Airway surface liquid antiviral activity in cystic fibrosis

Transcription

1 University of Iowa Iowa Research Online Theses and Dissertations Summer 2015 Airway surface liquid antiviral activity in cystic fibrosis Abigail Rae Berkebile University of Iowa Copyright 2015 Abigail Rae Berkebile This thesis is available at Iowa Research Online: Recommended Citation Berkebile, Abigail Rae. "Airway surface liquid antiviral activity in cystic fibrosis." MS (Master of Science) thesis, University of Iowa, Follow this and additional works at: Part of the Microbiology Commons

2 AIRWAY SURFACE LIQUID ANTIVIRAL ACTIVITY IN CYSTIC FIBROSIS by Abigail Rae Berkebile A thesis submitted in partial fulfillment of the requirements for the Master of Science degree in Microbiology in the Graduate College of The University of Iowa August 2015 Thesis Supervisor: Professor Paul B. McCray, Jr.

3 Copyright by ABIGAIL RAE BERKEBILE 2015 All Rights Reserved

4 Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL MASTER'S THESIS This is to certify that the Master s thesis of Abigail Rae Berkebile has been approved by the Examining Committee for the thesis requirement for the Master of Science degree in Microbiology at the August 2015 graduation. Thesis Committee: Paul B. McCray, Jr., Thesis Supervisor Stanley Perlman Steven Varga

5 ACKNOWLEDGEMENTS I would like to thank my mentor, Dr. Paul McCray for his guidance, support, and patience. I also want to thank the members of my committee, Drs. Stanley Perlman and Steven Varga for their feedback and guidance. I am grateful to Dr. Jennifer Bartlett for training me when I first came to the lab and helping me get this project started. I would like to thank Drs. Brian McCullagh and Robert Gray for their help in collecting pig airway surface liquid and their feedback during lab meetings. I am grateful to Dr. Brajesh Singh and my fellow graduate student, Ashley Cooney, for often being the first sounding board when I was planning a new experiment or analyzing new results. Thank you to Cystic Fibrosis Research Center for insightful and critical feedback and specifically to Dr. Mahmoud Abou Alaiwa and Viral Shah for your technical expertise. I would also like to acknowledge the support of the Department of Microbiology. Lastly, I would like to thank all of the members of the McCray lab for creating a wonderful environment in which to work and learn. I acknowledge the support of the National Science Foundation Graduate Research Fellowship (Grant No ). ii

6 ABSTRACT Cystic fibrosis (CF) is a lethal genetic disease that affects 30,000 people in the United States alone. While the disease affects organs throughout the body, it is the lung disease that is the primary cause of morbidity and mortality for people with the disease. CF lung disease is characterized by thick and sticky mucus that obstructs the airways, acute and chronic bacterial infections, and chronic inflammation and remodeling. Thanks to the creation of the CF pig, it is now possible to study the manifestations of CF lung disease at birth. The CF pig develops spontaneous lung disease, similar to that found in humans with CF, making it the ideal model for our studies. One of the critical findings that revealed in studies of the CF pig is that airway surface liquid (ASL) bactericidal activity is impaired in CF at birth, and this activity is ph dependent. Because infants and children with CF tend to suffer greater morbidity from respiratory viruses than non-cf infants and children, we sought to determine if ASL has antiviral activity and if that activity is reduced in newborn CF pigs. We found that pre-incubating either tracheal or nasal ASL from wild-type pigs reduced the infectivity of various recombinant viruses expressing an egfp or GFP reporter gene. Those viruses include Sendai virus (SeV-eGFP), respiratory syncytial virus (RSV-GFP), the PR8 strain of influenza virus A (PR8-eGFP), and adenovirus (Ad-eGFP), indicating ASL has broadspectrum antiviral activity. Nasal secretions from newborn CF pigs had strikingly reduced antiviral activity against SeV-eGFP and Ad-eGFP compared to nasal secretions from WT littermates. Unlike what was observed for ASL antibacterial activity, nasal secretion antiviral activity was not affected by ph, nor was it affected by bicarbonate concentration, one of the molecules that drives ph in the airways. However, when we mixed CF and WT nasal secretions at different ratios, we found the antiviral activity to follow a linear trend, with antiviral activity iii

7 increasing as the percentage of WT nasal secretions increased. This suggests that one or more components of nasal secretions are found less abundantly in CF nasal secretions compared to WT nasal secretions, leading to reduced antiviral activity in CF. The CF pig has facilitated a much greater understanding of the early stages of CF lung disease. This model will allow us to determine what antiviral components are lacking in the CF airways and why they are reduced in CF. iv

8 PUBLIC ABSTRACT Cystic fibrosis (CF) is a lethal genetic disease that affects 30,000 people in the United States alone. While the disease affects organs throughout the body, lung disease causes the majority of illness and death in CF patients. Thanks to the creation of the CF pig, it is now possible to study the development of CF lung disease starting from birth. One of the critical findings revealed by the CF pig is that in airway surface liquid (ASL) from these animals, antibacterial activity is impaired at birth. Because young CF patients tend to suffer more severe symptoms from respiratory viral infections than non-cf individuals, we sought to determine if ASL has antiviral activity and if that activity is reduced in newborn CF pigs. We found that pre-incubating ASL from non-cf pigs reduced the infectivity of various respiratory viruses. Additionally, nasal secretions from newborn CF pigs had strikingly reduced antiviral activity compared to nasal secretions from WT littermates. When we mixed CF and WT nasal secretions at different ratios, we found the antiviral activity increased as the percentage of WT nasal secretions increased. This suggests that one or more components of nasal secretions are less abundant in CF nasal secretions compared to WT nasal secretions, leading to reduced antiviral activity in CF. In future studies, this animal model will allow us to determine what antiviral components are lacking in CF airways and why they are reduced in CF v

9 TABLE OF CONTENTS LIST OF FIGURES... vii CHAPTER 1: INTRODUCTION.. 1 Cystic Fibrosis. 1 Antibacterial Defect of Cystic Fibrosis... 3 The Effect of Airway Surface Liquid ph in Cystic Fibrosis Host Defense 4 The role of ASL in the defense of the airways.. 4 CFTR helps regulate ASL ph... 4 Impact of ph on the activity of ASL antimicrobials. 5 Mucin viscosity is ph dependent... 6 The impact of reduced ASL ph on other aspects of host defense. 7 Targeting ASL ph as a possible therapeutic treatment for cystic fibrosis.. 9 Evidence of an Antiviral Defect in Cystic Fibrosis 10 CHAPTER 2: ANTIVIRAL ACTIVITY OF AIRWAY SURFACE LIQUID IN CYSTIC FIBROSIS.. 16 Summary 16 Introduction 16 Methods. 19 Viruses and tissue culture 19 Collection and processing of ASL and nasal secretions Viral inactivation assay Flow cytometry titering of viruses Adjusting the ph and bicarbonate concentrations of nasal secretions.. 22 Electrophoresis and silver stain Mixing WT and CF nasal secretions. 23 Results 23 ASL and nasal secretions reduce viral infectivity ASL antiviral activity is heat-labile. 24 Newborn pig serum has anti-sev and anti-ad activity Nasal secretions from CF pigs have reduced antiviral activity Bicarbonate and ph do not affect the antiviral activity of nasal secretions.. 26 Mixing WT and CF nasal secretions. 28 Antimicrobials found in porcine and human ASL have antiviral activity Discussion CHAPTER 3: CONCLUSIONS AND FUTURE DIRECTIONS REFERENCES vi

10 LIST OF FIGURES Figure 1: A simplified model of the airway epithelium Figure 2: Schematic of proteins contributing to acid and base transport across the apical membrane of the airway epithelium Figure 3: A scheme for how changes in ASL ph may influence CF pathogenesis Figure 4: Time dependence of ASL antiviral activity 34 Figure 5: Antiviral activity of porcine ASL Figure 6: The effect of methacholine stimulation on 3-week WT pig nasal secretions Figure 7: Antiviral activity of human nasal secretions Figure 8: Heat-lability of ASL antiviral activity 38 Figure 9: Antiviral activity of newborn WT pig serum.. 39 Figure 10: Antiviral activity of newborn non-cf, het, and CF pig nasal secretions 40 Figure 11: The role of ph and bicarbonate in newborn CF nasal secretions 41 Figure 12: Protein content of newborn WT, Het, and CF pig nasal secretions 42 Figure 13: Antiviral activity of cationic antimicrobial peptides found in ASL 43 Figure 14: Antiviral activity of antimicrobial proteins found in ASL.. 44 vii

11 CHAPTER 1: INTRODUCTION Cystic Fibrosis Cystic Fibrosis (CF) is an autosomal recessive disorder that affects multiple organs including the sweat ducts, lungs, intestines, pancreas, and liver. The primary cause of morbidity and death in CF is progressive lung disease caused by chronic bacterial infection and inflammation. CF is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene which encodes an anion channel regulated by nucleotide binding and camp-mediated phosphorylation. CFTR is localized to the apical membrane of cells of the surface airway epithelium and submucosal glands and it is especially abundant in ciliated cells (1). In the lung, CFTR conducts bicarbonate (HCO3 - ) and chloride (Cl - ) thereby helping regulate airway surface liquid (ASL) volume and composition. Loss of CFTR function has been implicated in an airways host defense defect, leading to impaired innate immunity and chronic bacterial colonization of the airways. In the respiratory tract, ASL is a first line of defense against inhaled or aspirated pathogens including bacteria, fungi, and respiratory viruses. ASL comprises two layers: an aqueous (sol) layer and a mucus (gel) layer (Figure 1). The aqueous periciliary layer covers the cilia, hydrating mucins and allowing for ciliary beating by distancing the mucus from the cell surface. The mucus layer is comprised of secreted and tethered mucins produced by surface goblet cells and submucosal gland epithelia. This material traps inhaled and aspirated microbes so they can be removed from the lung via mucociliary clearance. In addition to trapping pathogens, ASL also contains numerous antimicrobial peptides, proteins, and lipids, the secreted products of surface and submucosal gland epithelia and resident phagocytic cells (2). 1

12 While it has traditionally been thought that babies with CF are born with normal lungs, growing evidence indicates airway defenses are compromised early, perhaps as soon as the first month (3, 4). This defect contributes to lung disease progression during the first years of life and is characterized by colonization with bacteria (e.g. Haemophilus influenzae, Staphylococcus aureus, and Pseudomonas aeruginosa), and the onset of inflammation. In addition to difficulties with bacterial infections, infants with CF are more likely to suffer greater morbidity from common respiratory virus infections, though the total number of viral infections is not different from non-cf (5, 6). The availability of new CF animal models, including the pig (7) and ferret (8), has facilitated study of the early events of CF lung disease at the molecular and cellular level. While newborn CF pigs do not have pulmonary inflammation (7), they exhibit an impaired ability to eradicate bacteria compared to their non-cf littermates (9). BAL and lung tissues removed from newborn CF pigs were less likely to be sterile than non-cf samples from non-cf littermates. CF pigs also exhibited a reduced ability to clear S. aureus when challenged via aerosol. A recent study by Pezzulo et al indicates the CF host defense defect in newborns is caused, in part, by abnormal ASL ph (10). These studies showed that a reduction in ph leads to decreased ASL antimicrobial activity in CF pigs (10), though the mechanism(s) by which ph impair ASL antimicrobials is currently unknown. The CF ferret also demonstrates early abnormalities in host defense, with tracheal xenographs from newborn CF ferrets exhibiting defective camp-induced chloride permeability and decreased submucosal gland fluid secretion (11). Juvenile and adult CF ferrets have decreased mucociliary clearance, increased mucus obstruction, and increased bacterial infections compared to non-cf littermates (12). 2

13 The pathogenesis of CF lung disease is complex and changes as patients age. As the disease progresses, the secondary complications of chronic inflammation and the protease rich environment of the airways further compromise host defenses. The Antibacterial Defect of Cystic Fibrosis CF patients are infected and colonized with bacteria from an early age. Interestingly, bacterial diversity in the CF airway initially increases with age (13), but then decreases as lung function declines (13, 14). Organisms that are typically found in the CF airway include, S. aureus, P. aeruginosa, Burkholderia cepacia, Acrhomabacter spp, and Stenotrophomonas matophilia (15). Non-pathogenic oral bacteria may also colonize the CF airways at a greater density than the opportunistic infections typically associated with CF (15). Numerous factors may play a role in the susceptibility of CF airways to acute and chronic bacterial infections. At the forefront is the thick, sticky mucus that coats and clogs the CF airways. This mucus offers a perfect milieu for bacteria due to its resistance to clearance. Thick mucus may be caused by reduced ph in the CF airways (16, 17), impaired liquid secretion, dehydration (18-20), or a combination of those factors. Mucociliary clearance is also impaired in CF due to mucus that remains adherent to the submucosal glands (21) and/or ph-dependent impairment of ciliary beating (22, 23). Reduced ph in the CF airways also directly inhibits the antibacterial killing of ASL (10). The many and varied ways that ph can affect the killing and clearance of bacteria from the airways certainly warrants further discussion. 3

14 The Effect of Airway Surface Liquid ph in Cystic Fibrosis Host Defense The role of ASL in the defenses of the airways ASL plays a key role in the initial defense of the airways from pathogens. In addition to acting as a physical barrier to infection, ASL contains a number of peptide and protein antimicrobials. Some of the antimicrobials found in ASL include LL-37, lactoferrin, lysozyme, β-defensins, secretory leukocyte peptidase inhibitor (SLPI), and surfactant proteins A and D (SP- A and SP-D). Many of these proteins possess both antibacterial and antiviral activity. Some phagocytic cells of the innate immune system, including macrophages and neutrophils, are also found in ASL. CFTR helps regulate ASL ph CFTR conducts chloride (Cl - ), bicarbonate (HCO3 - ) (24), thiocyanate (SCN - ) (25), and other anions. Of note, HCO3 - secretion by airway epithelia helps regulate ASL ph. Loss of CFTR impairs HCO3 - secretion in CF and leads to a decreased ASL ph (26). A lower ph for CF compared to non-cf has been reported for ASL removed from human primary airway epithelial cells (26), cultured submucosal glands (27), exhaled breath condensate from human patients (28), and tracheal ASL from newborn CF pigs (10). The differences in ASL ph for non-cf and CF subjects may also depend on age and disease state. McShane and coworkers observed no differences in ASL ph between people with CF and non-cf controls aged 3 years or older (29). More recently, Abou Alaiwa and colleagues found that neonates with CF had a lower nasal ASL ph compared to non-cf neonates, whereas nasal ph in older CF children and adults was similar to values measured in people without CF (30). Further studies in CF animal models may aid in 4

15 understanding how and when changes in ASL ph occur and how such changes influence the onset and progression of CF lung disease. Further work in this area may also aid our understanding of how changes in ASL ph might contribute to other respiratory diseases. For example, asthmatic subjects have been reported to have a reduced breath condensate ph compared to healthy controls (31, 32). Impact of ph on the activity of ASL antimicrobials Antimicrobial peptides and proteins are key components of the innate immune response that help determine whether exposure to a pathogen leads to a disease state. Due to their ever ready, non-specific, and redundant nature, antimicrobials interact with microbes within minutes to hours of infection compared to the days or weeks required to mount an adaptive immune response. The CF pig model has demonstrated that an airway innate immune defect is present at birth in CF (9, 10). Despite the presence of this defect, no difference in the abundance of key airway antimicrobials including lysozyme, lactoferrin, and SP-A was observed in newborn CF pigs (10), suggesting that the animals susceptibility to bacterial infections lies in an adverse ASL environment. However, the abundance of all antimicrobial proteins in CF ASL has not been rigorously tested. Changes in ph can have a variety of effects on protein function. The reduced ph of CF ASL is predicted to negatively impact the activity of many antimicrobials. For example, in an acidic environment the cathelicidin, LL-37, undergoes a conformation change that decreases its antibacterial activity (33). SLPI (34) and human neutrophil peptide 1 (HNP-1) (35) both possess antiviral activity that decreases as ph is lowered. SP-D and SP-A can inhibit E. coli growth at ph 6.0 and 7.4 but not at ph 5.0 (36). An acidic environment also reduces the amount of human 5

16 β-defensin 1 produced by airway epithelial cells (37). Lactoferrin is not as ph sensitive as other antimicrobials in the airways and can chelate iron over a broad ph range down to ph 3.0 (38). The lytic activity of lysozyme is stable between ph 5.8 and 9.3 (39). We note that lactoferrin displays antimicrobial activity independent of its iron binding capacity, and that lysozyme also displays muramidase-independent antimicrobial activity. Further research is required to determine if these other antimicrobial properties are ph-dependent. In non-cf pigs, lowering the ph of tracheal ASL reduced bacterial killing, while raising the ph of tracheal ASL in CF pigs improved bacterial killing (10). The ph sensitivity of numerous airway antimicrobials is a probable causal link between reduced ASL ph in CF and the impaired antibacterial activity of CF ASL. Changes in ph also can disrupt the synergistic activity of antimicrobials (40). Mucin viscosity is ph dependent While secreted and tethered mucins are key factors in airway defense, the abnormally viscous mucus found in the CF airways impairs mucociliary clearance and contributes to airway obstruction and bacterial colonization. Mucins are large, anionic, polymeric glycoproteins that are the primary component of mucus. Their abundance increases in response to inflammatory stimuli and remodeling of the airway epithelium (41). Of note, mucins become more viscous at acidic ph (41, 42), and this may further contribute to the thick mucus that characterizes CF lung disease. Jayaraman and coauthors reported that CF submucosal gland secretions are more viscous than non-cf secretions, though they did not detect a difference in ph between CF and non-cf secretions (43). Recently, Gustaffson and colleagues reported that the addition of HCO3 - to mucus from the small intestines of CF mice reduces both mucus density and its adherence to 6

17 intestinal epithelium (16). This is consistent with previous data from Chen et al that demonstrated that HCO3 - reduces aggregation of porcine gastric mucins (44). In addition to their role in mucociliary clearance, mucins also physically interact with antimicrobials. LL-37, a positively charged cationic peptide, binds to negatively charged molecules and macromolecules including mucins (45), DNA, and F-actin bundles (46). Binding of LL-37 to these molecules leads to decreased antibacterial activity (45, 46). Due to the direct interaction of LL-37 with mucins, it can be hypothesized that changes in the physical properties of mucins may alter the antimicrobial activity of LL-37. The increased abundance and viscosity of mucins of the CF airways may trap LL-37 and therefore decrease its availability to interact with microbes. Other cationic antimicrobials that may interact similarly with mucins include the - and β-defensins, lactoferrin, lysozyme, and SLPI. The impact of reduced ASL ph on other aspects of host defense ASL ph may also influence other components of airway defenses. There is evidence that reductions in ph lead to reduced ciliary beat frequency in human airways (22, 23). While the mechanism behind this is not fully understood, ph has recently been linked to the regulation of ASL volume, with a reduction in ph favoring a reduced ASL volume (18). Therefore, in CF, a reduction in ASL volume may impair ciliary beating; the altered viscosity of CF mucus may further reduce the ability of cilia to clear mucus from the airways (23). Impaired mucus clearance provides a nidus for bacteria to remain in the airways, replicate, and colonize. Additionally, clearance of bacteria from the CF airways may be reduced by impaired phagocyte function. Ex vivo, alveolar macrophage phagocytic activity decreases as extracellular ph is reduced (47, 48). Activated alveolar macrophages also release reduced amounts of TNF-α 7

18 (48) and reactive oxygen species (ROS) (48, 49) when cultured at lower ph values. It is currently uncertain whether or not CF macrophages have intrinsically altered bactericidal activity. A recent study demonstrated that when cultured under identical conditions, human CF and non-cf alveolar macrophages do not differ in their bactericidal activity against P. aeruginosa or in their intracellular ROS production (50). However, earlier studies reported that monocyte derived macrophages (MDMs) from CF patients (51) and alveolar macrophages from CF mice (52) possess decreased bactericidal activity compared to non-cf macrophages. The study by Di et al also showed deficient phagosome and lysosome acidification in murine CF alveolar macrophages (52). However, since CF mice do not develop spontaneous lung disease similar to humans (53), it is difficult to reconcile findings in mice with the disease phenotype found in humans. Variations in the techniques used to isolate macrophages in the two human studies may also have contributed to the different findings. In addition, neither of the studies done with human samples address whether CF macrophages have impaired antimicrobial function prior to the onset of chronic lung disease, as both studies isolated macrophages from adult patients. Repeating these studies with macrophages from the CF pig and ferret models may shed light on whether or not there is an intrinsic defect in CF macrophage function and whether the environment of the CF airways, including reduced ASL ph, could reduce macrophage or neutrophil phagocytic activity or ROS production. As CF lung disease advances, neutrophil-dominated inflammation is established in the airways. Proteases released from neutrophils and epithelia perturb the protease-antiprotease balance (54). This imbalance further contributes to the CF host defense defect by promoting increased secretion of mucins and the degradation of antimicrobials (55). Some proteases, such as cathepsins, are activated by acidic ph and proteolytically cleave antimicrobials, including 8

19 lactoferrin and lysozyme (56), human β-defensins (57), and SP-A (58), reducing their antimicrobial activity. The airways have a complex host defense system that includes cough, the barrier properties of the epithelium, secreted mucus and antimicrobials, mucociliary clearance, and phagocytic cells. While changes in ASL ph are unlikely the sole explanation for the increased susceptibility of the CF airways to bacterial infection and colonization, it is a likely contributor to the early host defense defect. As discussed in this review, reduced ASL ph can increase the viscosity of mucus (41, 42), inhibit the activity of some endogenous antimicrobials (33-36), decrease ciliary beat frequency (22), and impair phagocytic cell function (47, 48). Each of these factors contributes to an airway environment in which bacteria are more likely to survive. The connections between loss of CFTR, reduced ASL ph, and the CF host defense defect provide a new therapeutic target for reducing the morbidity associated with CF lung disease. Targeting ASL ph as a possible therapeutic treatment for cystic fibrosis The work of Pezzulo et al. in the CF pig model suggests that increasing ASL ph may prevent or reduce airway infections (10). A phase 2 clinical trial from 2006 reported that patients administered inhaled HCO3 - expectorated three times more mucus than those given inhaled saline alone (ClinicalTrials.gov NCT ). However, to the best of our knowledge no published clinical studies demonstrate benefits of inhaled HCO3 - or other ph-altering interventions in the treatment of early or established CF lung disease. It is possible that the ph change associated with aerosolized HCO3 - may be too short lived to be therapeutically useful. It may be feasible to modulate the ph of ASL by activating other HCO3 - transport pathways or by inhibiting proton secretion (Figure 2). Further research will be needed to design 9

20 and test drugs or small molecules that specifically increase HCO3 - transport or inhibit proton secretion in the airways. Shamsuddin and Quinton reported Ca 2+ -mediated HCO3 - secretion in the small airways. This HCO3 - secretion occurs independently of the camp-mediated secretion via CFTR and may be mediated by Ca 2+ -activated chloride channels (CaCC) (59). Pendrin, a Cl - /HCO3 - exchanger, is another potential target for stimulating CFTR-independent HCO3 - secretion (60). Activating CaCC or pendrin in people with CF might stimulate sufficient HCO3 - secretion to increase the ASL ph and improve host defense. Another approach to alter ASL ph is by inhibiting H + secretion, including H + channels and ATP-driven H + pumps. Two ATP-dependent transporters, the H + /K + -ATPase and the vaculolar-type H + -ATPase (V-ATPase), contribute to apical proton secretion in both CF and non-cf airway epithelial cells (26, 61). Proton channels such as HCVN1 are also expressed in airway epithelia (62). It should be noted that any effect of ph modulation as a therapy is likely to depend on the disease state of the patient. It may be more beneficial to increase ph early in the disease course to prevent or limit initial colonization with microbes. Evidence of an antiviral defect in cystic fibrosis While most of the morbidity and mortality of CF is associated with acute and chronic bacterial infections of the airway, the impact of respiratory viruses should not be overlooked. Respiratory viral infections are no more prevalent in CF infants than in non-cf infants (3, 6). However, CF infants suffer greater morbidity due to infection (3, 63-65), resulting in increased hospitalizations, decreased lung function (6), and increased risk of bacterial colonization within the first years of life (3, 64). It has also been reported that children with CF have a 100-fold higher rhinovirus load in their lower airways compared to non-cf children (66). Rhinovirus 10

21 load was also found to negatively correlate with the amount of IFN-β and IFN-λ in the BALF of the CF children (66). Adults with CF also show evidence of an impaired antiviral defense, including evidence of increased mortality due to the 2009 influenza H1N1 pandemic (67), and a reduced rate of seroconversion when given the 2009/H1N1 vaccine (68). Risk factors associated with failure to seroconvert included malnutrition, taking the non-adjuvanted vaccine rather than the adjuvanted vaccine, and lung transplantation, likely due to the patients taking immunosuppressants (68). Most of the in vitro studies looking at the antiviral response of CF versus non-cf AECs have utilized AECs from adult donors (69, 70), or from children over the age of one (71). CF AECs support higher replication of RV (71) and PIV3 (72) compared to non-cf AECs, possibly due to an impaired apoptotic response (70, 71) and/or an impaired IFN response (70, 72, 73), leading to decreased NOS2 response (72, 73). The impaired NOS2 response may contribute to increased RSV replication in CF mice (74). Additionally, CF AECs exhibit greater induction of IL-8 compared to non-cf AECs when infected with IAV or RV (70, 72). While these studies indicate that there is an impaired innate immune response to viral infections in CF AECs, it is currently unknown whether this is a result of an inherent antiviral defect or if this defect is due to chronic inflammation and/or infection in the CF airways. There is also in vitro evidence that suggests bacterial and viral infections work synergistically in CF. CF airway epithelial cells preinfected with P. aeruginosa had a blunted type I and type III IFN response to rhinovirus compared to non-cf airway epithelial cells that were coinfected in the same manner (69). The same research group also observed that rhinovirus infection led to the release of planktonic P. aeruginosa from biofilms on CF airway epithelial cells and that these released cells were able to migrate from the apical to basolateral side of the 11

22 airway cells (75). RSV may also increase the adherence of P. aeruginosa to epithelial cells (76), thus contributing to P. aeruginosa colonization of the CF airways. Studies in new CF animal models should help determine if similar synergism between bacteria and viruses occurs in the CF lung in vivo. Many components of ASL have antiviral activity in addition to their antibacterial activity. HBDs and LL-37 can disrupt viral envelopes (77, 78). Lactoferrin interacts with the adenovirus capsid, inhibiting infection (79). Mucins can prevent virus diffusion to the cell surface (80-83). In the following chapter we sought to determine if replete ASL has antiviral activity and if that antiviral activity is impaired in CF secretions. We hypothesized that any observed differences in antiviral activity would be ph dependent, similar to what has been observed for ASL antibacterial activity (10). 12

23 Figure 1. A simplified model of the airway epithelium. The airway epithelium has two compartments, the surface epithelium and the submucosal gland (SMG) epithelium. The airway surface epithelium includes ciliated and nonciliated cells, goblet cells, and basal cells (not shown). SMG epithelium includes ciliated duct cells, mucus cells, and serous cells that secrete antimicrobial proteins. ASL provides a barrier between the epithelium and inspired air. ASL is composed of two layers, a mucus (gel) layer and periciliary liquid. The periciliary liquid covers the cilia, providing an environment for the beating of the cilia and ASL clearance when pathogens become trapped. Various antimicrobials are found in ASL. Figure previously printed in (84). 13

24 Figure 2. Schematic of proteins contributing to acid and base transport across the apical membrane of the airway epithelium. H + and HCO3 - secretion help regulate ASL ph. CFTR provides a major HCO3 - conductance that is lost in CF. Other sources of HCO3 - transport across the apical membrane of airway epithelial cells include Ca 2+ -mediated HCO3 - secretion (CaCC) and pendrin, a Cl - /HCO3 - exchanger. H + channels, V-ATPase, and H + / K + ATPase are three apical membrane proteins that contribute to the acidification of ASL. We note that basolateral transport mechanisms and the paracellular pathway are also important to these processes (not shown). Figure previously printed in (84). 14

25 Figure 3. A scheme for how changes in ASL ph may influence CF pathogenesis. CF is caused by loss of CFTR function, an anion channel that conducts Cl - and HCO3 -. Loss of CFTR function results in decreased HCO3 - conductance across airway epithelial cells leading to decreased ASL ph. Many antimicrobials have reduced activity at a low ph. Cilia beat frequency is also reduced at lower ph values and mucins increase in viscosity as ph falls, leading to decreased mucociliary clearance. Phagocytic cell function may also be reduced in environments with lower ph. This decrease in antimicrobial activity subsequently contributes to respiratory infections in the CF airway, caused by both viral and bacterial pathogens. Figure previously printed in (84). 15

26 CHAPTER 2: ANTIVIRAL ACTIVITY OF AIRWAY SURFACE LIQUID IN CYSTIC FIBROSIS Summary Although CF affects multiple organs, progressive lung disease caused by chronic bacterial infections and inflammation leads to significant morbidity and mortality in CF patients. Relatively little is known about the susceptibility of these patients to viral infections. Our study seeks to determine if ASL from newborn cystic fibrosis pigs has reduced antiviral activity and what might cause this reduction. ASL is the first line of defense that pathogens must overcome when invading the airways. Its antimicrobial activity can directly inactivate invading bacteria. However, this critical antibacterial activity is reduced in tracheal secretions from the CF pig model. Using SeV as our model virus, we demonstrate that ASL also has antiviral activity that is reduced in nasal secretions from CF pigs compared to their wild-type littermates. Interestingly, changing the ph and bicarbonate concentrations of the ASL did not affect its antiviral activity. The difference in antiviral activity between CF and WT nasal secretions may explain, in part, the increased morbidity suffered by cystic fibrosis patients when infected with respiratory viruses. Introduction Cystic fibrosis (CF) is the most common lethal genetic disease among Caucasians. The median life expectancy is in the early 40s (85). While CF is a multi-organ disease that affects the sinuses, lungs, pancreas, liver, intestines, sweat ducts and reproductive tract, the morbidity associated with CF is primarily caused by chronic lung disease that is characterized by acute and chronic bacterial infections and neutrophil-dominated inflammation (2). CF is caused by 16

27 mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene (86, 87), a camp-dependent anion channel that conducts Cl -, HCO3 - (24), and SCN - (25) across the airway epithelium. Much is still unknown about the link between CFTR dysfunction and CF lung disease progression. The first CF mouse model was developed in the early 1990s (88), and although the CF mouse models suffer from intestinal obstruction as some human CF infants do, they do not develop lung disease in the same manner as humans (53). It was not until the pig (7) and ferret (8) CF models were created that researchers had access to animal models that spontaneously develop progressive lung disease similar to human CF patients (9, 11). Thanks to these models, we can now investigate the early stages of CF airway disease. Although CF newborns were originally thought to have relatively normal lungs, growing evidence indicates that CF infants exhibit an impaired host defense as early as the first month after birth (3, 4). Since the CF pig and ferret models also exhibit an impaired host defense defect, studies using these animals have provided insight into the nature of CF lung disease at birth. Newborn CF pigs lack airway inflammation, but within hours of birth their airways exhibit an impaired antibacterial defense compared to non-cf littermates (9). Likewise, soon after birth CF ferrets rapidly develop respiratory distress and bronchopneumonia (12). The pig and ferret models strongly suggest that this host defense defect leads to lung disease onset and progression within the first years of a CF patient s life. While CF morbidity and mortality is primarily associated with acute and chronic bacterial infections of the airway and associated inflammation, the impact of respiratory viruses should not be overlooked. Although CF infants are not infected with respiratory viruses more often than non-cf infants (3, 6), CF infants suffer greater disease severity with increased hospitalizations, decreased lung function (6), and increased risk of bacterial colonization within 17

28 the first years of life (3, 64). In one study, children with CF had a higher viral load when infected with rhinovirus, a cause of the common cold, as compared to non-cf children, even though infection prevalence was similar between the two groups (66). Most of the in vitro studies looking at the antiviral responses of CF versus non-cf primary airway epithelial cells (AECs) utilized AECs from adult donors (69, 70), or from children over the age of one (71). CF AECs supported higher rhinovirus (71) and human parainfluenza virus 3 (hpiv3) (72) replication compared to non-cf AECs. Rhinovirus genome expression was increased 10-fold in infected CF AECs (71). hpiv3 titers recovered from infected CF AECs were 7-fold higher compared to non-cf AECs 24 hours post-infection (72). This may be due to an impaired apoptotic response (71) and/or an impaired IFN response (72, 73). Microarray analysis of CF and non-cf AECs infected with IAV also revealed decreased expression of IFN and apoptotic genes in CF AECs (70). While these studies indicate there is an impaired innate immune response to viral infections in CF AECs, it is unknown if this is a consequence of loss of CFTR function or if this defect is due to chronic inflammation and/or chronic infection in the CF airways. Likewise, no studies have looked at whether there is an extracellular antiviral defense defect in CF airways, even though airway surface liquid (ASL) from newborn CF pigs has reduced antibacterial activity compared to non-cf ASL (10). We hypothesize that ASL can also inactivate viruses and that the newborn CF pigs have diminished ASL antiviral activity compared to their non-cf littermates. We studied the antiviral properties of ASL because this liquid is among the first lines of defense viruses must overcome when they enter the airways. ASL is composed of two layers: the aqueous periciliary layer that covers the cilia and allows them to beat, and the viscous mucus layer that traps pathogens. There are also numerous antimicrobial proteins and peptides within 18

29 ASL, including lysozyme, lactoferrin, defensins, and cathelicidins, to name a few. Each of these proteins have been reported individually to have antiviral activity (89-92). In order to characterize ASL antiviral activity, we challenged several recombinant viruses with tracheal ASL collected from non-cf pigs and nasal secretions collected from non-cf pigs and humans. We also tested nasal secretions from newborn CF and non-cf pigs in order to determine if there was a genotype-dependent difference in antiviral activity. Methods Viruses and Tissue Culture Recombinant SeV encoding the egfp reporter gene (SeV-eGFP) was propagated by injecting the virus into the allantoic fluid of 10-day-old embryonated eggs. LLC-MK2 cells plated in 12-well plates at 2x10 5 cells/well were used for titrating SeV-eGFP and were maintained at 37 C, 5% CO2, in Opti-MEM medium (Life Technologies Corporation) containing 10% fetal bovine serum and 1% penicillin-streptomycin. Titers of SeV-eGFP are expressed as fluorescent focus-forming units (FFU) A recombinant respiratory syncytial virus encoding green fluorescent protein (RSV-GFP) upstream of the NS1 open reading frame (93) was kindly provided by Mark Peeples (Columbus Children s Research Institute and Ohio State University, Columbus, OH) and Peter Collins (NIH, Bethesda, MD). Recombinant adenovirus expressing egfp (Ad-eGFP) was prepared by the University of Iowa Gene Transfer Vector Core. RSV-GFP and Ad-eGFP were titrated on HEp-2 cells that had been seeded on a 48-well plate at 4x10 4 cells/well. HEp-2 cells were maintained in MEM medium (Life Technologies Corporation) containing 10% fetal bovine serum and 1% 19

30 penicillin-streptomycin, at 37 C, 5% CO2. Ad-eGFP and RSV-GFP titers are expressed as plaque forming units (PFU). Collection and processing of ASL and nasal secretions ASL was collected from 3-week old pigs that were anesthetized intramuscularly with ketamine (20mg/kg) and xylazine (2 mg/kg). Anesthesia was maintained intravenously with propofol (2 mg/kg). Tracheal secretions were stimulated by administering 2-4 doses of methacholine (2.5 mg/kg) intravenously over the course of the collection. A bronchoscope was introduced into the pigs trachea and a microsampling probe (model BC-401C; Olympus Optical Co, TD, Tokyo) was advanced into the trachea via the bronchoscope and positioned against the tracheal wall for 1 minute to collect ASL. The microsampling probe was then withdrawn from the bronchoscope. The probes were inserted into 1.5-ml tubes and the ASL was collected via centrifugation at 10,000 x g. The samples were gamma-irradiated at 8 krad to inactivate any bacterial contaminants and were stored at -80 C. Irradiation did not affect the antiviral activity of ASL (data not shown). Nasal secretions were collected from newborn CF and non-cf pigs and from 3-week old wild-type pigs that had been anesthetized with ketamine and xylazine as described above. A sterile polyester-tipped applicator (Puritan Medical Products Co.) was inserted into each nostril of the pig to collect nasal secretions. The probes were then placed in a microcentrifuge tube and the secretions were collected via centrifugation. Nasal secretions from human subjects were collected in a similar manner, albeit no anesthesia was required. Nasal secretions from both pigs and humans were gamma-irradiated at 8 krad and stored at -80 C. Protein concentrations of newborn CF and non-cf porcine nasal secretions were measured after collection using the 20

31 Coomassie protein assay kit (Bradford). The ph of CF and non-cf nasal secretions was measured using a Dx-pH probe (Restech) while at 37 C and 5% CO2 immediately before measurement of antiviral activity. Viral inactivation assay Viral inactivation by ASL and nasal secretions was measured by pre-incubating 10 6 FFU SeV-eGFP, 4x10 5 PFU RSV-GFP, or 2x10 5 PFU Ad-GFP with increasing amounts of ASL or nasal secretions for 2 hrs, unless otherwise noted. After pre-incubating the virus with the airway secretions, the virus-secretion mixture was brought up to 300 μl with 100 mm HEPES buffer (ph 7.4) and then applied to the appropriate titering cells. Titering was carried out as described above for each virus. To assess the antiviral activity of specific antimicrobials, the assay was also performed in a similar manner using increasing amounts LL-37 (Anaspec), protegrin-1 (Anaspec), recombinant human beta-defensin 2 (HBD2) (PeproTech), HBD3 (PeproTech), recombinant human lysozyme (Sigma), and iron-saturated and unsaturated human lactoferrin (Sigma). Antimicrobials were incubated with the viruses in 100 mm HEPES buffer at ph 6.8, 7.4, or 8.0 at a total volume of 250 μl. Serial dilutions of serum isolated from newborn WT pigs was incubated with 10 6 FFU SeV-eGFP or 2x10 5 PFU Ad-GFP at 37 C for 1 hour. The samples were then brought up to 250 μl for SeV-eGFP and 100 μl for Ad-eGFP in Opti-MEM medium and used to infect the appropriate titering cells for 1 hour at 37 C and 5% CO2. Serum was either untreated or heattreated at 56 C for 30 minutes prior to dilution. 21

32 Flow Cytometry Titering of Viruses SeV-eGFP, Ad-eGFP, and RSV-GFP were titered on the appropriate titering cells via flow cytometry. LLC-MK2 cells and HEp-2 cells were seeded in 12-well plates at 2x10 5 cells/well for SeV-eGFP and Ad-GFP titration, respectively. HEp-2 cells were seeded in 48-well plates at 2x10 4 cells/well to titer RSV. The next day, cells were washed twice with 1X PBS prior to applying virus to the surface of the cells. Cells were incubated with each virus for 1 hour at 37 C. The cells were then washed twice with 1X PBS to remove excess virus and the tissue culture media was replaced. After an overnight incubation at 37 C and 5% CO2, cells were detached with Accumax cell dissolution solution (Innovative Cell Technologies, Inc) and analyzed for GFP expression on a BD Accuri C6 FACS analyzer (BD Biosciences) using the program CFlow Plus. Adjusting the ph and Bicarbonate Concentrations of Nasal Secretions To assess the effect of ph on nasal secretions, we preincubated 1 μl of nasal secretions with 25 μl 100 mm HEPES buffer at ph 6.8, 7.4, or 8.0 for 5 minutes FFU SeV-eGFP was then added to the buffered nasal secretions and the viral inactivation assay was carried out as described above except after 2 hours the volume of the mixture was brought up to 250 μl with 100 mm HEPES buffer of the ph of the original mixture. To determine the effect of bicarbonate on nasal secretions, 1 μl of CF nasal secretions was incubated for 5 minutes with 25 μl 100 mm HEPES buffer supplemented with increasing amounts of NaHCO3 at ph 7.4. To keep the ionic strength consistent between samples, NaCl was added to each buffer for a final [Na + ] of 50 mm FFU SeV-eGFP was then added to the buffered samples. The viral inactivation assay was then carried out as described above. The 22

33 sample volumes were then brought up to 250 μl using buffers with the appropriate NaHCO3 concentrations. Lastly, ph of WT nasal secretions was reduced by incubating 1 μl of nasal secretions with 10 6 FFU/ml of SeV-eGFP at 5%, 10%, 15%, and ambient CO2 at 37 C in a humidified chamber. ph of the samples was measured using a needle-type ph microsensor (Presens). Electrophoresis and Silver Stain Nasal secretions were loaded on a 4-20% Tris-HCl gel (Bio-Rad Laboratories) at a concentration of 2 ug/lane. Samples were separated via electrophoresis and then stained using the Silver Stain for Mass Spectrometry (Pierce) as described by the manufacturer. Mixing WT and CF Nasal Secretions Nasal secretions from newborn WT and CF pigs were mixed at different WT:CF ratios (4:1, 3:1, 2:1, 1:1, 1:2, 1:3, 1:4). 1 μl of each mixture or 1 μl of WT or CF nasal secretions were incubated with 10 6 SeV-eGFP for 2 hours. The viral inactivation assay was then carried out as previously above. Results ASL and nasal secretions reduce viral infectivity Tracheal ASL (10) and nasal secretions (94, 95) have previously been shown to have antibacterial activity. However, whether complete airway secretions possessed antiviral activity remained unknown. To test this, we first investigated the antiviral activity of non-cf porcine airway secretions. We pre-incubated SeV-eGFP with nasal secretions from WT pigs at 37 C for 23

34 0, 0.5, 1, and 2 hours (Figure 4). Although the majority of antiviral activity seems to occur within the first 30 minutes of incubation, we used a 2 hour incubation time for our remaining experiments to maximize the amount of antiviral activity. Next, we pre-incubated SeV-eGFP (Figure 5a) and RSV-GFP (Figure 5b), members of the paramyxoviridae family of negative sense single stranded RNA viruses, with tracheal and nasal secretions that had been collected from 3-week old methacholine stimulated pigs. SeV-eGFP titers were reduced in a dosedependent manner when preincubated with tracheal and nasal secretions, whereas RSV titers were reduced by approximately 25% in the presence of ASL or 60% in the presence of nasal secretions. In addition to reducing the titers of members of the paramyxoviridae family, porcine nasal secretions also reduced the titer of the orthomyxovirus, IAV-PR8, in a dose-dependent manner (Figure 5c). The last virus we tested was serotype 5 Ad-eGFP. Unlike SeV, RSV, and IAV, Ad is an encapsidated DNA virus. Both tracheal ASL and nasal secretions reduced the titer of Ad-eGFP, indicating the WT secretions have broad-spectrum antiviral activity (Figure 5d). Methacholine stimulation did not affect the protein concentration (Figure 6a), ph (Figure 6b), or anti-sev activity (Figure 6c) of nasal secretions. Next, to extend these experiments to humans, we incubated SeV-eGFP (Figure 7a), RSV-GFP (Figure 7b), PR8-eGFP (Figure 7c), and Ad-eGFP (Figure 7d) with adult non-cf human nasal secretions. Human nasal secretions reduced the titer of all four viruses in a dosedependent manner similar to that observed in the pig. ASL antiviral activity is heat-labile Tracheal ASL from 3-week old farm pigs was heated for 30 minutes at 56 C, 65 C, and 95 C or was left on ice for 30 minutes (Figure 8). 56 C is sufficient to inactivate components of 24

35 the complement system and most proteins are denatured at 95 C. 65 C was chosen as an intermediate temperature. The heat-treated ASL was then incubated with SeV-eGFP for 2 hours before being applied to LLC-MK2 cells. ASL that had been incubated at 56 C had consistently, though not significantly different, reduced anti-sev activity compared to ASL incubated on ice. ASL anti-sev activity was significantly reduced when incubated with ASL that had been incubated at 65 C and 95 C, with 95 C showing the greatest reduction in ASL antiviral activity. The heat-lability of ASL antiviral activity suggests that proteins are responsible for this activity. In addition, the partial reduction in ASL antiviral activity from incubation at 56 C and 65 C suggests that multiple components are responsible for the antiviral activity. Newborn pig serum has anti-sev and anti-ad activity Proteins from serum can reach the airway through transudation or, in the case of IgA, specific transcellular transport pathways. For this reason, many host defense proteins typically assumed to be present in blood may be found in ASL or bronchoalveolar lavage fluid. We predicted that proteins such as complement or immunoglobulins might contribute to the antiviral activity of ASL. To test this hypothesis we pre-incubated SeV-eGFP (Figure 9a) and Ad-eGFP (Figure 9b) with serial dilutions of serum from newborn WT pigs. The serum was either untreated or heat-treated for 30 minutes at 56 C to inactivate complement. Untreated serum had dose-dependent antiviral activity against SeV-eGFP and Ad-eGFP that was more potent (on a per-volume basis) than that of ASL. Interestingly, while heat-treated serum lost its anti-sev activity (Figure 9a), the same heat-treated serum retained its anti-ad activity (Figure 9b), suggesting different components of serum are responsible for the antiviral activity against these two viruses. We speculate that the anti-sev activity of serum may be mediated primarily by 25

36 complement whereas Anti-Ad activity is due to molecules resistant to inactivation at 56 C, such as immunoglobulins or other host defense proteins. Both complement and immunoglobulins are present in the airway and may contribute to the antiviral activity we observed in ASL. Nasal secretions from CF pigs have reduced antiviral activity It has already been established that non-cf human ASL (94, 95) and pig ASL (10) have antibacterial activity, with CF pig ASL exhibiting impaired antimicrobial activity compared to non-cf ASL (10). We hypothesized that CF pig airway secretions would also have impaired antiviral activity compared to non-cf secretions. To test this hypothesis, we collected nasal secretions from newborn CFTR null pigs (CF pigs), and pigs that were CFTR homozygotes (WTs) or heterozygotes (Hets). We used nasal secretions primarily because the upper airways serve as a portal for many respiratory viral infections and because of the ease with which they can be collected. We found that CF nasal secretions exhibited less antiviral activity against SeVeGFP (Figure 10a) and Ad-eGFP (Figure 10b) compared to WT nasal secretions, but there was no difference in activity against RSV-GFP (Figure 10c). Bicarbonate and ph do not affect the antiviral activity of nasal secretions As previously reported, airway secretions from newborn pigs (10) and infant humans (30) have reduced ph for CF secretions compared to non-cf (Figure 11a). We hypothesized that increasing ph in CF nasal secretions could restore their antiviral activity to non-cf levels. Surprisingly, CF nasal secretion antiviral activity decreased when the ph was increased, whereas there were no differences observed when WT and Het secretions were diluted at different phs (Figure 11b). However, CF, Het, and WT secretions did not differ in their antiviral activity 26

37 when compared at identical ph values, indicating that diluting the samples ablated the genotypedependent differences of nasal secretion antiviral activity. We next reduced the ph of WT nasal secretions by incubating the virus-nasal secretion mixture at various CO2 concentrations in order to remove the effect of dilution (Figure 11c). However, the ph was only reduced when CO2 concentrations were increased from ambient CO2 to 5%. There was no difference in ph observed between 5% and 15% CO2. This may have been due to lower than physiological concentrations of HCO3 - in the nasal secretions which ranged from ~0.8 mm at ambient CO2 to ~4.4 mm at 15% CO2 as calculated using the Henderson-Hasselbalch equation (Figure 11d). Pezzulo et al estimated ASL bicarbonate concentrations to be 28.1 ± 4.2 mm in tracheal secretions from newborn pigs (10), similar to what was reported for fluid collected from forksolin-stimulated Calu-3 cells (~31 mm) (96). Increasing the concentration of CO2 did not significantly affect the antiviral activity of WT nasal secretions (Figure 11e), which was consistent with our results using HEPES buffer at various phs (Figure 11b). HCO3 - is a major driver of ph in the airways and helps determine the antimicrobial activity of various cationic antimicrobial peptides (97). Although our samples have reduced HCO3 - concentrations compared to physiological conditions, we wanted to determine if increasing the amount of HCO3 - would affect nasal secretion antiviral activity (Figure 11f). We diluted CF nasal secretions in 100 mm HEPES buffer containing increasing amounts NaHCO3. To maintain constant osmolality, NaCl was added to each buffer for a [Na + ] of 50 mm. ph was also clamped at 7.4. Increasing the HCO3 - concentration up to 50 mm had no effect on the antiviral activity of CF nasal secretions. Together, this data suggests neither reduced HCO3 - concentrations nor ph account for the decreased antiviral activity of CF nasal secretions. 27

38 Mixing WT and CF nasal secretions We next sought to determine if the antiviral activity of CF nasal secretions might be reduced due to a difference in protein composition between CF and WT secretions. While protein concentrations were not statistically different between WT, Het, and CF nasal secretions (Figure 12a), silver staining of equivalent amounts of total protein from WT, Het, and CF nasal secretions revealed a slightly lighter banding pattern in CF compared to WT and Het (Figure 12b). However, since the banding pattern is the same any perceived differences may have been caused by donor-variability between the different secretions and no conclusions could be drawn. While overall protein concentrations did not differ between WT and CF nasal secretions, when we mixed WT and CF secretions at various ratios and incubated 1 μl of each mixture with SeVeGFP there was a strong linear correlation between the amount of WT secretions and the amount of anti-sev activity (Figure 12c). This suggests that CF nasal secretions may have reduced levels of one or more key antiviral agents compared to WT secretions. Antimicrobials found in porcine and human ASL have antiviral activity ASL is a complex mixture of water, proteins, lipids, and mucus, including antiviral proteins such as lactoferrin (90, 98, 99), lysozyme (91), beta defensins (100), and cathelicidins (LL-37 in humans (78, 92, 101) and the protegrin proteins in pigs (102)). We screened selected antimicrobials for antiviral activity against SeV-eGFP and Ad-eGFP in order to begin identifying proteins that may contribute to ASL antiviral activity. Of the cationic peptides tested, LL-37 (Figure 13a, b), protegrin-1 (Figure 13c, d), and human beta defensin 3 (HBD3) (Figure 13e, f) exhibited antiviral activity against SeV-eGFP and Ad-eGFP when preincubated with either virus. Each of these peptides had activity that was enhanced at ph 6.8, except protegrin-1 against Ad- 28

39 egfp, though statistical differences in area under the curve were only observed for HBD-3 vs SeV-eGFP, and LL-37 or HBD-3 vs Ad-eGFP. This is consistent with our observation that reduced ph does not contribute to the difference between CF and non-cf antiviral activity. Human beta defensin 2 (HBD2) did not exhibit antiviral activity against SeV-eGFP (Figure 13g) but had some activity against Ad-eGFP that was significantly increased at ph 6.8 (Figure 13h). Lastly we tested lysozyme and lactoferrin for antiviral activity against SeV-eGFP and Ad-eGFP. Lysozyme had antiviral activity against both SeV-eGFP and Ad-eGFP. This activity was phindependent against SeV-eGFP (Figure 14a) and significantly enhanced by ph 6.8 against AdeGFP (Figure 14b). Lactoferrin that was saturated or unsaturated with iron demonstrated no antiviral activity against SeV-eGFP (Figure 14c), but did have ph-independent antiviral activity against Ad-eGFP (Figure 14d). All antimicrobials tested showed antiviral activity at physiologically relevant concentrations. Lactoferrin and Lysozyme are present in the airways at mg/ml(103). HBD2 is present at μg/ml in nasal secretions (94) and 8-10 μg/ml post inflammatory stimulation of ASL from cultured airway epithelia (104). LL-37 is present at mucosal surfaces at a concentration of 2-5 μg/ml (105). Discussion ASL is one of the first lines of defense that pathogens must overcome when invading the airway. Although it has previously been shown that human nasal secretions (94, 95) and porcine tracheal ASL (10) have antibacterial activity, this is the first study to show that airway secretions have antiviral activity as well. We tested tracheal and nasal ASL from 3-week old wild-type pigs against several recombinant respiratory viruses including SeV-eGFP, RSV-GFP, Ad-eGFP, and 29

40 PR8-eGFP. When ASL was pre-incubated with these test viruses, it reduced their titers in a dose-dependent manner, indicating ASL has broad antiviral activity since it can act on enveloped viruses (SeV, RSV, and PR8) as well as an encapsidated virus (Ad). Likewise, human nasal secretions possessed dose-dependent antiviral activity against all four viruses tested. It should be noted, however, that while we used an MOI of 5 for Ad-eGFP and SeV-eGFP and an MOI of 10 for RSV-GFP, we typically only observed GFP expression in 60% of cells infected with untreated SeV-eGFP and ~20% GFP expression in cells infected with untreated Ad-eGFP or RSV-GFP. For Ad-eGFP and RSV-GFP, this difference could be caused by the difference in titering assay (plaque assay vs measuring GFP expression via flow cytometry). More likely, the decrease in apparent titer for the three viruses was caused by extended incubation time (2 hours) at 37 C. We evaluated several individual airway antimicrobials for antiviral activity against SeVeGFP, our model enveloped virus, and Ad-eGFP, our model encapsidated virus. LL-37, protegrin-1, HBD-3, and lysozyme had antiviral activity against both SeV-eGFP and Ad-eGFP when preincubated with the virus. HBD-2 and lactoferrin decreased Ad-eGFP titers, but had minimal effect on SeV-eGFP. Many of these antimicrobial peptides and proteins have been shown to act by directly binding to viruses. For example, HBDs can interact with lipid bilayers (89) and even disrupt the viral envelope (77), perhaps accounting for their activity against SeVeGFP. HBD3 can also act as a lectin, binding viral and host glycoproteins (106). Like HBDs, LL-37 is able to interact directly with viruses, such as IAV, and has been shown to disrupt the viral envelope (78). Human alpha defensins, such as human defensin 5 (HD5), act against adenovirus by binding to capsid proteins and preventing capsid disassembly in the endosome during infection (107). HBDs may have a similar antiviral mechanism against adenovirus. 30

41 Similar to HD5, Lactoferrin inhibits adenovirus infection by interacting with viral capsid (79). However, lactoferrin inhibits hpiv2 infection by binding to cellular proteins rather than through direct interaction with the virus (99), perhaps explaining why preincubating SeV-eGFP with lactoferrin had no effect on viral infection. Lysozyme has been shown to inhibit HIV-1 infection and replication, though the exact antiviral mechanism is unknown (108, 109). In addition to the antimicrobial peptides and proteins tested in our study, ASL contains many other proteins with antiviral activity including SP-A and SP-D ( ). Mucins are also a major component of ASL and are inhibitory to viral infections (80-83). We speculate that together, all of these components give ASL broad-spectrum antiviral activity. Although we showed that ASL inhibits viral infection when preincubated with the virus, it would be interesting to see if it also impairs viral infection when preincubated with the cells. Using fluorescently labeled viruses would also allow us to track the virus s progress through the ASL on top of primary AECs and determine if the virus becomes stuck in mucus before it can reach the cell surface. Our studies show, for the first time, that CF ASL has impaired antiviral activity compared to non-cf ASL. While no difference in activity was observed against RSV, a human paramyxovirus, nasal secretions from newborn CF pigs had significantly reduced antiviral activity against SeV-eGFP, an enveloped paramyxovirus, and Ad-eGFP, an encapsidated virus. While characterizing the antiviral activity of CF and non-cf nasal secretions, we chose to use recombinant SeV-eGFP as our model virus because of its close antigenic and genetic homology to human PIV-1. HPIV-1, along with HPIV-3, causes one-third of the lower respiratory tract infections in children under the age of 5 in the United States (115). Of note, human 31

42 parainfluenza viruses are also implicated in pulmonary exacerbations in children with CF (3, 6, 116). CF nasal secretions were more acidic (~ph 6.84) compared to WT nasal secretions (~ph 7.13). This trend has also been reported for newborn human nasal secretions (30), airway breath condensate (28), human primary airway epithelial cells (26), and human submucosal glands (27). Interestingly, nasal secretions from het pigs had an intermediate ph (~ph 6.96). We hypothesized that reduced ph or reduced HCO3 - were responsible for the decreased antiviral activity in CF nasal secretions. However, changing the ph of the nasal secretions by diluting them in 100 mm HEPES buffers had no significant effect on activity, except in inhibiting the antiviral activity of CF nasal secretions at ph 8.0 compared to ph 6.8. Reducing the ph of WT nasal secretions with CO2 also did not significantly affect antiviral activity. Likewise, incubating CF nasal secretions with increasing amounts of bicarbonate, an anion contributing to ASL ph, had no effect on antiviral activity. Interestingly, many of the antimicrobials that we tested (LL- 37, PG-1, HBD3, and lysozyme) demonstrated increased antiviral activity at ph 6.8 against SeVeGFP, Ad-eGFP, or both, suggesting that the antiviral components of ASL are not inhibited by the reduced ph of CF secretions. Indeed, a more acidic ph enhanced the antiviral activity of some of the individual ASL antimicrobial peptides. This suggests that the effect(s) of ph against enveloped viruses or an encapsidated DNA virus are mechanistically different than that observed for bacteria (10, 84) Interestingly, the difference in the antiviral activity of WT and CF nasal secretions disappeared when the secretions were diluted in buffers with identical ph values. Therefore, we were curious to see if the difference in antiviral activity of CF nasal secretions might be due to a deficiency in one or more of the antiviral components of ASL. To test this, we mixed CF and 32

43 WT nasal secretions from newborn pigs at various ratios. The result was a nearly linear trend where increasing amounts of WT nasal secretions led to increased antiviral activity against SeVeGFP. Our next step will be to conduct quantitative mass spectrometry on WT and CF nasal secretions in hopes of determining which component, or components, may be reduced in CF. It is worth emphasizing that the nasal secretions we used in this study come from newborn CF and non-cf pigs. These pigs have had minimal exposure to environmental inflammatory stimuli or pathogens. Therefore, this study demonstrates for the first time that, at birth, CF pigs have decreased innate antiviral activity in their airway secretions compared to their non-cf littermates. While this alone may not account for the increased morbidity suffered by CF infants infected with respiratory viruses (3, 5, 65, 66), it may act together with other impaired defenses in the CF airway and contribute to disease onset and progression. In addition to studying extracellular antiviral defenses, the CF pig model offers the opportunity to study whether the impaired cellular antiviral response observed in adult human primary AECs from CF and non-cf donors and CF and non-cf cell lines (70-72) is present at birth in CF. 33

44 Figure 4. Time dependence of ASL antiviral activity. 2.5 μl nasal secretions from WT pigs ranging in age from 3 days to 3 weeks old was incubated with 10 6 FFU SeV-eGFP for 0, 0.5, 1, and 2 hours prior being brought up to 250 μl in 100 mm HEPES buffer (ph 7.4). LLC-MK2 cells plated in a 12-well plate were then infected with the virus-nasal secretion mix. Viral infection was quantified 24- hours post infection via flow cytometry analysis of GFP expression. Results are shown as mean ± SEM for 3 individual experiments, n=3. 34

45 Figure 5. Antiviral activity of porcine ASL. Tracheal secretions were collected from anesthetized 3-week old wild-type pigs that had been stimulated intravenously with methacholine. A microsampling probe was passed through a pediatric bronchoscope and touched against the tracheal wall for 1 minute. Nasal secretions were also collected from the 3-week old wild-type pigs by placing a Puritan sterile polyester tipped applicator in the nose for 2 minutes. The probes were then placed in a 1.5 ml tube and centrifuged for 5 minutes at 10,000 x g to collect the secretions. SeV-eGFP (A), RSV-GFP (B), IAV PR8 (C), or adenovirus (D), were incubated with increasing amounts of porcine tracheal or nasal secretions for 2 hours at 37 C and 5% CO2. After the 2 hour incubation, the virus-asl mixture was brought up to 250 μl in 100 mm HEPES buffer (ph 7.4) before the appropriate titering cells were infected with the virus-asl mixture. Viral infection was quantified 24-hours post infection via flow cytometry analysis of GFP expression. Results are shown as mean ± SEM for 3 individual experiments, n=3. 35

46 Figure 6. The effect of methacholine stimulation on 3-week WT pig nasal secretions. A. Protein concentrations of nasal secretions from 3-week old pigs were measured via the Coomassie (Bradford) protein assay kit by reading the samples at 595 nm on a VersaMax microplate reader, n=3 B. The ph of nasal secretions was measured ex vivo with a Restech Dx-pH probe immediately prior to mixing with SeV-eGFP, n=3. C. SeV-eGFP was incubated with increasing amounts of porcine nasal secretions for 2 hours at 37 C and 5% CO2. After the 2 hour incubation, the virus-asl mixture was brought up to 250 μl in 100 mm HEPES buffer (ph 7.4) before the appropriate titering cells were infected with the virus-asl mixture. Viral infection was quantified 24-hours post infection via flow cytometry analysis of GFP expression. Results are shown as mean ± SEM for 3 individual experiments, n=3. All differences were statistically insignificant as determined by the Student s t-test (p>0.05). 36

47 Figure 7. Antiviral activity of human nasal secretions. Human nasal secretions were collected from adult donors by placing a Puritan sterile polyester tipped applicator in the nose for 5 minutes. The probes were then placed in a 1.5 ml tube and centrifuged for 5 minutes at 10,000xg to collect the secretions. SeVeGFP (A), RSV-GFP (B), IAV PR8 (C), or adenovirus (D), were incubated with increasing amounts of human nasal secretions for 2 hours at 37 C and 5% CO2. After the 2 hour incubation, the virus-asl mixture was brought up to 250 μl in 100 mm HEPES buffer (ph 7.4) before the appropriate titering cells were infected with the virus-asl mixture. Viral infection was quantified 24-hours post infection via flow cytometry analysis of GFP expression. Results are shown as mean ± SEM for 3 individual experiments, n=

48 Figure 8. Heat-lability of ASL antiviral activity. Tracheal ASL was heated at 56 C, 65 C, 95 C, or was incubated on ice for 30 minutes. After the ASL cooled to room temperature, 2.5 μl were incubated with 10 6 FFU SeVeGFP for 2 hours at 37 C. The virus-asl mixture was brought up to 250 μl in 100 mm HEPES buffer (ph 7.4) before LLC-MK2 cells were infected with the virus-asl mixture. Viral infection was quantified 24-hours post infection via flow cytometry analysis of GFP expression. Results are shown as mean ± SEM, for 3 individual experiments, n=3. *, p<0.05 in comparison to On Ice ; #, p<0.05, ###, p< in comparison to 0 μl ASL as determined by one-way ANOVA. Only statistically significant differences are marked. 38

49 Figure 9. Antiviral activity of newborn WT pig serum. Serum from newborn WT pigs was either untreated or heat-treated at 56 C for 30 minutes. The serum was then serially diluted 1:2 (starting with 25 μl) and incubated with A. SeV-eGFP and B. Ad-eGFP for 1 hour at 37 C. The samples were then brought up to 250 μl for SeV-eGFP or 100 μl for Ad-eGFP in Opti-MEM medium. The appropriate titering cells were then infected and viral infection was quantified 24-hours post infection via flow cytometry analysis of GFP expression. Results are shown as mean ± SEM for 3 individual experiments, n=3. 39

50 Figure 10. Antiviral Activity of newborn non-cf, het, and CF pig nasal secretions. Nasal secretions were collected from anesthetized newborn pigs by placing a puritan sterile polyester tipped applicator in the nose for 2 minutes. The probes were then placed in a 1.5 ml tube and centrifuged for 5 minutes at 10,000xg to collect the secretions. A. SeV-eGFP, B. Ad-eGFP, and C. RSV-GFP were incubated with increasing amounts of nasal secretions from newborn WT, Het, and CF pigs for 2 hours at 37 C and 5% CO2. After the 2 hour incubation, the virus-secretion mixture was brought up to 250 μl in 100 mm HEPES buffer (ph 7.4) before the appropriate titering cells were infected with the virus-ns mixture. Viral infection was quantified 24-hours post infection via flow cytometry analysis of GFP expression, n=6-8 for SeV-eGFP and n=3 for Ad-eGFP and RSV-GFP. Results are shown as mean ± SEM for at least 3 individual experiments. *, p<0.05 and **, p<0.01 as determined by one-way ANOVA. Only statistically significant differences are marked. 40

51 Figure 11. The role of ph and HCO3 - in the anti-sev activity of newborn WT and CF porcine nasal secretions. A. The ph of nasal secretions was measured ex vivo with a Restech Dx-pH probe immediately prior to mixing with SeV-eGFP, n=3-8. B. The ph of nasal secretions was adjusted by mixing 1 μl nasal secretions with 25 μl 100 mm HEPES buffer at ph 6.8, 7.4, or 8.0. The viral inactivation assay was then carried out as for Figure 1. n=3-4. C. ph of WT nasal secretions and SeV-eGFP mixtures incubated at various CO2 concentrations was measured using a needle-type ph microsensor. D. Henderson-Hasselbalch equation was used to calculate the HCO3 - concentrations of the WT nasal secretion and SeV-eGFP mixtures at each CO2 concentrations. The dotted line represents the previously published calculated HCO3 - concentration of fresh tracheal ASL from newborn WT pigs (10). E. WT nasal secretions and SeV-eGFP were incubated together for 2 hours at various CO2 concentrations before SeV-eGFP was titered on LLC-MK2 cells. F. 1 μl of CF nasal secretions was incubated for 5 minutes in 100 mm HEPES buffer containing increasing amounts of NaHCO3 and decreasing amounts of NaCl ([Na + ]= 50 mm). SeV-eGFP was then added to the mixture and the viral inactivation assay was carried as for Figure 1. All results are shown as mean ± SEM for 3 individual experiments. *, p<0.05 as determined by one-way ANOVA. Only statistically significant differences are marked. 41

52 Figure 12. The role of protein composition in the anti-sev activity of newborn WT and CF porcine nasal secretions. A. Protein concentrations were measured via the Coomassie (Bradford) protein assay kit by reading the samples at 595 nm on a VersaMax microplate reader, n=8-11. Results shown as mean ± SEM. Lack of statistical significance determined by one-way ANOVA. B. Nasal secretions were loaded on a 4-20% Tris-HCl gel at a concentration of 2ug/lane. Samples were separated via electrophoresis and then stained using the Silver Stain for Mass Spectrometry (Pierce), n=3 for each genotype. C. WT and CF nasal secretions were mixed in various ratios. 1 μl of each mixture, as well as pure WT, CF, or HET nasal secretions were incubated with SeV-eGFP prior to titering on LLC-MK2 cells. After 2 hours, the virus-secretion mixture was brought up to 250 μl with 100 mm HEPES (ph 7.4). Results are shown as mean ± SEM for 5 individual experiments, n=5, R 2 =