesis and calcium homeostasis. Disruptions to ER Ca2+ concentrations can cause protein misfolding, and accumulation of these unfolded proteins can lead to ER stress [58, 59]. UPR signaling cascades are initiated in response to this stress and restoration of homeostasis is achieved by attenuating translation, restoring protein folding, or degrading misfolded proteins [59]. The ER stress response and UPR can initiate inflammation through induction of cytokine production or activation of transcriptional regulators of inflammatory genes (Figure 1.2). UPR pathways have been shown to activate the JNK-AP-1 and NF-κB-IKK pathways and the production of reactive oxygen species (ROS) [58, 60-62]. Both AP-1 and NF-κB are transcription factors. Upon activation, they drive the transcription of inflammatory mediators and cytokines that are involved in recruitment and activation of leukocytes [63-65]. ROS production has many functional roles, including immune modulation, activation of adaptive immunity, and induction of autophagy [66, 67]. ER stress and UPR have been implicated in many immune-related diseases including IBD, diabetes, chronic obstructive pulmonary disease (COPD), arthritis, and neurodegenerative inflammatory diseases [68]. It is currently poorly understood whether ER stress is an underlying cause of disease or if its induction is a result of chronic inflammation. Indeed, it is possible that environment factors such as infection or inhalation of smoke particles can activate the UPR, triggering the onset of lung disease in genetically predisposed individuals [69]. However, it is also probable that ER stress is exacerbated by inflammation and contributes to the perpetuation of the disease. The UPR protein CHOP is implicated in lung inflammation in vivo, highlighting the importance of understanding the role of the UPR 10  network of pathways in lung disease [70]. Despite these findings, however, the link between asthma and ER stress has yet to be established.  1.6 HYPOTHESIS AND OBJECTIVES The goal of the present thesis is to investigate the functional role of ORMDL3 in innate immunity. Specifically, the role of ORMDL3 in innate immunity in airway cells was explored. Given that ORMDL3 expression levels in asthma patients are elevated and that the gene potentially functions as a negative regulator of ER homeostasis, it is hypothesized that increased ORMDL3 levels result in heightened inflammatory responses. Increased levels of ORMDL3 protein may in turn disrupt ER homeostasis, leading to ER overload and UPR responses that initiate inflammatory responses. The genes located on chromosome 17q21 – GSDMA, GSDMB, and ORMDL3 – are all potential candidate genes in asthma pathogenesis. However, it was important to first confirm that ORMDL3 was the optimal gene to pursue for the purposes of this investigation. GDSMA and GSDMB are both known to have low expression in bronchial epithelial cells [38, 43]. Functionally, they both have roles in cancer development [43]. ORMDL3, in contrast, is potentially involved in the activation of the UPR and inflammatory responses. The objective of this research was to explore how this locus affects the innate immune response. Since GSDMA and GSDMB are not highly expressed in the lung, we believed that their functional role may have less impact than ORMDL3 in airway cells. Using quantitative PCR, expression profiles of the genes were established in two different airway cell lines, A549 and 1HAE.  11  Experimentally, ORMDL3 expression was manipulated in vitro in airway alveolar (A549) or epithelial (1HAE) cells and its relationship with the inflammatory response was established. The effects of ORMDL3 expression levels on inflammatory responses were assessed using various assays. Next, the effect of knockdown of ORMDL3 on the activation of the ER stress and the UPR pathways was explored. Specific genes within the response pathways were chosen as markers of UPR induction. A significant strength of this study is the ability to manipulate ORMDL3 expression. In doing so, the difference in gene expression that has been established between asthmatics and healthy individuals could be mimicked. Although this was done in cell lines in vitro, this method ensured control and the confidence that any effect on the innate immune response was in fact correlated with a change in ORMDL3 expression levels. If the same experiments were performed on ex vivo airway cells of patients, genetic and other differences between individuals could have affected the results. Overall, the results of this investigation help to further our understanding of asthma pathogenesis. This research provides insight into ORMDL3 function and how genetic variation affecting expression levels contributes to inflammation in asthma patients. Elucidating the role of ORMDL3 in immunity will contribute to identifying new therapeutic targets that may improve clinical outcomes of autoimmune diseases such as asthma.  12       Figure 1.1   World prevalence of asthma. © GINA. From the Global Burden of Asthma Report used with permission from the Global Initiative for Asthma (GINA), www.ginasthma.org [7].  13                    Figure 1.2   ER stress-induced inflammatory responses. Activation of any of the three pathways of the unfolded protein response can lead to inflammatory responses, inducing production of pro-inflammatory cytokines such as IL-6 and IL-8. © Elsevier. Cell, 2010, adapted by permission [58].      P IRE1      P  P  P ATF6 PERK eIF2α Reduced IκBα ATF6-N JNK XBP1 splicing AP-1 Reduced IκBα Nrf2 ROS, NO CHOP Apoptosis  Traf2  IKK NF-κB INFLAMMATORY RESPONSE   14  2     METHODS AND MATERIALS Cell transfection with small interfering RNA (siRNA) or overexpression plasmid were the chosen methods for manipulating expression of ORMDL3 in vitro. Though transient, these tools allowed us to effectively alter gene expression and investigate the functional outcomes. siRNA silences gene expression post-transcriptionally. The siRNA sequence is approximately 21-nucleotides in length and complementary to a portion of the target gene mRNA. After transfection into the cell, the siRNA is incorporated into a protein complex called RISC (RNA-induced silencing complex). The siRNA serves as a template for the RISC complex that guides it to complementary mRNA. Complementation of the siRNA to the target mRNA initiates the RNA interference pathway, leading to degradation of the mRNA by endonucleases. It may also prevent protein synthesis by repressing translation [71, 72]. As a result, protein expression is reduced and the gene is effectively “knocked down”. Overexpression plasmids were used to express high levels of a particular protein, encoded by the plasmid, in cells. In this investigation, we used the pEGFP-N1 plasmid driven by the human cytomegalovirus (CMV) promoter. The strong CMV allowed for constitutive expression of ORMDL3-eGFP. To transfect the plasmid or siRNA into the cells, we employed Nucleofection® technology. The Nucleofector® system is a non-viral electroporation method that uses cell- type specific reagents and programs to transfect substrates into the cell. This method is advantageous for plasmid transfection because of its ability transfect substrates directly into the cell nucleus, allowing for more rapid expression. Other conventional transfection 15  methods allow the substrate to enter the cytoplasm, but require cell division for the substrate to enter the nucleus. The overall experimental approach is depicted in Figure 2.1. Unless otherwise stated, cells were transfected with plasmid and/or siRNA at t = 0h and allowed to recover for 24 hours. At t = 24h, cells were stimulated for 24 hours. Supernatants, RNA, and protein were harvested at t = 48h.   2.1 CELL CULTURE A549 cells (adenocarcinomic human alveolar basal epithelial cells) were cultured in F-12K medium supplemented with 10% fetal calf serum (FCS), 2mM L-glutamine, and 1mM sodium pyruvate (HyClone). 1HAEo¯ (1HAE) cells (SV40-transformed normal human airway epithelial cells) were cultured in DMEM-high glucose medium with 10% fetal calf serum (FCS), 2mM L-glutamine, and 1mM sodium pyruvate (HyClone)  [73]. Cells were incubated in a 37°C, 5% CO2 incubator.   2.2 CLONING ORMDL3 CDNA INTO PEGFP-N1 VECTOR The ORMDL3 gene was amplified from A549 cell line cDNA using forward primer 5’-CTAAGAATTCATGAATGTGGGCACAGCGCAC-3’ and reverse primer 5’- TACTGGTACCCCGTACTTATTGATTCCAAAAATCCGGACT-3’, introducing EcoRI and KpnI restriction endonuclease sites, respectively. The ORMDL3 PCR product was then inserted into a pEGFP-N1 eukaryotic expression vector (Clontech). ORMDL3 and eGFP are in frame and produce a fusion protein with eGFP expressed at the C-terminus of ORMDL3.   16  2.2.1 CLONING PROTOCOL ORMDL3 was amplified by PCR from A549 cDNA template with Phusion® high- fidelity DNA polymerase (NEB). The following cycle was used: 98°C for 30s, [98°C for 10s, 71°C for 30s, 72°C for 20s] x 30, 72°C for 5mins, hold at 4°C. The amplified product was PCR purified and the size was confirmed by gel electrophoresis (1% agarose, 100V for ~30mins). PCR product and pEGFP-N1 vector were then digested separately in 50µL reactions with KpnI restriction enzyme (NEB1 buffer, 1 hour at 37°C), followed by digestion with EcoRI-HF (NEB4 buffer, 1 hour at 37°C). Digested DNA was then gel purified (1% agarose, 100V for ~30mins) (Omega Bio-Tek). Approximate intensities of each were determined qualitatively after gel electrophoresis. The vector DNA was approximately 10 times more intense than the insert and is also 10 times larger than the insert. Insert and vector were ligated with 3:1 and 9:1 ratios at room temperature for 1 hour. For 2µL of vector per ligation reaction, the volume of insert to be added was calculated as follows: 𝑉𝑜𝑙𝑢𝑚𝑒  𝑜𝑓  𝑖𝑛𝑠𝑒𝑟𝑡 = 𝑟𝑎𝑡𝑖𝑜   𝑒.𝑔. 3  𝑓𝑜𝑟  3: 1 ×𝑣𝑜𝑙𝑢𝑚𝑒  𝑜𝑓  𝑣𝑒𝑐𝑡𝑜𝑟  ×    𝑆𝑖𝑧𝑒  𝑜𝑓  𝑖𝑛𝑠𝑒𝑟𝑡𝑆𝑖𝑧𝑒  𝑜𝑓  𝑣𝑒𝑐𝑡𝑜𝑟× 𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦  𝑜𝑓  𝑣𝑒𝑐𝑡𝑜𝑟𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦  𝑜𝑓  𝑖𝑛𝑠𝑒𝑟𝑡 Ligation reactions were performed for vector + insert and digested vector alone (negative control). Volumes for PCR, digestion and ligation reactions are listed in Table 2.1. Ligation mixtures were subsequently transformed into DH5α competent bacteria (Invitrogen) and plated on kanamycin (30µg/mL) agar (Sigma). Colonies were obtained from plates the next day, grown in liquid culture over-night, and the plasmids were isolated by mini-prep (Omega Bio-Tek). The construct sequence was verified by sequencing using CMV-forward and EGFP-N vector sequencing primers (Table 2.2). This construct is denoted as pEGFP-ORMDL3. Protein is denoted as ORMDL3-eGFP. 17    2.3 CELL TRANSFECTION A549 and 1HAE cell lines were transfected with pEGFP-ORMDL3, non-specific (scramble) or ORMDL3-specific siRNA (pre-designed by Qiagen) using Amaxa® Cell Line Nucleofector® Kit T (Lonza). Sequences for the siRNA used are listed in Table 2.3. Two ORMDL3 siRNAs were used. Concentrations used for transfection represent pooled siRNA concentration. For A549 cells, 1x106 cells/reaction were transfected using program X-001. For 1HAE cells, 5x105 cells/reaction were transfected using program T-020 or W-001. Program T-020 was used for siRNA alone transfections, whereas program W-001 was used for plasmid and siRNA co-transfections. Cells were seeded into a 24-well plate (BD Biosciences) at a density of 2x105 cells/well for A549 cells or 1x105 cells/well for 1HAE cells.   2.3.1 TRANSFECTION PROTOCOL After cell counting, the required number of cells was centrifuged at 90 x g for 10mins at room temperature. Supernatant was subsequently removed and the cell pellet was resuspended in room temperature Nucleofector® Solution (100µL/reaction). The cell solution was aliquoted into 100µL samples and plasmid DNA or siRNA was added. The samples were then transferred into certified cuvettes, inserted into the Nucleofector® cuvette holder, and the selected program applied. After transfection, ~400µL of pre-warmed culture medium was added to the cuvette and the sample was divided among 5 wells of a 24-well plate containing 1mL pre-warmed medium/well. Cells were incubated in a 37°C, 5% CO2 incubator. 18    2.4 MICROSCOPY  Cells were visualized by bright field and fluorescent microscopy with a Nikon® Eclipse TS100 microscope with Epi-Fluorescence attachment. Images were obtained with a DS-Fi1 camera head, DS-U3 camera control unit, and NIS Elements F 3.2 software (Nikon).   2.5 CELL STIMULATION AND IMMUNE RESPONSE QUANTIFICATION After transfection, cells were left for 24 hours before stimulation with TNFα (200ng/mL) (eBioscience), E.coli K12 LPS (100µg/mL) (InvivoGen), S. typhimurium flagellin (10-200ng/mL) (InvivoGen), or IL-1β (200ng/mL) (eBioscience). Cells were stimulated for 24 hours. Supernatants were collected and analyzed for cytokine secretion. Pro-inflammatory cytokines, IL-6 and IL-8, were detected and quantified using Human IL-6 and IL-8 Ready-Set-Go!® ELISA kits (eBioscience).   2.5.1 ELISA PROTOCOL • Capture antibody For each plate, 100µL of diluted antibody (20µL of capture antibody in 10mL coating buffer) was added per well. Plates were sealed and refrigerated overnight or longer.   • Blocking The next day, plates were washed three times with PBST wash buffer and then blocked with 200µL of assay diluent/well for 1 hour at room temperature. 19  • Sample, standard preparation and incubation Samples were diluted appropriately in assay diluent and 2x serial dilutions of standards were prepared. After blocking, plates were washed three times with PBST wash buffer. 100µl/well of samples and standards were added in duplicate to the plates and incubated for 2 hours at room temperature. • Detection antibody Plates were washed three times with PBST wash buffer. For each plate, 100µL of diluted antibody (20µL of capture antibody in 10mL assay diluent) was added per well. Plates were incubated for 1 hour at room temperature. • Avidin-HRP binding Plates were washed three times with PBST wash buffer. For each plate, 100µL of diluted avidin-HRP (20µL of capture antibody in 10mL assay diluent) was added per well. Plates were incubated for 30mins at room temperature. • Detection Plates were washed five times with PBST wash buffer. 100µl of TMB substrate solution was added to each well. Reactions were stopped with the addition of 50µl of 1M H2SO4 stop solution. Plates were read by absorbance detection using Spectra Max Plus 384 plate reader.       2.6 ER STRESS AND UPR ACTIVATION  Cells were stimulated with tunicamycin (200µg/mL) or thapsigargin (10µM) to induce ER stress and activate the UPR. RNA was extracted at 2 or 4 hours after stimulation. 20  Expression of genes XBP-1u, XBP-1s, and CHOP were then quantified as markers of UPR activation.  2.7 RNA ISOLATION AND REVERSE TRANSCRIPTION RNA was extracted from lysates using E.Z.N.A.® Total RNA Kit (Omega Bio-Tek). Protocol as described by the manufacturer was followed. RNA was eluted in 40µL H2O. Extracted RNA was reverse transcribed into cDNA using the SuperScript® VILO™ cDNA Synthesis Kit (Invitrogen). Complement DNA was diluted to 200ng/µL prior to quantification of gene expression by qPCR. This method was followed for all samples, unless otherwise stated (see section 2.10 PCR Array).   2.8 QUANTIFICATION OF MRNA EXPRESSION Gene expression was calculated relative to ACTB, GAPDH or PPIA housekeeping genes and was quantified by SYBR Green chemistry (PerfeCTa™ qPCR SuperMix, ROX, Quanta Biosciences) using a 7300 Real Time PCR System (Applied Biosystems). Reactions were done in triplicate and volumes (per reaction) are listed in Table 2.4. The primers used are listed in Table 2.5. Reactions were performed using the following cycling conditions: 50°C for 2mins, 95°C for 10mins, [95°C for 15s, 60°C for 1min] x 40. Gene expression was calculated by the Pfaffl method [74].  2.9 IMMUNOBLOT ANALYSIS Cells were lysed in 50µl RIPA Buffer + 1x HALT™ protease inhibitor (Thermo Scientific). Cell debris was removed by centrifugation: 18,000 x g for 10mins at 4°C. SDS 21  load buffer was added to the lysate and the sample was boiled for 5mins. Total protein was loaded onto 10% SDS-PAGE gel. The gel was run at 150V for ~1 hour in 1x SDS running buffer. Proteins were transferred onto Immobilon®-FL transfer membrane (Millipore) by wet transfer and run at 100V for 1 hour on ice. Antibodies used for Western blot analysis were: monoclonal anti-GFP antibody 1:10,000 (Clontech), anti-ACTB antibody 1:6,000 (Cell Signaling), and IRDye® 680 or 800 secondary antibodies 1:8000 (Li-cor). Two methods, classical and SNAP i.d.® (Millipore), were used for protein detection.   2.9.1 CLASSICAL IMMUNOBLOT METHOD Membranes were blocked at room temperature in 5% milk TBST for 2-3 hours and incubated at 4°C with primary antibody (in 5% milk TBST) over-night. The next day, primary antibody solution was removed and the membrane was washed three times with TBST at room temperature for 10mins per wash. Secondary antibody in 5% milk TBST was added to the membrane and incubated at room temperature for 1h. The secondary antibody is light sensitive, so the membrane was covered for all subsequent steps. Following, the membrane was again washed three times with TBST at room temperature for 10mins per wash. Membrane was washed once with TBS and then visualized using an Odyssey Infrared Imaging System (Li-cor).   2.9.2 SNAP I.D.® IMMUNOBLOT METHOD All volumes listed are for a single well blot holder. 22  Antibody concentrations were scaled as follows: for example, for 1:10,000 antibody concentration as used in the classical method (1µL in 10mL solution), the same absolute volume of antibody (1µL) was added to 3mL 0.5% milk TBST. The membrane was placed with protein side down in the pre-wet blot holder. The spacer was placed on top of the membrane and air bubbles were rolled out. Blot holder was closed and placed in the SNAP i.d. instrument. Membrane was blocked with 30mL 0.5% milk TBST. Vacuum was applied to the system until well was emptied and then turned off. Primary antibody was added evenly over the entire membrane surface and left to incubate for 10mins at room temperature. After 10mins, vacuum was applied. Vacuum was left running while washing the membrane – membrane was washed three times with 30mL TBST per wash. Vacuum was turned off. Secondary antibody was added to the membrane surface and incubated for 10mins at room temperature. After 10mins, vacuum was applied and the membrane was washed three times with 30mL TBST per wash. Finally, the membrane was washed once with 30mL TBS then visualized using an Odyssey Infrared Imaging System.   2.10 PCR ARRAY TNFα-stimulated 1HAE cells transfected with pEGFP-ORMDL3 and ORMDL3 siRNA (low ORMDL3 expression) were compared to stimulated cells transfected with pEGFP-ORMDL3 and scramble siRNA (high ORMDL3 expression). Extracted RNA was reverse transcribed into first strand cDNA using the RT2 First Strand Kit (SABiosciences, Qiagen). Protocol as described by the manufacturer was followed. Two RT2 Profiler PCR arrays (SABiosciences, Qiagen), profiling expression of 84 genes each, were used: Human Cytokines & Chemokines and Allergy & Asthma. 23  Complementary DNA template was mixed with RT2 SYBR® Green qPCR Mastermix (SABiosciences, Qiagen) as follows: 1350µL SYBR Green Master Mix, 1248µL nuclease- free H2O, and 102µL cDNA (~200ng/µL). Note: these volumes were used as recommended by the manufacturer for use with a 7300 Real Time PCR System (Applied Biosystems). Template was then aliquoted into PCR plates containing pre-dispensed primers. Cycler program as provided by the manufacturer was used. Results were analyzed using the PCR Array Data Analysis Web Portal.   2.11 GENE EXPRESSION VALIDATION  Primers were designed for the genes that had a fold change greater than 1.5. A complete list of the primers used is shown in Table 2.5. Differential gene expression was validated on cDNA from TNFα-stimulated (24 hour stimulation) 1HAE cells with ORMDL3 knockdown using qPCR analysis (see section 2.8 Quantification of mRNA expression). Expression was quantified to a housekeeping gene, PPIA. Results were obtained from three experimental repeats (n=3).  2.12 STATISTICAL ANALYSIS Data are shown as mean ± SEM of three experimental repeats (n=3), unless otherwise stated. Results were analyzed using one-way ANOVA with Bonferroni post-test. This test was chosen because it could correct for multiple comparisons. Statistical analysis was performed using GraphPad Prism5 (GraphPad Software, Inc.). After normalization, differences with p < 0.05 were considered significant.  24  Table 2.1   Volumes for PCR, digestion, and ligation reactions   Reagent Volume (µl) PCR Nuclease-free H20 12.4  5x HF buffer 4  10mM dNTP 0.4  10µM forward primer 1  10µM reverser primer 1  cDNA (200ng/µL) 1   Phusion DNA polymerase 0.2 Digestion 10x NEB buffer 5  10x BSA 5  DNA up to 1µg  enzyme 1   Nuclease-free H20 to 50µl Ligation vector 2  insert x  10x DNA ligase buffer 2  T4 DNA ligase 1   Nuclease-free H20 to 20µl       Table 2.2   Primer sequences for sequencing Primer name Sequence (5'→3') CMV forward  CGCAAATGGGCGGTAGGCGTG EGFP-N CGTCGCCGTCCAGCTCGACCAG              25  Table 2.3   Small interfering RNA sequences Target gene Product name (Qiagen) Sequence (5'→3') Non-specific Ctrl_AllStars_1 Proprietary, not disclosed ORMDL3 Hs_ORMDL3_1 CAGGCTTGGAGGGTTAATTTA ORMDL3 Hs_ORMDL3_6 CAGCTTCTACACTAAGTACGA MAPK1 Mm/Hs_MAPK1_3 AATGCTGACTCCAAAGCTCTG       Table 2.4   Volumes per qPCR reaction (well) Component Volume (µl) 2x SYBR Green Master Mix 12.5 Nuclease-free H20 10.5 cDNA (200ng/µL) 1 10µM forward primer 0.5 10µM reverser primer 0.5                          26  Table 2.5   Primer sequences for qPCR Gene Primer Sequence (5'→3') PPIA forward TAAAGCATACGGGTCCTGGCATCT   reverse ATCCAACCACTCAGTCTTGGCAGT GAPDH forward GCACCGTCAAGGCTGAGAACGG   reverse CGACGTACTCAGCGCCAGCATC ACTB forward GTTGCGTTACACCCTTTCTT  reverse ACCTTCACCGTTCCAGTTT ORMDL1 forward AATGGCTGGTCCTTCAAGTGCT   reverse ACCCTCACTGTGATGCCCTTTA ORMDL2 forward ACACACTGGGAGCAAATGGACT   reverse AGTGCGCAGCATCATACTTGGT ORMDL3 forward TCAGGCAGCCAAAGCACTTTAACC   reverse ACCCATCCCACACTTGCTTCCATA GSDMA forward AGCAGCTTACCAAGGCCTCCTAAT   reverse TGGGCAGGTGGAAGTTGGTTATCA GSDMB forward TGGGTTCGGAGGATTCCAGAAACA   reverse AATATCCTCCTTGCCGAGGCACTT CCL2 forward TCGCTCAGCCAGATGCAATCAATG   reverse TGGAATCCTGAACCCACTTCTGCT CCL5 forward TGCCTGTTTCTGCTTGCTCTTGTC   reverse TGTGGTAGAATCTGGGCCCTTCAA ADRB2 forward TCATCATGGGCACTTTCACCCTCT   reverse AGCTCCTGGAAGGCAATCCTGAAA BCL6 forward ACAATCCCAGAAGAGGCACGAAGT   reverse GCTCGAAATGCAGGGCAATCTCAT CSF2 forward AAATGTTTGACCTCCAGGAGCCGA   reverse GGTGATAATCTGGGTTGCACAGGA IL13RA1 forward GTCCCAGTGTAGCACCAATGA   reverse CAGTCACAGCAGACTCAGGAT IL12A forward ATGATGGCCCTGTGCCTTAGTAGT   reverse AGGGCCTGCATCAGCTCATCAATA VEGFA forward TTCAGGACATTGCTGTGCTTTGGG  reverse TGGGCTGCTTCTTCCAACAATGTG     27                  Figure 2.1   Experimental approach. Cells were transfected with pEGFP-ORMDL3 and/or siRNA at time = 0 hour and subsequently stimulated at time = 24 hours. At time = 48 hours, supernatants, RNA, and protein were collected for ELISA, qPCR, and immunoblot analysis, respectively. This approach was followed throughout the presented thesis, unless otherwise stated.   t = 0h t = 24h t = 48h Transfect cells with plasmid, siRNA Stimulate cells Harvest supernatant, RNA, protein 28  3     RESULTS 3.1 ORMDL3 EXPRESSION IN AIRWAY EPITHELIAL CELLS To determine the gene expression levels of the three genes in the indicated cells lines, RNA was extracted from unstimulated cells and reverse transcribed to cDNA. The cDNA was subsequently used in qPCR and relative gene transcript levels were compared. Our results confirmed that GSDMA and GSDMB have low expression in airway cells. We also show, in Figure 3.1, that GSDMA and GSDMB mRNA expression are much lower compared to that of ORMDL3. ORMDL3 is highly similar to ORMDL1 and ORMDL2. Table 3.1 shows the percent identity between the three genes at the transcript and protein levels. Quantitative PCR was performed on cDNA from both airway cell lines, in addition to Caco-2 (epithelial colorectal adenocarcinoma) and Thp-1 (monocytic leukemia) cells. Results indicate that ORMDL3 is least abundantly expressed compared to ORMDL1 and ORMDL2 in all cell types tested (Figure 3.2). This presented a challenge for detecting knockdown of ORMDL3 protein because the commercially available antibodies have poor specificity and are polyclonal. The three proteins also have similar sizes: 17,371Da (ORMDL1), 17,363Da (ORMDL2), and 17,495Da (ORMDL3). Several antibodies were tested and, in all cases, Western blot analysis could not distinguish between the three proteins of the ORMDL family. Because our qPCR data showed that ORMDL3 had lower expression than ORMDL1 and ORMDL2, the abundant levels of ORMDL1 and ORMDL2 proteins would prevent us from detecting the changes in ORMDL3 protein levels.     29  3.2 TRANSFECTION OPTIMIZATION  An optimized transfection protocol was provided by the manufacturer (Lonza) for A549 cells. According to manufacturer data, transfection of 1x106 cells with program X-001 was optimal for transfection efficiency (~75%) and cell viability (~80%, 24 hours post- transfection). Using this protocol, we validated this method of transfection and optimized the concentration of siRNA required to knockdown a gene. A549 cells were transfected with a Qiagen validated MAPK1 siRNA in increasing concentrations. This MAPK1 siRNA served as a positive control for knockdown using the Nucleofector®. Knockdown efficacy of MAPK1 was determined by qPCR. Results indicate that concentrations of 100nM-500nM were sufficient to achieve good knockdown (~40-80%) of MAPK1 in A549 cells (Figure 3.3). We also confirmed that the non-specific (scramble) siRNA did not affect gene expression – no difference was observed in cells with scramble siRNA versus the cells only control. These siRNA concentrations were later applied and verified for transfections with ORMDL3 siRNA. Concentrations of siRNA in this range (below 500nM) were used for all subsequent transfections. For 1HAE cells, no optimized protocol was provided by the manufacturer. Therefore, we tested several programs that were recommended for epithelial cells. Cells were transfected with 200nM ORMDL3 siRNA using five different programs that differed in time and voltage parameters: S-005, T-013, T-020, W-001, and U-017. ORMDL3 transcript expression was then compared 24 hours post-transfection. We determined that program T- 020 yielded the highest knockdown – using 200nM ORMDL3 siRNA, we obtained approximately 90% knockdown (Figure 3.4). Because such a high gene knockdown was 30  achieved with 200nM siRNA, we used lower concentrations of siRNA for transfection in 1HAE cells (50nM-200nM).  A similar optimization protocol, as above, was followed for transfection with pEGFP- ORMDL3. We tested programs S-005, T-020, U-017, and W-001. We determined that transfection with program W-001 yielded the best overall results with good transfection efficiency and good cell morphology (Figure 3.5). For all subsequent transfections with pEGFP-ORMDL3 (alone or with siRNA), we used program W-001.  3.3 GENE KNOCKDOWN  3.3.1 ORMDL3 KNOCKDOWN VALIDATION In order to determine functional impact of ORMDL3, knockdown of the gene was performed using siRNAs. As previously mentioned, ORMDL3 is very similar to ORMDL1 and ORMDL2. At the transcript level, the three genes are approximately 45% identical (Table 3.1). To confirm specificity of the siRNAs used in the experiment, we transfected A549 cells with a scramble siRNA or increasing concentrations of ORMDL3 siRNA. Using primers designed specifically to each ORMDL gene, expression levels were determined by qPCR 72 hours post-transfection. Both ORMDL1 and ORMDL2 expression remained at levels similar to the negative control (transfection with scramble siRNA), while ORMDL3 was successfully knocked down (Figure 3.6). A titration of increasing ORMDL3 knockdown with increasing siRNA is observed. We achieved approximately 40-70% knockdown of ORMDL3 (30-60% expression of cells only control) using siRNA concentrations of 50nM- 500nM.  31  3.3.2 KNOCKDOWN EFFECT ON THE UPR  The effect of ORMDL3 knockdown on activation of the UPR was explored in A549 cells. Initiation of the UPR is mediated by one or more of the ER-membrane protein sensors: PKR-like eukaryotic initiation factor 2α kinase (PERK), inositol requiring enzyme 1 (IRE1), and activating transcription factor-6 (ATF6) [58]. Activation of any of the three pathways initiates signaling cascades that mediate changes to relieve ER stress. Expression changes in three genes that signify UPR activation, XBP-1u, XBP1-s, and CHOP, were determined by qPCR. The gene XBP-1 is a substrate for IRE1 ribonuclease [59]. Upon activation of the IRE1 pathway, the IRE1 ribonuclease removes a 26-bp intron from the unspliced variant, XBP-1u, which results in the spliced variant, XBP-1s [75]. This spliced variant is the active form of the gene that contributes to ER stress responses. CHOP transcription, in contrast, can be induced by the PERK and ATF6 pathways [59]. Of particular interest was the PERK pathway, as higher levels of eIF2α phosphorylation was found in cells overexpressing ORMDL3 [54]. As positive controls for UPR activation, we stimulated 1HAE cells with tunicamycin or thapsigargin. Both are inducers of ER stress – tunicamycin inhibits N-linked glycosylation and thapsigargain inhibits the SERCA pump causing ER calcium stores to be depleted [76]. Quantitative measurement of transcript levels showed that both XBP-1s and CHOP increased, while XBP1-u decreased (Figure 3.7) upon stimulation. Because the half-life of ORMDL3 protein is unknown, we waited 72 hours after transfection in A549 cells to ensure knockdown at a protein level. UPR stress responses have also been shown to take up to three days to occur, post viral infection [75]. We then extracted RNA to measure transcript levels of XBP-1u, XBP1-s, and CHOP. Quantitative PCR 32  confirmed that ORMDL3 remained knocked down at this time point (Figure 3.6). Figure 3.8 shows that transcript levels of XBP-1u and XBP1-s were not altered with ORMDL3 knockdown, which supports previous findings by Cantero-Recasens et al. [54]. However, transcript levels of CHOP also did not vary in comparison to the negative control. This result was unexpected, as CHOP transcription is regulated in part by eukaryotic initiation factor 2α (eIF2α) phosphorylation in the PERK pathway.  3.3.3 KNOCKDOWN, CELL STIMULATION, AND CYTOKINE PRODUCTION Twenty-four hours after 1HAE cells were transfected with scramble (200nM) or ORMDL3 (50, 100, 200nM) siRNA, cells were stimulated with various stimuli for an additional 24 hours. Stimulants and their concentrations were chosen based on published literature or past experiments [77-79]. Supernatants were collected 48h post-transfection and secretion of pro-inflammatory cytokines, IL-6 and IL-8, were measured by ELISA. RNA was also collected for quantification of ORMDL3 expression by qPCR. Knockdown of ORMDL3 transcript was confirmed in the cells, as shown in Figure 3.9. Similar knockdown of the gene was achieved in all four experimental repeats, highlighting the technical success of our chosen methodology. ELISA results (Figure 3.10) show that cells stimulated with TNFα responded with good production of both IL-6 and IL-8. In contrast, cells stimulated with LPS and flagellin produced background levels of IL-6 and IL-8. Altering the expression level of ORMDL3, however, did not affect cytokine production.    33  3.4 GENE OVEREXPRESSION  3.4.1 TRANSIENT EXPRESSION OF ORMDL3-EGFP An eGFP-tagged ORMDL3 protein was developed to address the issue of ORMDL3 detection at the protein level. The genes are in frame, linked by 10 amino acids. The pEGFP- ORMDL3 construct, depicted schematically in Figure 3.11, was transiently transfected into 1HAE cells to ensure proper transcription and expression of ORMDL3-eGFP protein. Cells were also transfected with an empty pEGFP plasmid as a positive control for eGFP expression. Cells were visualized under fluorescence microscopy 24 hours post-transfection. RNA and protein were harvested and analyzed by qPCR and Western blot, respectively. Under the microscope, both pEGFP and pEGFP-ORMDL3 transfected cells fluoresced as shown in Figure 3.12A. This indicates that the cells were able to express the fusion protein ORMDL3-eGFP. We also confirmed expression of ORMDL3-eGFP by Western blot (Figure 3.12B). Using a α-GFP monoclonal antibody, eGFP protein was detected at ~27kDa in pEGFP-transfected cells. In contrast, the antibody detected eGFP protein species at ~44kDa in cells transfected with pEGFP-ORMDL3. This is consistent with the size of the ORMDL3-eGFP protein. RNA extracted from the cells was reverse transcribed into cDNA and analyzed by qPCR. At 24 hours post-transfection, however, there was no significant increase in ORMDL3 expression in cells transfected with pEGFP-ORMDL3 compared to cells with pEGFP. Both conditions had ORMDL3 levels similar to cells with no plasmid (relative ORMDL3 expression = 1). This suggests that although ORMDL3-eGFP protein was synthesized, the cells were unable to overexpress the fusion gene. Results are depicted in Figure 3.13. 34  Although ORMDL3 mRNA levels did not appear to increase at 24 hours with transfection of pEGFP-ORMDL3, this method allowed us to detect ORMDL3 protein in future experiments.  3.4.2 OVEREXPRESSION, CELL STIMULATION, AND CYTOKINE PRODUCTION Cells were stimulated and IL-6/IL-8 production was measured by ELISA, as above. When 1HAE cells were transfected with pEGFP-ORMDL3, cytokine production did not change compared to cells transfected with pEGFP (Figure 3.14).  3.5 CO-TRANSFECTION  3.5.1 PLASMID AND SIRNA CO-TRANSFECTION  One advantage to constructing a fusion protein, ORMDL3-eGFP, is that eGFP is only expressed with expression of ORMDL3. Therefore, when cells are co-transfected with pEGFP-ORMDL3 and siRNA, knockdown of ORMDL3 could subsequently be detected by Western blot using a α-GFP antibody. This allows us to ensure that the siRNA used in this experiment can also affect protein expression levels. It should be noted that the siRNA cannot discriminate between endogenous and exogenous ORMDL3 transcripts. Although it is not known how eGFP affects the function of ORMDL3, endogenous ORMDL3 is still expressed and assumed to function normally.  1HAE cells were co-transfected with pEGFP-ORMDL3 and siRNA (scramble or ORMDL3). The transfection reactions were as follows: 2µg pEGFP-ORMDL3 + 500nM scramble siRNA 2µg pEGFP-ORMDL3 + 100nM, 200nM, or 500nM ORMDL3 siRNA 35  Protein and RNA were harvested at 24h post-transfection for analysis by Western blot and qPCR, respectively. Cells were also visualized under microscope for eGFP expression. By direct imaging, pEGFP-ORMDL3 and scramble siRNA-transfected cells appeared to have the most cells expressing eGFP (Figure 3.15). As ORMDL3 siRNA concentration was increased, the number of green cells appeared to decrease. We did not attempt to formally quantify these images. This observation was confirmed by Western blot of whole cell lysate collected from the cells in each condition. Figure 3.16B shows knockdown of ORMDL3-eGFP protein, confirming that ORMDL3 siRNA affects both mRNA and protein expression. At the transcript level, we observed a small but significant increase (p < 0.05) in ORMDL3 expression in the cells transfected with pEGFP-ORMDL3 + scramble siRNA compared to cells alone (Figure 3.16A). Transfection with pEGFP-ORMDL3 and increasing concentrations of ORMDL3 siRNA resulted in a titration effect of decreasing ORMDL3 mRNA expression. We also measured gene transcript expression at 3 and 6 hours post- transfection (data not shown). Similar results were obtained.  3.5.2 CO-TRANSFECTION, CELL STIMULATION, AND CYTOKINE PRODUCTION 1HAE cells co-transfected with pEGFP-ORMDL3 and siRNA (scramble or ORMDL3) were stimulated. The same timeline was followed, as above (t = 0h transfect, t = 24h stimulate, t = 48h harvest). Supernatants were collected and production of pro- inflammatory cytokines, IL-6 and IL-8, were measured by ELISA. Again, despite confirmation of ORMDL3 protein knockdown, we did not observe any effect on pro- 36  inflammatory cytokine production after stimulation (Figure 3.17). This result was expected and confirms the results we obtained previously.  3.5.3 EFFECT ON THE UPR  The effect of ORMDL3 knockdown on expression of UPR markers (XBP-1u, XBP1-s, and CHOP) was determined in plasmid + siRNA co-transfected 1HAE cells. Transfected cells were stimulated with tunicamycin, thapsigargin, or TNFα, and RNA was extracted. Gene expression was quantified by qPCR and no change in UPR marker expression was observed (Figure 3.18). These results support our earlier findings in unstimulated, siRNA transfected A549 cells.  3.5.4 REPETITION IN A549 CELLS  A549 cells, co-transfected with plasmid and siRNA, yielded the same results as those obtained in the 1HAE cells. No meaningful difference was observed in IL-6 and IL-8 pro- inflammatory cytokine production after knockdown and stimulation of the cells (Figure 3.19).  3.5.5 PCR ARRAY  The functional effect of altering ORMDL3 expression on the expression of various other genes was determined. We performed PCR arrays to profile expression of cytokines, chemokines, and key genes involved in asthma and allergy. Gene expression was compared between TNFα-stimulated (for 24 hours) 1HAE cells with high and low ORMDL3 expression (Figure 3.20). Cells considered to have “high” ORMDL3 were cells transfected with 2µg 37  pEGFP-ORMDL3 + 500nM scramble siRNA. Cells considered to have “low” ORMDL3 expression were cells transfected with 2µg pEGFP-ORMDL3 + 500nM of ORMDL3 siRNA. Based on results obtained from the array, thirty-six genes were identified to be differentially regulated by more than 1.5 fold, summarized in Table 3.2. Of these identified genes, only eight were expressed at relatively high levels (average threshold cycle <30) in the cell line used. These genes are identified by bold text in Table 3.2. We were unable, however, to validate those results (n=3). We did not observe any significant change in expression in the identified genes (Figure 3.21). Expression of CCL2 appeared to decrease with ORMDL3 knockdown when we compared highest to lowest ORMDL3 expression. However, this difference was small and not significant. When we looked at secreted protein levels of CCL2, similar results were obtained (Figure 3.22). The same arrays were done on cells stimulated for 2 hours with TNFα. The same conditions for “high” and “low” ORMDL3 expression were used. From our results, we identified only one gene, IL23A, that was differentially regulated by more than 1.5 fold (fold regulation -2.37) and amplified at cycle <30. Validation is currently being done. No other genes, as identified in the previous arrays, were found to be differentially regulated at this time point.  38  Table 3.1   Percent identity of ORMDL1 and ORMDL2 compared to ORMDL3     Table 3.2   Genes differentially regulated in cells with low ORMDL3 expression Gene Fold Regulation   Gene Fold Regulation ADRB2 -2.16  CCL24 -1.94 BCL6 -2.15  IL10 -1.68 CCL2 -2.11  IL17F -1.77 CCL5 -1.75  IL5 -2.32 CSF2 -2.12  OSM -2.16 IL12A -1.81  TNFRSF11B -1.80 IL13RA1 -1.63  TNFSF13B -2.85 VEGFA -1.53  CCL22 -2.66 EPX -2.57  CMA1 1.82 FOXP3 -1.91  GATA3 1.77 IL12B -1.72  IL1RL1 1.97 ARG1 -1.94  IL4 1.56 IL3RA -1.68  RNASE3 2.24 IL5 -2.09  SATB1 1.93 TBX21 -1.62  IL12B 1.64 TNFSF4 -4.88  IL16 3.46 TSLP -3.60  IL9 2.26 BMP4 -3.19  TNFSF11 2.17   Gene mRNA identity (%) Protein identity (%) ORMDL1 45.4 84.3 ORMDL2 42.4 80.4 39        Figure 3.1   Gene expression profile of GSDMA, GSDMB, and ORMDL3 in A549 and 1HAE cell lines. Genes were quantified by qPCR to a housekeeping gene and then compared relative to ORMDL3. Experiment was performed once (n=1). O R M D L3 G SD M A G SD M B O R M D L3 G SD M A G SD M B 0.0 0.5 1.0 A549 1HAE Gene R el at iv e ge ne  e xp re ss io n 40       Figure 3.2   Gene expression profile of ORMDL1, ORMDL2, and ORMDL3 in various cell lines. A549 are adenocarcinomic human alveolar basal epithelial cells, 1HAE are normal human airway epithelial cells, Caco-2 are human epithelial colorectal adenocarcinoma cells, and Thp-1 are human acute monocytic leukemia cells. Genes were quantified by qPCR to a housekeeping gene and then compared relative to ACTB (expression = 1). Experiment was performed once (n=1). O R M D L1 O R M D L2 O R M D L3 O R M D L1 O R M D L2 O R M D L3 O R M D L1 O R M D L2 O R M D L3 O R M D L1 O R M D L2 O R M D L3 0.00 0.01 0.01 0.02 0.02 0.03 A549 1HAE Caco-2 Thp-1 Gene R el at iv e ge ne  e xp re ss io n 41       Figure 3.3   MAPK1 gene knockdown in A549 cells. Cells were transfected with scramble (non-specific) or MAPK1 siRNA and compared to a cells only control. Relative gene expression was determined by qPCR. Experiment repeat was performed once (n=1).  C el ls  o nl y Sc ra m bl e si R N A  5 00 nM M A PK 1 si R N A  1 00 nM M A PK 1 si R N A  2 50 nM M A PK 1 si R N A  5 00 nM 0.0 0.5 1.0 1.5 R el at iv e M A PK 1 ex pr es si on 42       N on e S- 00 5 T- 01 3 T- 02 0 W -0 01 U -0 17 0.0 0.5 1.0 1.5 Program R el at iv e O R M D L3 ex pr es si on      Figure 3.4   Transfection optimization with siRNA in 1HAE cells. Cells were transfected with 200nM of ORMDL3 siRNA using different Nucleofector® electroporation programs. ORMDL3 transcript levels were compared relative to cells treated with siRNA, but were not transfected. RNA was harvested 24 hours post-transfection. Experiment was performed once (n=1). 43                       Figure 3.5   Transfection optimization with pEGFP-ORMDL3 in 1HAE cells. Four Nucleofector® electroporation programs were compared for transfection of 1HAE cells with pEGFP-ORMDL3. Cells were visualized by bright field and fluorescent microscopy.  None S-005 T-020 U-017 W-001 Bright field microscopy Fluorescent microscopy Program: 44        Figure 3.6   Specificity of ORMDL3 siRNA. Relative expression levels of ORMDL1, ORMDL2, and ORMDL3 were quantified 72 hours after transfection with ORMDL3 siRNA in A549 cells. Gene expression is compared relative to cells transfected with scramble siRNA, a non-specific siRNA (negative control). Data represent the mean ± SEM of three experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test. *** signifies p < 0.001.  Sc ra m bl e 50 0n M O R M D L3  1 00 nM O R M D L3  2 50 nM O R M D L3  5 00 nM Sc ra m bl e 50 0n M O R M D L3  1 00 nM O R M D L3  2 50 nM O R M D L3  5 00 nM Sc ra m bl e 50 0n M O R M D L3  1 00 nM O R M D L3  2 50 nM O R M D L3  5 00 nM 0.0 0.5 1.0 1.5 ORMDL1 ORMDL2 ORMDL3 *** *** *** siRNA R el at iv e m R N A ex pr es si on 45                    Figure 3.7   Unfolded protein response activation. ER stress was induced in 1HAE cells stimulated with A) tunicamycin or B) thapsigargin for 2 or 4 hours. Relative expression levels of XBP-1u, XBP-1s, and CHOP were quantified and compared to unstimulated cells. Data represent the mean ± SEM of three experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test. *** signifies p < 0.001, ** signifies p < 0.01, * signifies p < 0.05.  Tunicamycin un sti mu lat ed 2h 4h 0 2 4 6 8 10 XBP-1u XBP-1s CHOP*** *** *** *** ** ns Re la tiv e m RN A ex pr es si on Thapsigargin un sti mu lat ed 2h 4h 0 10 20 30 XBP-1u XBP-1s CHOP *** (XBP-1s) *** *** *** * ns (CHOP) Re la tiv e m RN A ex pr es si on A B 46         Figure 3.8   Expression of UPR gene markers in unstimulated A549 cells with ORMDL3 knockdown. Three gene markers, XBP-1u, XBP-1s, and CHOP, were quantified 72 hours post-transfection. XBP-1 is spliced upon activation of the IRE1 branch of the UPR and may lead to inflammatory responses. CHOP is induced through the PERK and ATF6 pathways and may lead to apoptosis. Data represent the mean ± SEM of three experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test.  Sc ra m bl e 50 0n M O R M D L3  1 00 nM O R M D L3  2 50 nM O R M D L3  5 00 nM Sc ra m bl e 50 0n M O R M D L3  1 00 nM O R M D L3  2 50 nM O R M D L3  5 00 nM Sc ra m bl e 50 0n M O R M D L3  1 00 nM O R M D L3  2 50 nM O R M D L3  5 00 nM 0.0 0.5 1.0 1.5 2.0 XBP-1u XBP-1s CHOP siRNA R el at iv e m R N A ex pr es si on 47        Figure 3.9   ORMDL3 gene knockdown in 1HAE cells. Expression levels are compared relative to cells transfected with scramble siRNA (negative control) 48 hours post- transfection. Data represent the mean ± SEM of four experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test. *** signifies p < 0.001.  Sc ra m bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM 0.0 0.5 1.0 1.5 *** *** *** siRNA R el at iv e O R M D L3  e xp re ss io n 48  A  B   Figure 3.10   Cytokine production in siRNA-transfected 1HAE cells. A) Secreted IL-6 levels and B) secreted IL-8 levels after cell stimulation. Cells were stimulated with TNFα, LPS, flagellin, or DMSO (vehicle control) for 24 hours. Data represent the mean ± SEM of four experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test. Sc ra m bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM Sc ra m bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM Sc ra m bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM Sc ra m bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM 0 500 1000 1500 2000 TNFα (200ng/mL) LPS (100µg/mL) Flagellin (100ng/mL) DMSO (1%) siRNA IL -6  (p g/ m L) Sc ra m bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM Sc ra m bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM Sc ra m bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM Sc ra m bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM 0 25000 50000 75000 100000 125000 TNFα (200ng/mL) LPS (100µg/mL) Flagellin (100ng/mL) DMSO (1%) siRNA IL -8  (p g/ m L) 49       Figure. 3.11   Schematic depiction of pEGFP-ORMDL3 plasmid. © Xiaoli Dong, Paul Stothard, Ian J. Forsythe, and David S. Wishart “PlasMapper: a web server for drawing and auto-annotating plasmid maps” Nucleic Acids Res. 2004 Jul 1;32(Web Server issue):W660- 4. Created using PlasMapper by permission.  50  A          B  Scramble siRNA 200nM +    ORMDL3 siRNA 200nM   +   pEGFP 2µg   +  pEGFP-ORMDL3 2µg  +      Figure 3.12   Expression of eGFP or ORMDL3-eGFP in 1HAE cells. Cells were transfected with 200nM scramble siRNA, 2µg pEGFP, or 2µg pEGFP-ORMDL3. Protein expression was confirmed by A) fluorescent microscopy visualization or B) Western blot analysis on whole cell lysate 24 hours post-transfection. eGFP ORMDL3-eGFP β-­‐actin Scramble siRNA pEGFP pEGFP-ORMDL3 51          Figure 3.13   Relative ORMDL3 expression in plasmid-transfected 1HAE cells. Cells were transfected with 2µg pEGFP or pEGFP-ORMDL3 and ORMDL3 transcript levels were quantified 24 hours post-transfection. Data represents the mean ± SEM of three experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post- test.  pE G FP pE G FP -O R M D L3 0.0 0.5 1.0 1.5 R el at iv e O R M D L3  e xp re ss io n 52  A  B   Figure 3.14   Cytokine production in plasmid-transfected 1HAE cells. A) Secreted IL-6 levels and B) secreted IL-8 levels after cell stimulation. Cells were stimulated with TNFα, LPS, flagellin, or DMSO (vehicle control) for 24 hours. Data represent the mean ± SEM of three experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test.  pE G FP pE G FP -O R M D L3 pE G FP pE G FP -O R M D L3 pE G FP pE G FP -O R M D L3 pE G FP pE G FP -O R M D L3 0 500 1000 1500 2000 TNFα (200ng/mL) LPS (100µg/mL) Flagellin (100ng/mL) DMSO (1%) IL -6  (p g/ m L) pE G FP pE G FP -O R M D L3 pE G FP pE G FP -O R M D L3 pE G FP pE G FP -O R M D L3 pE G FP pE G FP -O R M D L3 0 20000 40000 60000 80000 100000 TNFα (200ng/mL) LPS (100µg/mL) Flagellin (100ng/mL) DMSO (1%) IL -8  (p g/ m L) 53                     Figure 3.15   Visualization of 1HAE cells co-transfected with plasmid and siRNA using fluorescent microscopy. Cells were co-transfected with 2µg plasmid and varying concentrations of siRNA. Images were compared to a negative control (“none”) of cells transfected with no plasmid or siRNA. Columns of panels (from left to right) show the cells under 4x, 10x, or 25x magnification. None pEGFP-ORMDL3 + 500nM scramble siRNA pEGFP-ORMDL3 + 100nM ORMDL3 siRNA pEGFP-ORMDL3 + 200nM ORMDL3 siRNA pEGFP-ORMDL3 + 500nM ORMDL3 siRNA 4x 10x 25x 54  A  B Cells only +     pEGFP-ORMDL3 + 500nM scramble siRNA  +    pEGFP-ORMDL3 + 100nM ORMDL3 siRNA  +   pEGFP-ORMDL3 + 200nM ORMDL3 siRNA  +  pEGFP-ORMDL3 + 500nM ORMDL3 siRNA  +      Figure 3.16   ORMDL3 expression in plasmid and siRNA co-transfected 1HAE cells. A) Relative ORMDL3 transcript levels. Data represents the mean ± SEM of three experimental repeats. B) Protein expression of ORMDL3-eGFP. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test. * signifies p < 0.05, *** signifies p < 0.001. Ce lls  on ly Sc ram ble  si RN A 50 0n M OR MD L3  si RN A 10 0n M OR MD L3  si RN A 20 0n M OR MD L3  si RN A 50 0n M 0.0 0.5 1.0 1.5 pEGFP-ORMDL3 2µg + *** *** **** R el at iv e O R M D L3  e xp re ss io n ORMDL3-eGFP β-actin 55  A  B  Figure 3.17   Cytokine production in plasmid and siRNA co-transfected 1HAE cells. A) Secreted IL-6 levels and B) secreted IL-8 levels after cell stimulation. Cells were stimulated with TNFα, LPS, flagellin, or IL-1β for 24 hours. Data represent the mean ± SEM of three experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test. C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM 0 2000 4000 6000 8000 10000 pEGFP-ORMDL3 2µg + TNFα (200ng/mL) LPS (100µg/mL) Flagellin (10ng/mL) IL-1β (200ng/mL) No stimulation IL -6  (p g/ m L) C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM 0 50000 100000 150000 TNFα (200ng/mL) LPS (100µg/mL) Flagellin (10ng/mL) IL-1β (200ng/mL) No stimulation pEGFP-ORMDL3 2µg + IL -8  (p g/ m L) 56                    Figure 3.18   Expression of UPR gene markers in stimulated 1HAE cells with ORMDL3 knockdown. Three gene markers, XBP-1u, XBP-1s, and CHOP, were quantified and compared between knockdown conditions. Cells were stimulated with A) Tunicamycin for 4 hours, B) Thapsigargin for 4 hours, or C) TNFα for 24 hours. Data represent the mean ± SEM of three experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test.  Thapsigargin S cr am bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM S cr am bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM S cr am bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM 0.0 0.5 1.0 1.5 XBP-1u XBP-1s CHOP siRNA R el at iv e m R N A  e xp re ss io n Tunicamycin S cr am bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM S cr am bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM S cr am bl e 20 0n M O R M D L3  5 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM 0.0 0.5 1.0 1.5 XBP-1u XBP-1s CHOP siRNA R el at iv e m R N A  e xp re ss io n TNFα C el ls  o nl y S cr am bl e 50 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM O R M D L3  5 00 nM C el ls  o nl y S cr am bl e 50 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM O R M D L3  5 00 nM C el ls  o nl y S cr am bl e 50 0n M O R M D L3  1 00 nM O R M D L3  2 00 nM O R M D L3  5 00 nM 0.0 0.5 1.0 1.5 XBP-1u XBP-1s CHOP pEGFP-ORMDL3 2µg + R el at iv e m R N A  e xp re ss io n A B C 57  A  B  Figure 3.19   Cytokine production in plasmid and siRNA co-transfected A549 cells. A) Secreted IL-6 levels and B) secreted IL-8 levels after cell stimulation. Cells were stimulated with TNFα, flagellin, or IL-1β for 24 hours. Data represent the mean ± SEM of three experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test. C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM 0 500 1000 1500 2000 TNFα (200ng/mL) IL-1β (200ng/mL) Flagellin (200ng/mL) No stimulation pEGFP-ORMDL3 2µg + IL -6  (p g/ m L) C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM C el ls  o nl y Sc ra m bl e si R N A  5 00 nM O R M D L3  s iR N A  1 00 nM O R M D L3  s iR N A  2 00 nM O R M D L3  s iR N A  5 00 nM 0 50000 100000 150000 TNFα (200ng/mL) IL-1β (200ng/mL) Flagellin (200ng/mL) No stimulation pEGFP-ORMDL3 2µg + IL -8  (p g/ m L) 58     0.0001 0.001 0.01 0.1 1 10 0.0001 0.001 0.01 0.1 1 10 CCL2 IL13RA1 VEGFA CSF2 CCL5 ADRB2 BCL6 IL12A Log10 (high ORMDL3 expression 2 -ΔCt) Lo g 1 0 (lo w O R M D L3  e xp re ss io n 2- Δ C t )     Figure 3.20   PCR array. Results shown are genes profiled from two PCR arrays: Allergy & Asthma and Human Cytokines & Chemokines.1HAE cells with high or low ORMDL3 expression were stimulated with TNFα for 24 hours and differences in gene expression were compared. Comparing low ORMDL3 expression to high ORMDL3 expression, black circles are genes with less than 1.5 fold change and green circles are genes down-regulated by more than 1.5 fold.  59                       Figure 3.21   Validation of differentially expressed genes. 1HAE cells were stimulated with TNFα for 24 hours and RNA was extracted for analysis of gene expression. Data represent the mean ± SEM of three experimental repeats. Statistical analysis was performed using one- way ANOVA with Bonferroni post-test, comparing ORMDL3 siRNA columns to scramble siRNA column.  ADRB2 Ce lls  on ly Sc ram ble  si RN A 50 0n M OR MD L3  si RN A 10 0n M OR MD L3  si RN A 20 0n M OR MD L3  si RN A 50 0n M 0.0 0.5 1.0 1.5 pEGFP-ORMDL3 2µg + R el at iv e m R N A ex pr es si on BCL6 Ce lls  on ly Sc ram ble  si RN A 50 0n M OR MD L3  si RN A 10 0n M OR MD L3  si RN A 20 0n M OR MD L3  si RN A 50 0n M 0.0 0.5 1.0 1.5 pEGFP-ORMDL3 2µg + R el at iv e m R N A ex pr es si on CSF2 Ce lls  on ly Sc ram ble  si RN A 50 0n M OR MD L3  si RN A 10 0n M OR MD L3  si RN A 20 0n M OR MD L3  si RN A 50 0n M 0.0 0.5 1.0 1.5 2.0 pEGFP-ORMDL3 2µg + R el at iv e m R N A ex pr es si on CCL2 Ce lls  on ly Sc ram ble  si RN A 50 0n M OR MD L3  si RN A 10 0n M OR MD L3  si RN A 20 0n M OR MD L3  si RN A 50 0n M 0.0 0.5 1.0 1.5 pEGFP-ORMDL3 2µg + R el at iv e m R N A ex pr es si on CCL5 Ce lls  on ly Sc ram ble  si RN A 50 0n M OR MD L3  si RN A 10 0n M OR MD L3  si RN A 20 0n M OR MD L3  si RN A 50 0n M 0.0 0.5 1.0 1.5 pEGFP-ORMDL3 2µg + R el at iv e m R N A ex pr es si on VEGFA Ce lls  on ly Sc ram ble  si RN A 50 0n M OR MD L3  si RN A 10 0n M OR MD L3  si RN A 20 0n M OR MD L3  si RN A 50 0n M 0.0 0.5 1.0 1.5 pEGFP-ORMDL3 2µg + R el at iv e m R N A ex pr es si on IL13RA1 Ce lls  on ly Sc ram ble  si RN A 50 0n M OR MD L3  si RN A 10 0n M OR MD L3  si RN A 20 0n M OR MD L3  si RN A 50 0n M 0.0 0.5 1.0 1.5 pEGFP-ORMDL3 2µg + R el at iv e m R N A ex pr es si on IL12A Ce lls  on ly Sc ram ble  si RN A 50 0n M OR MD L3  si RN A 10 0n M OR MD L3  si RN A 20 0n M OR MD L3  si RN A 50 0n M 0.0 0.5 1.0 1.5 pEGFP-ORMDL3 2µg + R el at iv e m R N A ex pr es si on 60          Figure 3.22   CCL2 production in plasmid and siRNA co-transfected 1HAE cells. Cells were stimulated with TNFα, LPS, flagellin, or IL-1β for 24 hours. Data represent the mean ± SEM of three experimental repeats. Statistical analysis was performed using one-way ANOVA with Bonferroni post-test.  C el ls  o nl y Sc ra m bl e si R N A  5 00 nM  O R M D L3  s iR N A  1 00 nM  O R M D L3  s iR N A  2 00 nM  O R M D L3  s iR N A  5 00 nM  C el ls  o nl y Sc ra m bl e si R N A  5 00 nM  O R M D L3  s iR N A  1 00 nM  O R M D L3  s iR N A  2 00 nM  O R M D L3  s iR N A  5 00 nM  C el ls  o nl y Sc ra m bl e si R N A  5 00 nM  O R M D L3  s iR N A  1 00 nM  O R M D L3  s iR N A  2 00 nM  O R M D L3  s iR N A  5 00 nM  C el ls  o nl y Sc ra m bl e si R N A  5 00 nM  O R M D L3  s iR N A  1 00 nM  O R M D L3  s iR N A  2 00 nM  O R M D L3  s iR N A  5 00 nM  C el ls  o nl y Sc ra m bl e si R N A  5 00 nM  O R M D L3  s iR N A  1 00 nM  O R M D L3  s iR N A  2 00 nM  O R M D L3  s iR N A  5 00 nM  0 5000 10000 15000 20000 25000 TNFα (200ng/mL) LPS (100µg/mL) Flagellin (10ng/mL) IL-1β (200ng/mL) No stimulation pEGFP-ORMDL3 2µg + C C L2   ( pg /m L) 61  4     DISCUSSION Asthma is a complex disease affecting many individuals in the developed world. Genome-wide association studies have recently been used to identify genetic causes for such complex diseases. One particular gene, ORMDL3, is of interest because of its association with asthma, IBD, and most recently Type I Diabetes – all of which are caused by immune- mediated pathology [20, 36, 41, 80, 81]. The gene ORMDL3 encodes a transmembrane ER protein. Functionally, ORMDL3 is potentially involved in Ca2+-signaling in the ER and sphingolipid synthesis [53, 54, 56]. It has also been correlated to activation of the UPR, though the mechanisms remain unclear [54]. ER stress and inflammation have been linked, which may explain the association of the gene with several inflammatory diseases. However, the functional role of ORMDL3 in the pathogenesis of asthma has yet to be elucidated.  Airway epithelial cells play an important role in innate immunity and in the development of asthma. Upon exposure to environmental factors, the airway epithelial cells are among the first types of cells to initiate an inflammatory response. Current findings in literature indicate that ORMDL3 is involved in immunity and that asthmatics have higher expression of the gene than non-asthmatics . We therefore investigated the effect of changes in ORMDL3 expression levels on the innate immune responsiveness of airway cells. Using plasmid and siRNA technologies, we were able to manipulate gene expression levels in vitro. At the transcript level, we achieved up to 70% gene knockdown. We confirmed knockdown at the protein level by immunoblot analysis using a α-GFP antibody. However, we expected a larger increase in ORMDL3 mRNA with transfection of the overexpression plasmid. These results suggest that although ORMDL3-eGFP was expressed, the cells had difficulty producing large quantities of the protein. Many reasons, including low mRNA stability or 62  poor translation efficiency could have contributed to this effect. Despite this limitation, co- transfection of the fusion protein-encoding plasmid with siRNA enabled us to knockdown the gene and detect gene expression changes at both transcript and protein levels. In this study, we were able to successfully manipulate ORMDL3 expression in vitro. This experimental approach allowed us to explore the effects of variation in ORMDL3 expression on the innate immune response. Cells were transfected and left for 24 hours to allow for knockdown to take effect. Post-transfection, cells were stimulated with various immune response-inducing factors. TNFα and IL-1β were chosen because both are early response cytokines that perpetuate acute inflammatory processes. LPS and flagellin, in contrast, are common microbial antigens recognized by the innate immune system. The airway cells responded well to stimulation with TNFα and IL-1β and were less responsive to LPS and flagellin. This is likely because the cells express low levels of the receptors TLR4 and TLR5 which mediate responses to LPS and flagellin, respectively. There is also evidence that airway cells express low levels of necessary TLR4 co-receptors, MD2 and CD14, thus contributing to the lack of response to LPS [82, 83]. Analysis by ELISA showed no differences in pro-inflammatory cytokine production, which suggests that variation in ORMDL3 expression levels does not affect innate immune production of IL-6 and IL-8 in airway cells. We next explored the effects of ORMDL3 expression on activation of the UPR. Three UPR activation markers were chosen and quantified after knockdown of ORMDL3: XBP-1u, XBP-1s, and CHOP. We tested two cell lines, A549 and 1HAE, at different time points and under different conditions, stimulated or unstimulated. All experiments showed that ORMDL3 knockdown did not cause differential UPR activation. We are limited, however, by 63  the genes that we chose as UPR activation markers. It is possible that ORMDL3 knockdown affects other genes at either transcript or protein levels within the UPR pathway network. Detection of UPR proteins such as phosphorylated eIF2α could be explored in future experiments. We are limited in this study in that the protein function of ORMDL3 is poorly understood. Although we successfully manipulated levels of ORMDL3, we are uncertain of its effect on ER homeostasis. Assays measuring calcium signaling are one possible way of detecting functional changes associated with changes in ORMDL3 expression. Work that has been done by Cantero-Recasens et al. showed that knockdown of ORMDL3 resulted in a higher ER calcium signal [54]. As a future experiment, we will employ a similar assay to measure ER calcium signal in cells with ORMDL3 knockdown. Taking a broader approach, we performed PCR arrays looking at expression of common immunity genes. Keeping in mind that some genes are not commonly produced by airway epithelial cells, these arrays identified a number of candidate genes affected by ORMDL3 knockdown: CCL2, TSLP, CSF2, CCL5, VEGFA, ADRB2, IL1RL1, and IL13RA1. We thought that although ORMDL3 levels may not affect the production of IL-6 or IL-8 cytokines, perhaps it was altering gene expression of other important immune genes. Verification of differential expression of these genes at a transcript level, however, did not show any meaningful difference when ORMDL3 was knocked down. Although trends similar to the initial PCR array were observed in some genes, the differences were minor and not statistically significant. It may also be possible that because we looked at transcript levels of the genes 24 hours after stimulation, transcript levels may have already peaked and returned to normal levels. For this reason, another PCR array was performed to look at changes in 64  gene expression after two hours of stimulation with TNFα. Under this condition, we identified one gene, IL23A, which was differentially regulated. Validation of this result also showed no significant difference with ORMDL3 knockdown. From these data, it is clear that the genes that were identified at 24 hours stimulation were not differentially regulated at the earlier time point as previously thought. This suggests that altering ORMDL3 expression does not have a profound effect on the expression of innate immune genes upon stimulation in the airway epithelia. A major research goal in the field of asthma is the identification of biomarkers. Biomarkers can be defined as metabolites, genes, transcripts, or proteins that can indicate a biological condition. Examples include prognostic biomarkers which facilitate the prediction of individuals who will likely develop allergies/asthma, and diagnostic biomarkers which determine the subtype or severity of a patient’s asthma. Therapeutic markers are used to determine the likelihood of a patient’s response to a particular therapeutic intervention strategy, whereas mechanistic biomarkers identify a direct cause of the susceptibility to asthma. ORMDL3 has been consistently found associated with childhood asthma susceptibility in genome-wide and candidate-gene association studies. Although we were unable to determine the functional role of ORMDL3 in innate immunity, this gene remains an important candidate gene for asthma susceptibility. More research is required to elucidate its functional role in asthma pathogenesis and its potential role as an initial trigger of inflammation. If ORMDL3 does have a significant role in airway inflammation, future drug therapy targets may include ORMDL3 or other molecules in the immune response pathway(s) it may be affecting. In such case, ORMDL3 will have implications in pharmacogenomics, which uses an individual’s genotype to optimize drug therapy. If ORMDL3 does not have an 65  important role in inflammation, it may still have the potential to serve as a biomarker in the prognosis or diagnosis of asthma. Based on my data, future experiments should focus on exploring the role of ORMDL3 in other cell types. It is possible that the effects of variation in ORMDL3 expression are a cell type-dependent phenomenon. While no effect on the inflammatory response was detected in airway cells, other cells types such as dendritic cells or T cells may be affected by altered ORMDL3 expression. Other phenotypes should also be measured to fully understand the functional role of ORMDL3 in immunity. A link has already been established between expression of the gene and UPR activation. Although we did not detect UPR activation from the genes we investigated, different gene markers or effects of UPR pathway induction, such as autophagy, should be examined. The effects of ORMDL3 knockdown on ER Ca2+- signaling and sphingolipid synthesis could also be explored. These data highlight that variation in ORMDL3 expression is not correlated with differential innate immune response to stimuli in airway cells. This investigation is focused and more experiments should be done to better understand the role of ORMDL3 in asthma pathogenesis. By increasing our understanding of the mechanisms responsible for the allergic and atopic disease development, new treatments can then be developed. Thus, we can reduce inflammatory responses by targeting the potential triggers, rather than the symptoms, of the disease. In doing so, we can also reduce the morbidity, mortality, and socio-economic burden of asthma and related allergic diseases. 66  REFERENCES  1. Turvey, S.E. and D.H. Broide, Innate immunity. J Allergy Clin Immunol, 2010. 125(2 Suppl 2): p. S24-32. 2. Janeway, C.A., Jr. and R. Medzhitov, Innate immune recognition. Annu Rev Immunol, 2002. 20: p. 197-216. 3. Karin, M., T. Lawrence, and V. Nizet, Innate immunity gone awry: linking microbial infections to chronic inflammation and cancer. Cell, 2006. 124(4): p. 823-35. 4. Bals, R. and P.S. Hiemstra, Innate immunity in the lung: how epithelial cells fight against respiratory pathogens. Eur Respir J, 2004. 23(2): p. 327-33. 5. Proud, D. and R. Leigh, Epithelial cells and airway diseases. Immunol Rev, 2011. 242(1): p. 186-204. 6. Cho, S.H., Pharmacogenomic approaches to asthma treatment. Allergy Asthma Immunol Res, 2010. 2(3): p. 177-82. 7. Masoli, M., et al., The global burden of asthma: executive summary of the GINA Dissemination Committee Report. Allergy, 2004. 59(5): p. 469-478. 8. Subbarao, P., et al., Epidemiology of asthma: risk factors for development. Expert Rev Clin Immunol, 2009. 5(1): p. 77-95. 9. Weiss, K.B. and S.D. Sullivan, The health economics of asthma and rhinitis. I. Assessing the economic impact. J Allergy Clin Immunol, 2001. 107(1): p. 3-8. 10. Green, R.H., et al., Management of asthma in adults: current therapy and future directions. Postgrad Med J, 2003. 79(931): p. 259-67. 11. Barnes, P.J., Anti-inflammatory actions of glucocorticoids: molecular mechanisms. Clin Sci (Lond), 1998. 94(6): p. 557-72. 12. Lee, E., et al., Leukotriene receptor antagonists and synthesis inhibitors reverse survival in eosinophils of asthmatic individuals. Am J Respir Crit Care Med, 2000. 161(6): p. 1881-6. 13. Davies, D.E., R. Djukanovic, and S.T. Holgate, Application of functional genomics to study of inflammatory airways disease. Thorax, 1999. 54(1): p. 79-81. 14. Gu, M.L., X.Q. Dong, and J. Zhao, New insight into the genes susceptible to asthma. J Asthma, 2010. 47(2): p. 113-6. 15. Ober, C. and S. Hoffjan, Asthma genetics 2006: the long and winding road to gene discovery. Genes Immun, 2006. 7(2): p. 95-100. 67  16. Strachan, D.P., The role of environmental factors in asthma. Br Med Bull, 2000. 56(4): p. 865-82. 17. Zhang, J., P.D. Pare, and A.J. Sandford, Recent advances in asthma genetics. Respir Res, 2008. 9: p. 4. 18. Shifren, A., et al., Mechanisms of remodeling in asthmatic airways. J Allergy (Cairo), 2012. 2012: p. 316049. 19. Cohn, L., J.A. Elias, and G.L. Chupp, Asthma: mechanisms of disease persistence and progression. Annu Rev Immunol, 2004. 22: p. 789-815. 20. Moffatt, M.F., et al., A large-scale, consortium-based genomewide association study of asthma. N Engl J Med, 2010. 363(13): p. 1211-21. 21. Kim, H.Y., R.H. DeKruyff, and D.T. Umetsu, The many paths to asthma: phenotype shaped by innate and adaptive immunity. Nat Immunol, 2010. 11(7): p. 577-84. 22. Umetsu, D.T., et al., Asthma: an epidemic of dysregulated immunity. Nat Immunol, 2002. 3(8): p. 715-20. 23. Strachan, D.P., Family size, infection and atopy: the first decade of the "hygiene hypothesis". Thorax, 2000. 55 Suppl 1: p. S2-10. 24. Finn, P.W. and T.D. Bigby, Innate immunity and asthma. Proc Am Thorac Soc, 2009. 6(3): p. 260-5. 25. Broide, D.H., et al., Cytokines in symptomatic asthma airways. J Allergy Clin Immunol, 1992. 89(5): p. 958-67. 26. Konno, S., et al., Cytokine concentrations in sputum of asthmatic patients. Int Arch Allergy Immunol, 1996. 109(1): p. 73-8. 27. Nocker, R.E., et al., Interleukin-8 in airway inflammation in patients with asthma and chronic obstructive pulmonary disease. Int Arch Allergy Immunol, 1996. 109(2): p. 183-91. 28. Virchow, J.C., Jr., et al., Inflammatory determinants of asthma severity: mediator and cellular changes in bronchoalveolar lavage fluid of patients with severe asthma. J Allergy Clin Immunol, 1996. 98(5 Pt 2): p. S27-33; discussion S33-40. 29. Feghali, C.A. and T.M. Wright, Cytokines in acute and chronic inflammation. Front Biosci, 1997. 2: p. d12-26. 30. Gabay, C., Interleukin-6 and chronic inflammation. Arthritis Res Ther, 2006. 8 Suppl 2: p. S3. 68  31. Vercelli, D., Discovering susceptibility genes for asthma and allergy. Nat Rev Immunol, 2008. 8(3): p. 169-82. 32. Bouzigon, E., et al., Meta-analysis of 20 genome-wide linkage studies evidenced new regions linked to asthma and atopy. Eur J Hum Genet, 2010. 18(6): p. 700-6. 33. Ferreira, M.A., et al., Association between ORMDL3, IL1RL1 and a deletion on chromosome 17q21 with asthma risk in Australia. Eur J Hum Genet, 2010: p. 1-7. 34. Galanter, J., et al., ORMDL3 gene is associated with asthma in three ethnically diverse populations. Am J Respir Crit Care Med, 2008. 177(11): p. 1194-200. 35. Leung, T.F., et al., Asthma and atopy are associated with chromosome 17q21 markers in Chinese children. Allergy, 2009. 64(4): p. 621-8. 36. Moffatt, M.F., et al., Genetic variants regulating ORMDL3 expression contribute to the risk of childhood asthma. Nature, 2007. 448(7152): p. 470-3. 37. Sleiman, P.M., et al., ORMDL3 variants associated with asthma susceptibility in North Americans of European ancestry. J Allergy Clin Immunol, 2008. 122(6): p. 1225-7. 38. Verlaan, D.J., et al., Allele-specific chromatin remodeling in the ZPBP2/GSDMB/ORMDL3 locus associated with the risk of asthma and autoimmune disease. Am J Hum Genet, 2009. 85(3): p. 377-93. 39. Wu, H., et al., Genetic variation in ORM1-like 3 (ORMDL3) and gasdermin-like (GSDML) and childhood asthma. Allergy, 2009. 64(4): p. 629-35. 40. Ober, C. and T.C. Yao, The genetics of asthma and allergic disease: a 21st century perspective. Immunol Rev, 2011. 242(1): p. 10-30. 41. McGovern, D.P., et al., Genome-wide association identifies multiple ulcerative colitis susceptibility loci. Nat Genet, 2010. 42(4): p. 332-7. 42. Neurath, M.F., S. Finotto, and L.H. Glimcher, The role of Th1/Th2 polarization in mucosal immunity. Nat Med, 2002. 8(6): p. 567-73. 43. Saeki, N., et al., Distinctive expression and function of four GSDM family genes (GSDMA-D) in normal and malignant upper gastrointestinal epithelium. Genes Chromosomes Cancer, 2009. 48(3): p. 261-71. 44. Carl-McGrath, S., et al., Differential expression and localisation of gasdermin-like (GSDML), a novel member of the cancer-associated GSDMDC protein family, in neoplastic and non-neoplastic gastric, hepatic, and colon tissues. Pathology, 2008. 40(1): p. 13-24. 69  45. Komiyama, H., et al., Alu-derived cis-element regulates tumorigenesis-dependent gastric expression of GASDERMIN B (GSDMB). Genes Genet Syst, 2010. 85(1): p. 75-83. 46. Lin, Y.N., et al., Loss of zona pellucida binding proteins in the acrosomal matrix disrupts acrosome biogenesis and sperm morphogenesis. Mol Cell Biol, 2007. 27(19): p. 6794-805. 47. Halapi, E., et al., A sequence variant on 17q21 is associated with age at onset and severity of asthma. Eur J Hum Genet, 2010. 18(8): p. 902-8. 48. Hirota, T., et al., Genetic polymorphism regulating ORM1-like 3 (Saccharomyces cerevisiae) expression is associated with childhood atopic asthma in a Japanese population. J Allergy Clin Immunol, 2008. 121(3): p. 769-70. 49. Tavendale, R., et al., A polymorphism controlling ORMDL3 expression is associated with asthma that is poorly controlled by current medications. J Allergy Clin Immunol, 2008. 121(4): p. 860-3. 50. Fang, Q., et al., Association of genetic variants in chromosome 17q21 and adult-onset asthma in a Chinese Han population. BMC Med Genet, 2011. 12: p. 133. 51. Hrdlickova, B. and L.I. Holla, Relationship between the 17q21 locus and adult asthma in a Czech population. Hum Immunol, 2011. 72(10): p. 921-5. 52. Lluis, A., et al., Asthma-associated polymorphisms in 17q21 influence cord blood ORMDL3 and GSDMA gene expression and IL-17 secretion. J Allergy Clin Immunol, 2011. 127(6): p. 1587-94 e6. 53. Hjelmqvist, L., et al., ORMDL proteins are a conserved new family of endoplasmic reticulum membrane proteins. Genome Biol, 2002. 3(6): p. RESEARCH0027. 54. Cantero-Recasens, G., et al., The asthma-associated ORMDL3 gene product regulates endoplasmic reticulum-mediated calcium signaling and cellular stress. Hum Mol Genet, 2009. 19(1): p. 111-21. 55. Mahn, K., et al., Diminished sarco/endoplasmic reticulum Ca2+ ATPase (SERCA) expression contributes to airway remodelling in bronchial asthma. Proc Natl Acad Sci U S A, 2009. 106(26): p. 10775-80. 56. Han, S., et al., Orm1 and Orm2 are conserved endoplasmic reticulum membrane proteins regulating lipid homeostasis and protein quality control. Proc Natl Acad Sci U S A, 2010. 107(13): p. 5851-6. 57. Breslow, D.K., et al., Orm family proteins mediate sphingolipid homeostasis. Nature, 2010. 463(7284): p. 1048-53. 70  58. Hotamisligil, G.S., Endoplasmic reticulum stress and the inflammatory basis of metabolic disease. Cell, 2010. 140(6): p. 900-17. 59. McGuckin, M.A., et al., ER stress and the unfolded protein response in intestinal inflammation. Am J Physiol Gastrointest Liver Physiol, 2010. 298(6): p. G820-32. 60. Cullinan, S.B. and J.A. Diehl, Coordination of ER and oxidative stress signaling: the PERK/Nrf2 signaling pathway. Int J Biochem Cell Biol, 2006. 38(3): p. 317-32. 61. Deng, J., et al., Translational repression mediates activation of nuclear factor kappa B by phosphorylated translation initiation factor 2. Mol Cell Biol, 2004. 24(23): p. 10161-8. 62. Hu, P., et al., Autocrine tumor necrosis factor alpha links endoplasmic reticulum stress to the membrane death receptor pathway through IRE1alpha-mediated NF- kappaB activation and down-regulation of TRAF2 expression. Mol Cell Biol, 2006. 26(8): p. 3071-84. 63. Rahman, I., et al., Oxidative stress and TNF-alpha induce histone acetylation and NF-kappaB/AP-1 activation in alveolar epithelial cells: potential mechanism in gene transcription in lung inflammation. Mol Cell Biochem, 2002. 234-235(1-2): p. 239- 48. 64. Whitmarsh, A.J. and R.J. Davis, Transcription factor AP-1 regulation by mitogen- activated protein kinase signal transduction pathways. J Mol Med (Berl), 1996. 74(10): p. 589-607. 65. Barnes, P.J., Nuclear factor-kappa B. Int J Biochem Cell Biol, 1997. 29(6): p. 867- 70. 66. Lam, G.Y., J. Huang, and J.H. Brumell, The many roles of NOX2 NADPH oxidase- derived ROS in immunity. Semin Immunopathol, 2010. 32(4): p. 415-30. 67. Kohchi, C., et al., ROS and innate immunity. Anticancer Res, 2009. 29(3): p. 817-21. 68. Hasnain, S.Z., et al., The interplay between endoplasmic reticulum stress and inflammation. Immunol Cell Biol, 2012. 69. Adair-Kirk, T.L., J.J. Atkinson, and R.M. Senior, Smoke particulates stress lung cells. Nat Med, 2008. 14(10): p. 1024-5. 70. Endo, M., et al., The ER stress pathway involving CHOP is activated in the lungs of LPS-treated mice. J Biochem, 2005. 138(4): p. 501-7. 71. Pratt, A.J. and I.J. MacRae, The RNA-induced silencing complex: a versatile gene- silencing machine. J Biol Chem, 2009. 284(27): p. 17897-901. 71  72. Meister, G. and T. Tuschl, Mechanisms of gene silencing by double-stranded RNA. Nature, 2004. 431(7006): p. 343-9. 73. Gruenert, D.C., et al., Characterization of human tracheal epithelial cells transformed by an origin-defective simian virus 40. Proc Natl Acad Sci U S A, 1988. 85(16): p. 5951-5. 74. Pfaffl, M.W., A new mathematical model for relative quantification in real-time RT- PCR. Nucleic Acids Res, 2001. 29(9): p. e45. 75. Isler, J.A., A.H. Skalet, and J.C. Alwine, Human cytomegalovirus infection activates and regulates the unfolded protein response. J Virol, 2005. 79(11): p. 6890-9. 76. Samali, A., et al., Methods for monitoring endoplasmic reticulum stress and the unfolded protein response. Int J Cell Biol, 2010. 2010: p. 830307. 77. Palmberg, L., et al., Induction of IL-8 production in human alveolar macrophages and human bronchial epithelial cells in vitro by swine dust. Thorax, 1998. 53(4): p. 260-4. 78. Yang, Y., et al., Regulation of interleukin-1beta and interleukin-1beta inhibitor release by human airway epithelial cells. Eur Respir J, 2004. 24(3): p. 360-6. 79. Shanks, K.K., et al., Interleukin-8 production by human airway epithelial cells in response to Pseudomonas aeruginosa clinical isolates expressing type a or type b flagellins. Clin Vaccine Immunol, 2010. 17(8): p. 1196-202. 80. Lees, C.W., et al., New IBD genetics: common pathways with other diseases. Gut, 2011. 60(12): p. 1739-53. 81. Saleh, N.M., et al., Genetic association analyses of atopic illness and proinflammatory cytokine genes with type 1 diabetes. Diabetes Metab Res Rev, 2011. 27(8): p. 838-43. 82. Becker, M.N., et al., CD14-dependent lipopolysaccharide-induced beta-defensin-2 expression in human tracheobronchial epithelium. J Biol Chem, 2000. 275(38): p. 29731-6. 83. Jia, H.P., et al., Endotoxin responsiveness of human airway epithelia is limited by low expression of MD-2. Am J Physiol Lung Cell Mol Physiol, 2004. 287(2): p. L428-37.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0072762/manifest

Comment

Related Items