METABOLISM OF HYPOTHIOCYANOUS ACID UNDERPIN A NOVEL THERAPY FOR INFECTIOUS AND INFLAMMATORY LUNG DISEASES
by
JOSHUA D. CHANDLER
B.A. Drury University, Springfield, MO, 2009
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment
of the requirements for the degree of Doctor of Philosophy
Toxicology Program 2014
This thesis for the Doctor of Philosophy degree by Joshua D. Chandler
has been approved for the Toxicology Program
by
Christopher C. Franklin, Chair Brian J. Day, Advisor
David P. Nichols Dennis R. Petersen
David Ross
Chandler, Joshua D. (Ph.D., Toxicology)
The Biological Chemistry of Thiocyanate and Mammalian Metabolism of
Hypothiocyanous Acid Underpin a Novel Therapy for Inflammatory and Infectious Lung Diseases
Thesis directed by Professor Brian J. Day. ABSTRACT
Thiocyanate (SCN-) functions in mammalian host defense by contributing to the microbicidal activity of haloperoxidases (e.g., myeloperoxidase, eosinophil peroxidase and lactoperoxidase), which catalytically reduce hydrogen peroxide and oxidize SCN- to hypothiocyanous acid (HOSCN). HOSCN is a thiol-targeting oxidant that inhibits pathogens but paradoxically is tolerated by host cells, unlike related species formed during inflammation (e.g., hypochlorous acid (HOCl)). However, the basis of mammalian cell tolerance of HOSCN has been poorly understood since its discovery. The focus of this thesis is to elucidate
biochemical mechanisms by which mammalian cells benefit from SCN- and HOSCN at antimicrobial doses and to utilize SCN- as a novel therapy to drive HOSCN formation in place of damaging oxidants in mammalian models of lung infection in order to decrease injury but maintain host defense.
Thioredoxin reductase (TrxR) is phylogenetically divergent in mammals, expressed as a selenoprotein with glutathione reductase homology unlike bacterial TrxR. Mammalian TrxR was observed to function as an
was inhibited by HOSCN. Expression of HOSCN reductase activity from cell lysates correlated with improved cell viability following HOSCN exposure while pharmacologic inhibition of mammalian TrxR sensitized lung epithelia to HOSCN. In contrast, HOCl was toxic to both bacterial and mammalian cells and was not metabolized by mammalian TrxR. These results demonstrate the protective role of mammalian TrxR in inflammation, which is mediated by the formation of HOSCN but not HOCl.
SCN- is of interest as a potential therapy for lung diseases, particularly since the discovery that it is dysregulated in cystic fibrosis and the therapeutic potential of SCN- was tested in multiple tissue culture and in vivo lung disease models using a concentration of ≥400 µM to drive the near-total replacement of HOCl with HOSCN during inflammation or inflammation-like conditions. WT and βENaC mice nebulized with SCN-, tested either with or without intratracheal infection with clinically isolated P. aeruginosa, fared significantly better in multiple outcomes including inflammatory infiltrate, pro-inflammatory cytokines and
bacterial load. SCN- also inhibited LPS-induced lung injury. These findings
suggest SCN- is a promising candidate for therapy in infectious and inflammatory lung diseases.
The form and content of this abstract are approved. I recommend its publication. Approved: Brian J. Day
DEDICATION
To anyone whose potential has been limited by sickness and to those who strive to improve all lives with new advances in medicine. May we continue to create a better world for everyone.
ACKNOWLEDGEMENTS
I would like to thank my mentor, Brian J. Day, Ph.D., for his guidance, support and ultimately his willingness to take a chance on me and on this project. Brian pushed me to express my ideas and communicate strong rationales for my choices at each new step. In addition to all the techniques I have learned in his lab, I feel Brian’s continuing efforts to make me express myself and develop my own ideas were the greatest training I could have received. He also deserves credit for being a thorough and thoughtful investigator every time a new experiment is designed, pushing his students to consider the science behind each experiment and take nothing for granted.
I also want to thank the members of my thesis committee—Christopher C. Franklin, Ph.D. (Chair); David P. Nichols, M.D.; Dennis R. Petersen, Ph.D. and David Ross, Ph.D. Each one of them has gone out of the way to provide me with helpful advice and guidance beyond my doctoral training. I also want to thank David Nichols for collaborating on many of my projects and Chris Franklin for being such an enthusiastic supporter of my research.
I also want to thank the other Day lab members for making life easier, especially Elysia Min for laughing at my stupid jokes and never complaining about all the in vivo work I put her through and Jie Huang for always being so friendly and willing to help. Neal Gould trained me on several of the methods I still use when I first joined the lab and was a great teacher. Cameron McElroy
was always willing to pitch in if I needed extra help. I want to thank the Nichols, Nick, Kettle and Bratton labs for collaboration on various projects. I also want to thank Manisha Patel, Ph.D. and her lab for frequently providing suggestions and advice regarding my project.
Lastly I want to thank my family and friends, both in Denver and elsewhere. You’ve all been a great source of support and I don’t express that enough. Mom and Dad, thanks for all you did to help me get where I am. Gregg and Terry, thanks for being a bit of much needed “home away from home” in Denver. And to my wife, Megan—you truly are tough as nails and I love you. Thanks for putting up with me while I completed this thesis.
TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION ... 1
Thiocyanate ... 2
Distribution and elimination of SCN- in healthy individuals ... 2
Origin of endogenous SCN- ... 4
Early 20th century administration of SCN- as antihypertensive ... 7
Biochemical reactions of SCN- and HOSCN ... 9
Effects of SCN- and HOSCN on cell viability ... 17
Antimicrobial properties of HOSCN ... 21
Cystic fibrosis ... 25
Brief history of CF and current interventions ... 25
Molecular implications of CF and biological models ... 27
Overview ... 28
II. MAMMALIAN THIOREDOXIN REDUCTASE CATALYTICALLY REDUCES HYPOTHIOCYANOUS ACID PROMOTING LUNG INNATE IMMUNITY AND CELL VIABILITY ... 30
Introduction ... 30
Methods ... 34
Mammalian cell culture ... 34
Flow cytometry for quantification of apoptosis and necrosis ... 35
Preparation of cell lysates ... 36
Sources of purified TrxR and GR ... 36
Enzyme kinetics and inhibition studies ... 37
Bacterial strains and culture ... 37
Generation of HOSCN ... 38
HOSCN reductase assay ... 38
Oxidase-peroxidase enzyme system ... 38
AFN exposure ... 39
HOCl ... 39
Acute oxidant exposure ... 39
Intracellular thiol measurement ... 40
Statistics ... 40
Results ... 40
Cell viability after enzymatic oxidant exposure ... 40
J774A.1 cell death after HOCl exposure and rescue by SCN- ... 42
Cell H-TrxR activity is associated with HOSCN, but not HOCl, tolerance ... 44
Differential effects of HOSCN on mammalian H-TrxR and bacterial L-TrxR ... 46
Expression of H-TrxR associates with HOSCN tolerance at bactericidal doses ... 49
H-TrxR inhibition sensitizes 16HBE cells to HOSCN ... 50
Discussion ... 52
III. NEBULIZED THIOCYANATE IMPROVES LUNG INFECTION OUTCOMES IN WILD TYPE MICE ... 61
Introduction ... 61
Chemicals ... 63
Animals ... 64
Fibrin plug model of retained P. aeruginosa lung infection ... 65
Measurement of pyocyanin ... 66
Isotonic SCN- formulation, nebulization and pharmacokinetics ... 66
Airway cell analysis ... 67
Cytokine measurement ... 67 Peroxidase activity ... 67 Measurement of SCN- ... 68 Measurement of GSH ... 68 Statistics ... 69 Results ... 69
Nebulization of SCN- efficiently bolsters ELF and plasma concentrations ... 69
Nebulized SCN- decreases lung P. aeruginosa bacterial burden and associated morbidity ... 71
Nebulized SCN- decreases infection-mediated lung neutrophilia .. 73
Additional studies on anti-inflammatory markers and effects ... 76
Consumption of SCN- during lung infection ... 76
Discussion ... 79
IV. NEBULIZED THIOCYANATE IMPROVES LUNG INFLAMMATION WITH AND WITHOUT INFECTION IN A βENAC-TRANSGENIC MOUSE MODEL ... 88
Introduction ... 88
Methods ... 91
Chemicals ... 91
Infection ... 92
SCN- formulation and administration ... 92
Cytology of BAL cells ... 93
Cytokines ... 93
Myeloperoxidase ELISA ... 93
Measurement of GSH and GSSG ... 94
Measurement of glutathione sulfonamide (GSA) ... 94
Measurement of SCN- ... 95
Measurement of TrxR activity ... 95
Western blotting ... 96
Statistics ... 96
Results ... 96
Nebulized SCN- ablates inflammation and oxidative stress in βENaC mice ... 96
Nebulized SCN- improves lung infection outcomes in βENaC and WT mice ... 101
Direct evidence of decreased HOCl formation in vivo with SCN- ... 105
Mammalian H-TrxR is induced by airway inflammation ... 106
Discussion ... 108
V. THIOCYANATE DECREASES LUNG INJURY IN A LIPOPOLYSACCHARIDE MODEL OF ACUTE AIRWAY NEUTROPHILIA ... 120
Introduction ... 120
Methods ... 123
LPS exposure ... 123
IgM ELISA ... 124
Human neutrophil isolation and culture ... 124
IL-8 ELISA ... 125
Results ... 125
Nebulized SCN- decreases lung injury post-LPS exposure in mice ... 125
SCN- protects neutrophil viability by decreasing necrosis with LPS ... 127
Discussion ... 131
VI. OVERVIEW AND DISCUSSION ... 135
LIST OF TABLES Table
1.1: Physiologic concentrations of SCN- in extracellular fluids ... 2
2.1: HOSCN reductase kinetics of screened enzymes ... 47
2.2: Cell viability and cell lysate HOSCN reductase activity ... 51
3.1: Nebulized SCN- pharmacokinetics in plasma and ELF of mice ... 71
4.1: BALF pro-inflammatory cytokines in uninfected βENaC mice ... 100
4.2: Pro-inflammatory cytokine data from infected mice not shown in Figure 4.3 ... 105
LIST OF FIGURES Figure
1.1: SCN- regulates the concentration of HOX in inflammatory milieu ... 12 1.2: Chemical and enzymatic pathways driving the host defense and
antioxidant properties of SCN- ... 18 1.3: Transport of SCN- by secretory epithelium and its protective role in the
lumen during infection and inflammation ... 22 2.1: Mammalian cell viability after sustained enzymatic oxidant exposure ... 41 2.2: SCN- protects J774A.1 cells from necrotic HOCl-mediated cell death ... 43 2.3: Mammalian H-TrxR activity correlates positively with HOSCN tolerance .... 45 2.4: Spectrophotometric trace of NADPH oxidation by catalytically active
H-TrxR2 ... 47 2.5: Differential effects of HOSCN on H-TrxR and L-TrxR ... 48 2.6: H-TrxR expression tolerizes 16HBE to bactericidal doses of HOSCN ... 51 2.7: Inhibition of H-TrxR by AFN sensitizes 16HBE cells to HOSCN thiol
oxidation and viability loss ... 53 2.8: Selective metabolism of HOSCN by secretory epithelium and host
defense implications ... 60 3.1: Pharmacokinetic analysis of nebulized SCN- ... 70 3.2: Nebulized SCN- reduces morbidity and bacterial burden in infected mice ... 72 3.3: Nebulized SCN- reduces infection-mediated airway inflammation ... 74 3.4: Nebulized SCN- diminishes infection-mediated markers of lung
inflammation ... 75 3.5: Nebulized SCN- preserves infection-mediated GSH adaptive responses
in the airway ... 77 3.6: ELF SCN- is depleted during infection ... 78 3.7: Proposed mechanism of action of nebulized SCN- on lung infection and
4.1: Constitutive inflammation and oxidative stress in the βENaC mouse lung .. 97 4.2: Nebulized SCN- rescues airway inflammation and oxidative stress in
βENaC mice ... 99 4.3: Nebulized SCN- rescues βENaC and WT mice from prolonged
morbidity, airway inflammation and infection ... 102 4.4: The HOCl biomarker GSA is increased in BAL fluid of inflamed and
infected mice and decreased by SCN- ... 107 4.5: H-TrxR is induced by inflammation and infection in mouse lung tissue ... 109 4.6: Nebulized SCN- does not affect the induction of lung tissue TrxR activity
by infection and inflammation ... 110 5.1: SCN- inhibits macrophage loss and protects murine airway from acute
injury induced by intratracheal LPS ... 126 5.2: SCN- enhances neutrophil viability and function by decreasing necrosis .. 128 6.1: Targets for therapeutic intervention in neutrophilic inflammation ... 141
LIST OF ABBREVIATIONS ACN Acetonitrile
ANOVA Analysis of variance APS Ammonium persulfate
ARDS Acute respiratory distress syndrome ASK1 Apoptosis signaling kinase 1
BAL Bronchoalveolar lavage BALF Bronchoalveolar lavage fluid BCRP/ABCG2 Breast cancer related protein BSA Bovine serum albumin
CaCC Ca2+-dependent Cl- channel
CF Cystic fibrosis
CFTR Cystic fibrosis transmembrane conductance regulator COPD Chronic obstructive pulmonary disease
CVD Cardiovascular disease
CXCLx Chemokine (C-X-C motif) ligand (x) 3-Cl-Tyr 3-chlorotyrosine
DAMP Damage-associated molecular pattern DMEM Dulbecco's modified eagle medium DMSO Dimethyl sulfoxide
DNA Deoxyribonucleic acid
DTNB 5,5'-dithiobis-2-nitrobenzoic acid
DUOX Dual oxidase
ε Extinction coefficient
ECL Enhanced chemiluminescence EDTA Ethylenediaminetetraacetic acid ELISA Enzyme linked immunosorbent assay
ELF Epithelial lining fluid (synonymous with airway surface liquid) EPO Eosinophil peroxidase
FAD Flavin adenine dinucleotide FBS Fetal bovine serum
FITC Fluorescein isothiocyanate GPO Gastric peroxidase
GR Glutathione reductase
Grx Glutaredoxin
GSA Glutathione sulfonamide GSH Reduced glutathione GSSG Oxidized glutathione HBE Human bronchial epithelia HBSS Hank‘s balanced salt solution
HCAEC Human coronary artery endothelial cells
HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid H2O2 Hydrogen peroxide
HOBr Hypobromous acid HOCl Hypochlorous acid HOSCN Hypothiocyanous acid
HOX Hypohalous acid (e.g., HOCl or HOBr) HPLC High performance liquid chromatography HRP Horseradish peroxidase
HUVEC Human umbilical vein endothelial cells
IgM Immunoglobulin M
IL-x Interleukin (x)
KC Keratinocyte chemoattractant; murine CXCL1 homolog LC-MS/MS Liquid chromatography with tandem mass spectrometry
detection
LDH Lactate dehydrogenase
LPO Lactoperoxidase LPS Lipopolysaccharide
MPO Myeloperoxidase
NADH Nicotinamide adenine dinucleotide
NADPH Nicotinamide adenine dinucleotide phosphate NEM N-ethylmaleimide
NF-κB Nuclear factor kappa light chain enhancer of B cells NMR Nuclear magnetic resonance
NO･ Nitric oxide (radical) NO2･ Nitrogen dioxide (radical) NO2- Nitrite
NOX NADPH-oxidase
O2 Dioxygen (radicals not indicated) O2･
Superoxide (radical)
OH･ Hydroxyl radical
ONOO- Peroxynitrite
OSCN- Hypothiocyanite
PAMP Pathogen-associated molecular pattern PRR Pattern-recognition receptor
PS Phosphatidylserine
PTP Protein tyrosine phosphatases RBC Red blood cells
Nrf2 Nuclear factor (erythroid derived-2)-like 2
OSCN Hypothiocyanite
PBS Phosphate-buffered saline
PBS-T Phosphate-buffered saline with 0.05% (v/v) Tween-20
PI Propidium iodide
PMN Polymorphonuclear leukocytes
Prx Peroxiredoxin
PSSG Protein-bound glutathione
RNHCl Monochloramine
RNHX Monohalamine
RNS Reactive nitrogen species ROS Reactive oxygen species RSe- Selenolate
RSeH Selenol
RSeOH Selenenic acid
RSeO2H Seleninic acid RSeO3H Selenonic acid
RSeSCN Selenenyl thiocyanate
RSeSeR Diselenide
RSeSR Selenosulfide
RS- Thiolate
RSH Thiol
RSOH Sulfenic acid RSO2H Sulfinic acid RSO3H Sulfonic acid
RSSCN Sulfenyl thiocyanate
RSSR Disulfide
SCN- Thiocyanate
SCN2 Thiocyanogen
SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis
Sec L-Selenocysteine
SEM Standard error of the mean TEMED Tetramethylethylenediamine TGR Thioredoxin glutathione reductase TNFα Tumor necrosis factor alpha TNB 5-thio-2-nitrobenzoic acid
Tris Tris(hydroxymethyl)aminomethane
Trx Thioredoxin
TrxR Thioredoxin reductase
H-TrxR High molecular weight thioredoxin reductase L-TrxR Low molecular weight thioredoxin reductase
CHAPTER I INTRODUCTION1
Thiocyanate (SCN-) is an acidic pseudohalide thiolate that is ubiquitously found in the extracellular fluids of mammals over a wide range of concentrations (e.g., from <5 µM to >2 mM). SCN- has been a drug of interest in human
medicine since the early 20th century when it was first used in the treatment of hypertension. Later it was discovered that SCN- is oxidized by mammalian halogenating heme peroxidases (i.e., haloperoxidases) to an antimicrobial species later determined to be hypothiocyanous acid (HOSCN). The work in this thesis provides promising new evidence for SCN- as a novel therapy for
infectious and inflammatory lung diseases such as cystic fibrosis (CF),
respiratory infection, acute lung inflammation and related illnesses. Furthermore, these results are underpinned by the discovery that HOSCN is tolerated by mammalian cells but not most bacteria due to the catalytic detoxification of this antimicrobial agent by the mammalian thioredoxin reductase (TrxR) enzyme family, which has phylogenetically diverged from the TrxR enzymes of bacteria that lack a comparable HOSCN detoxification function.
1 Some of the figures, tables and text in this chapter are originally published in Chandler, J.D. and Day, B.J. (2012) Thiocyanate: a potentially useful therapeutic agent with host defense and antioxidant properties. Biochem. Pharmacol. 84, 1381-1387
Thiocyanate
Distribution and elimination of SCN- in healthy individuals
Mammalian extracellular fluids are abundant sources of SCN- (Table 1.1). Plasma values of SCN- typically range between 5 and 70 µM in human non- smokers (1). In contrast, SCN- in human airway epithelial lining fluid (ELF) may be concentrated many fold higher with a mean value of 460 µM (2), while nasal lining fluid (NLF) has been reported at similar concentrations with wide
interpersonal variance (3). A study in young children reported a dilute bronchoalveolar lavage fluid (BALF) concentration of 280 nM, which would roughly predict an ELF SCN level around 30 µM (4). BALF corrected with the urea dilution factor (expressed as ELF) was reported at 100 µM in C57BL/6 mice (5) and sampling of undiluted airway secretions produced a mean value of 160
Table 1.1. Physiologic concentrations of SCN- in extracellular fluids. Compartment [SCN-] (µM) In Situ Peroxidasesa References
Plasma 5-70 VPO 1, 3
ELF 30-650 LPO, VPO 2, 4-6
NLF 100-1200 LPO 3
Gastric fluid 250-300 GPO 13
Saliva 670-2500 LPO 10-12
Tears 150 LPO 10
Breast milk 0.1-20 LPO 14, 15
This is a modified version of a table published in reference 174.
aMPO and EPO localize with activated neutrophils, monocytes and eosinophils at sites of inflammation. Induction of DUOX peroxidase activity may also contribute to oxidative consumption of SCN-.
µM (6) in sheep. These findings suggest that airway SCN- is concentrated from the plasma compartment, which has been suggested to occur through active transport of the basolateral sodium-iodide symporter (NIS) and apical anion channels such as the cystic fibrosis transmembrane conductance regulator (CFTR) (7, 8) and cytokine-regulated channels SLC26A4 (pendrin, an
electroneutral halide-exchange channel) and TMEM16A (Ca2+-dependent Cl -channel (CaCC), an active transporter of halides) (9).
The saliva and oral cavity are even more concentrated with SCN- owing to their heavy demand for a complex and potent mixture of antimicrobial defenses. The oral cavity ranges from 670 to 2500 µM SCN- (10-12) and utilizes the same active transport system as found in the airway making saliva the most SCN--rich matrix known in the body (12). The high concentration of SCN- in the oral cavity inhibits colonization of many bacterial species as a substrate for haloperoxidase activity and HOSCN formation. SCN- has also been reported in human tears at about 150 µM (10) and in the alimentary tract ranging 250-300 µM (13). Human breast milk is relatively SCN- poor, with a median value of approximately 1 µM (14, 15).
Elimination of SCN- occurs in the kidneys at a half-life of 3 days in healthy individuals due to a 90% reuptake rate of SCN- from glomeruli filtrates (16). Mean relative volume of distribution in healthy subjects is 0.25 L kg-1. Renal
half-life to a week or more (16). SCN- is excreted in the urine of healthy individuals at a concentration <10 µM (14, 17).
Origin of endogenous SCN-
In healthy individuals, SCN- is thought to originate primarily from the diet either as an intact molecule or resulting from the metabolism of cyanide by
sulfurtransferases (e.g., rhodanese and mercaptopyruvate sulfurtransferase) (18). Rhodanese-like enzyme activity (which catalyzes the formation of SCN- and
sulfite from cyanide and thiosulfate (19)) is chiefly located in the kidney, liver, brain, lung, muscle and stomach of humans (20). Tissue distribution of
rhodanese activity varies widely from species to species and is in part a product of the intrinsic diet of a given species of animal (e.g., increased alimentary rhodanese expression in ruminants) (21).
Dietary intake of SCN- varies between ethnic and cultural groups and is associated with glucosidic cyanogen-rich plants, such as cassava, yam, maize, sugar cane, sorghum and linseed, as well as glucosinolate metabolism (17, 22). These cyanogenic compounds also yield N-conjugated thiocyanates, the
structurally related but biochemically distinctive isothiocyanates (e.g.,
sulforaphane) and nitriles (23). Because cyanogen metabolism produces a wide milieu of biomolecules, the effects of cyanogens on human health cannot be inferred as the direct effects of SCN-.
The average daily intake of SCN- from diets rich in cyanogens is far below the amount required to observe adverse events in cyanogen-exposed animals (23). In a study of vegetarian and vegan individuals (groups considered at risk of iodine deficiency) it was reported that median SCN- excretion in urine was higher in non-smoking vegans than in vegetarians (10.9 and 5.8 µM, respectively) (17). However, thyroid-stimulating hormone (TSH) and free thyroxine (FT4) were not significantly different between the two groups and did not correlate with SCN -even when multiple variables were adjusted (17). Similarly, goitrous individuals given milk supplemented with 0.1 mg/L iodine and 328 µM SCN- daily over four weeks maintained normal thyroxine, triiodothyronine and thyroid-stimulating hormone (24). A study of mothers and breastfed infants found that SCN- in breast milk, maternal urine or infant urine is not predictive of infant TSH or FT4 (14).
Smoking significantly increases plasma SCN- due to the body’s increased utilization of sulfurtransferases to detoxify the resulting cyanide (1, 12, 25-27). Non-smokers generally fall below 70 µM plasma SCN-, while smokers may range from 70 to 300 µM with most mean reported values ranging from 100-150 µM. Some drugs and supplements such as nitroprusside and cyanocobalamin may also contribute to SCN- concentration by releasing cyanide (28). Studies of patients on nitroprusside did not identify toxicity caused by SCN-, even though plasma concentrations in some patients approached 1 mM (16, 29).
As with cyanogen exposure from the diet that may increase [SCN-] in the body, the outcomes of smoking and other behaviors that result in heightened
cyanide exposure, and thus elevated concentrations of SCN-, cannot be inferred as the effects of SCN- on purely associative grounds. “Correlation does not imply causation.” In two postmortem studies, elevated plasma SCN- was used to distinguish smokers from non-smokers and correlated with atherogenic
biomarkers such as oxidized low density lipoprotein, apolipoprotein E deposition and macrophage foam cells (30, 31). However, plasma SCN- actually correlated with less mortality in human myocardial infarction survivors (32) and protected against atherosclerosis in rabbits, which received 20-60 mg kg-1 d-1 K+SCN- orally (33). Similarly, plasma SCN- in smokers and non-smokers was not predictive of atherosclerotic coronary artery disease risk (26). In another study increased plasma SCN- was implicated as a risk factor for excess thiol oxidation by manipulating the plasma to mimic inflammation, but smokers and non-smokers had similar concentrations of plasma thiol (27). Thus, unlike circulating MPO levels (32, 34), increased plasma SCN- does not seem to be predictive of cardiovascular disease and mortality but may actually be a protective factor.
Currently it is unknown whether SCN- may also be synthesized from an endogenous source, but the ubiquity of SCN- in biologic systems and its exceptional concentration in many extracellular fluids suggest this possibility. Some bacteria have been shown to generate cyanide from glycine (35), which would act as a metabolic precursor for endogenous SCN-, but it is unclear whether any similar pathways exist in eukaryotes or how cyanide or a similar precursor molecule would be compartmentalized and regulated in order to avoid
deleterious effects on metabolism.
The discovery that SCN- is apically transported by the cystic fibrosis transmembrane conductance regulator (CFTR) (7) led to observations of its deficiency and dysregulation in primary lung epithelia, mouse and pig ELF and human saliva (3, 5, 12, 36, 37). Furthermore, SCN- concentration in NLF correlated with greater lung function in CF patients (3). Thus, SCN- may be a useful therapeutic in CF, which is one subject of this thesis. Previous use of SCN -as an antihypertensive provides insight into human tolerance and toxicology of SCN- as a drug.
Early 20th century administration of SCN- as an antihypertensive
In 1903, Wolfgang Pauli introduced SCN as a therapeutic agent for the treatment of hypertension, but it was not investigated thoroughly as an
antihypertensive until a quick and reliable method to measure plasma levels of SCN- was introduced in the 1940s (38, 39). Values originally reported as mg/100 dL will be reported here as the converted molar value. The efficacy of SCN- as an antihypertensive agent varied in reports of different authors during this time, with both enthusiastic supporters and concerned skeptics. By the mid-20th century SCN- again fell out of favor with physicians to be replaced by better
anti-hypertensives.
Patients treated with SCN- were commonly observed with plasma concentrations of 400–2000 µM, with additional case reports of higher
concentrations up to 7 mM (40, 41). SCN- was given as a potassium salt between 400 and 1000 mg per day and therapy could last from weeks to years (38, 42). Plasma SCN- was regularly maintained between 1400 to 2000 µM to sustain therapy and avoid toxicity according to Barker and associates (41). Studies were mixed regarding the efficacy of SCN- in hypertension but the contemporary literature suggests that these exceptionally high dosing regimens may have produced modest beneficial effects (43). Cardiac pathology also correlated inversely with therapeutic efficacy of SCN- so that hypertensive subjects with the least chronic disease symptoms were the most likely to experience benefits (43).
Toxicologic effects of SCN-, such as thyroid dysfunction, dermatitis and in extreme cases psychosis, were noted with excessive or uncontrolled dosing (40, 42, 44). Toxicology reports included descriptions of the very high exposures of plasma SCN- the patients experienced, up to 7 mM or over 2 orders of magnitude higher than the normal human plasma concentration (44). Most toxicity reporting came from patients with sustained plasma SCN- of 1 mM or more. In many cases from the literature it is difficult to isolate the effects of SCN- from other medical conditions, such as nervous disorders or renal impairment, and other treatments that may have caused toxicity, such as bromide sedatives (40, 44). It is also worth mention that these reports come from the time period when salt iodization was relatively new (increasing the chance of individual vulnerability to
interference of SCN- with I- transport). However, the apparent 1 mM plasma SCN -threshold to observe adverse events in humans is several times higher than the
effective concentration to fully alter most known biochemical and redox
processes related to SCN- (detailed in the next section), suggesting it was not these pathways that contributed to the reported toxicity but a different effect caused by the extremely high and prolonged doses of SCN- that hypertensive patients were exposed to.
Biochemical reactions of SCN- and HOSCN
SCN- shares many chemical properties with monatomic halides (e.g., Cl-, Br-, I-) but has a 58 Da molecular mass, consisting of a sulfur atom sharing a sigma bond with a nitrile carbon which results from the enzymatic transfer of a thiolate from a sulfur donor to cyanide. SCN- is not to be confused with the related isothiocyanate, an unstable resonance structure of SCN- limited to alkyl-conjugated species in nature (e.g. sulforaphane, iberin) (22, 23). The pKa of the sulfur atom of SCN- has been reported in the sub-zero range so, like the halides, SCN- is anionized even in the most acidic physiologic fluids (45). SCN- is
relatively inert prior to oxidation, which is the major biochemical step necessary for its physiologically relevant effects.
Attention was first brought to the role of SCN- in host defense when it was discovered that its oxidation by mammalian haloperoxidases in the catalytic reduction in H2O2 yields antimicrobial effects (46). The product of this reaction was later identified as HOSCN, a close chemical relative of the hypohalous acids (HOX) (47, 48). Although milk was the first well-studied biological matrix shown to
utilize SCN- for microbicidal activity via the action of peroxidases (46), it is now known that saliva, ELF, NLF, gastric fluid and tears do so as well (2, 3, 10, 13, 49, 50). Thiocyanogen (SCN2) exists in equilibrium with HOSCN and SCN- but is difficult to isolate at physiologic pH due to rapid equilibrium and may not readily occur under physiologic conditions (51).
The identity and structure of HOSCN has since been confirmed as the major product of SCN- oxidation by nuclear magnetic resonance and mass spectrometry (52, 53). HOSCN is generated from SCN- at rates approaching the diffusion limit whether occurring enzymatically via a haloperoxidase or by the non-enzymatic scavenging of hypohalous acids (e.g. HOCl, HOBr) by SCN-. The initial step of both pathways is the reduction of H2O2 by haloperoxidases, which occurs at a rate of ≥107 M-1 s-1 (54). The second step in the enzymatic pathway is two-electron oxidation of SCN- to regenerate the heme. The second order rate constant for the oxidation of SCN- by MPO compound I is 9.6 x 106 M-1 s-1, while the same rates for EPO and LPO are 10-20-fold greater (54, 55). The direct oxidation of SCN- by HOX follows oxidation of X- by MPO, EPO or (rarely) LPO (54), and thus is a three-step HOSCN-generating process in parallel to the enzymatic pathway. The reaction of HOCl with SCN- has a second order rate constant of 2.3 x 107 M-1 s-1 (56) and SCN- reacts with HOBr two orders of
magnitude even more rapidly at 2.3 x 109 M-1 s-1 (57) which corresponds with the relative electrophilicity of these HOX species (51). Furthermore, SCN- is the preferred electron donor (analyzed by specificity constant, kcat Km-1) of compound
I of all haloperoxidases that have been tested, including lactoperoxidase (LPO) (46), eosinophil peroxidase (EPO) (58), MPO (55, 59), gastric peroxidase (GPO) (13) and the recently discovered vascular peroxidase (VPO) (60) (all reviewed in detail with the exception of GPO and VPO in (54)). The salivary peroxidase
(SPO) referred to in many publications is likely identical or nearly identical to LPO, encoded by the same gene (61). Certain haloperoxidases are also relatively efficient in using Cl- and Br-, such as MPO, while others, such as LPO, will only oxidize SCN- in large amounts in physiologic conditions (54, 59, 61). Although I -is another potential electron donor for compound I of haloperoxidases, it -is too scarce in vivo (100 nM in serum) to significantly contribute to halogenation reactions. An exception to this rule is the thyroid, in which I- is concentrated 20-40 fold to promote synthesis of T3 and T4 via the peroxidatic activity of the tissue-specific thyroid peroxidase (TPO) (8).
In either case of enzymatic or non-enzymatic generation, the reaction to form HOSCN is limited by the extent of H2O2 reduction by haloperoxidases
(regulated by NADPH oxidase (NOX) and dual oxidase (DUOX) expression (62)) and the concentration of SCN-. Interestingly, airway epithelial DUOX expression has been suggested to be another source of HOSCN, owing to the extracellular DUOX peroxidase domain (63). Given that H2O2 and haloperoxidases are available and SCN- is sufficiently concentrated, HOSCN will be the dominant oxidizing species in an inflammatory milieu (Figure 1.1). However, the
Figure 1.1. SCN- regulates the concentration of HOX in inflammatory milieu. Release of MPO from neutrophilic granules is concurrent with oxidative burst, driving the formation of HOX. SCN- is the preferred electron donor of MPO so that increasing concentration of SCN- increases the relative proporition of H2O2 converted to HOSCN instead of HOCl or HOBr.
Proportions are approximate. This scheme does not depict the 1-electron radicals generated by the peroxidation cycle of MPO, which account for <10% of H2O2 consumed by MPO. Proportions are based on computational and observational data from publications arising from this thesis and from other works: (27, 64, 152, 153).
HOSCN comprises only 5-50% of oxidants produced during inflammation (27, 64). Concentrations of SCN- ≥400 µM drive the totality of oxidant output by MPO to generate HOSCN in vitro (64).
The major physiologic target of HOSCN is cysteine thiol, which makes up an ubiquitous component of peptides and proteins with large variations in
abundance (e.g., GSH versus Trx (65)) and chemical properties (e.g., free cysteine versus a catalytic residue influenced by protein context (66)). HOSCN reacts with anionized thiolates (RS-) in its protonated form resulting in the
formation of a sulfenic acid (RSOH) and transient inactivation of affected catalytic residues and/or changes in cell signaling regulation (67). However, the pKa of HOSCN is 4.85 (68) so the unreactive conjugate base hypothiocyanite (OSCN-) predominates in most physiologic fluids and limits rates of reaction. Due to this relative scarcity of the protonated HOSCN molecule in physiologic conditions and its >2-fold decrease in reduction potential compared with HOX (69), HOSCN is a more discriminatory and less potent reaction partner with biologic thiols than HOX (e.g., 103-105 M-1 s-1 at pH 7.4 for HOSCN vs. >107 M-1 s-1 for HOX under similar conditions) (68, 70). Greater acidity of a thiol will increase that thiol’s rate of reaction with HOSCN. For example, the second order rate constants for the reaction of HOSCN with 5-thio-2-nitrobenzoic acid (TNB, pKa=4.38, 3.8-4.4 x 105 M-1 s-1) is 5-15 times faster than with glutathione (GSH, pKa=8.7, 2.5-8.0 x 104 M -1
s-1) at pH 7.4 (68, 71). This finding is in marked contrast to that of HOCl, which reacts equally quickly with thiols ranging from pKa=4-9 (72). Mildly acidic pH
increases the rate of reaction of HOSCN with thiols and significantly enhances its selectivity for those with low pKa (e.g., kTNB/kGSH=15.2 at pH 7.4, 52.2 at pH 6.7 and 193 at pH 6.0) (68, 71); this is another trait that distinguishes HOSCN from HOX in the physiologic pH range (73). Furthermore, HOSCN reacts 1-2 orders of magnitude more rapidly with the chemically related selenolates (RSe-; present in selenoproteins, or proteins that express selenocysteine (Sec) in their primary structure) than an otherwise identical thiolate (74). Selenolates have an advantage over thiolates in that their increased acidity (pKa≤6) does not negatively influence the enhanced nucleophilicity of the selenium atom, which has been suggested to be a trade-off in an otherwise identical sulfur thiol (66). The enhanced nucleophilicity and acidity of selenium compared to sulfur (75, 76) is a logical chemical rationale for enhanced rates of reaction of HOSCN with selenols (e.g., HOSCN reacts nearly 16-fold faster with Sec than normal cysteine and 68-fold faster with γ-Glu-Sec-Gly than γ-Glu-Cys-Gly) (74).
Reaction of a thiol with HOSCN is thought to produce an unstable sulfenyl thiocyanate (RSSCN) (though this step may require an SCN2 intermediate (77)) which rapidly cleaves to form sulfenic acid (RSOH) and disulfides (RSSR) that can be repaired through enzymatic mechanisms (77, 78). Alternatively, HOSCN may bypass the RSSCN intermediate and directly oxidize RSH to RSOH with SCN- as the immediate leaving group. Higher oxidation states of sulfur (e.g., sulfinic (RSO2H) and sulfonic (RSO3H) acids) that require more advanced repair mechanisms such as sulfiredoxin conjugation and subsequent reduction (79, 80)
have been identified in vitro by direct reaction of HOSCN with purified proteins that express one or more catalytic thiols (81). Selenols undergo similar redox chemistry, likely forming a selenenyl thiocyanate (RSeSCN) intermediate, then resolving to or directly forming selenenic acid (RSeOH), diselenide (RSeSeR) or a heterogeneous selenosulfide (RSeSR) before regeneration of the selenolate or further oxidation (82).
Selenoproteins are especially good targets for HOSCN in vivo due to the reasons listed above. Though present in all kingdoms of life, selenoproteins are most abundant in high order animals such as mammals (83). In mammals, thioredoxin reductase (TrxR) is expressed as a selenoprotein with high structural homology to glutathione reductase (GR), in contrast to bacteria, plants and most single celled eukaryotes (84). TrxR is distinguished from GR by a C-terminal redox motif that assists the main active site in catalysis and expresses a penultimate Sec residue (84). TrxR is the chief redox regulator of thioredoxin (Trx), which in turn regulates DNA synthesis, peroxiredoxin (Prx) and H2O2, gene transcription and cell viability among other functions (85). However, TrxR can also act directly to reduce other targets (e.g., lipoic acid, selenite) (86). Thus, HOSCN may react directly with selenoproteins such as TrxR resulting in catalytic reduction, particularly if Sec expression makes the reaction more favorable. This subject is explored in Chapter II of this thesis.
Other reactions have also been reported for HOSCN from in vitro experimentation, though many are much less favorable in vivo owing to
significantly slower rates of reaction and the abundance of more favorable thiol and selenol targets of HOSCN in biological systems. Aqueous HOSCN ca. 200 µM has a half-life of 2-3 hours at room temperature (47), undergoing accelerated decomposition at acidic pH (48). When lacking a reaction partner, aqueous
HOSCN will eventually disproportionate into 4 parts SCN-, 1 part sulfuric acid and 1 part cyanide (47, 87). Complicating this scheme, aqueous HOSCN reacts with cyanide to form SCN- and OCN- (47, 88). Oxidation of tryptophan by HOSCN in phosphate buffer has also been reported (89), but the concentration of HOSCN required to observe the effect in a plasma matrix (>1.25 mM) is far above the reported physiologic range of this oxidant (11) (and is also well above the concentration of precursor SCN- in plasma, including smokers (1, 27)). Indirect evidence suggests ascorbate may react with HOSCN or modulate its generation by MPO (89, 90). The apparent reaction of ascorbate in buffered saline may comport with the speculation of a radical species arising from 1-electron oxidation of SCN- or OSCN- by peroxidase compounds I and II in balanced salt solutions (91, 92), but the in vivo relevance of such a reaction is not clear considering the 1-electron oxidation of SCN- is much less thermodynamically favorable than its 2-electron oxidation (69, 93).
Effects of SCN- and HOSCN on toxicity and cell viability
Another important consideration regarding the generation of HOSCN from SCN- is what molecules HOSCN is replacing or displacing when SCN- is oxidized by haloperoxidase compound I or HOX (Figure 1.2). The favorable interaction of SCN- with compound I of MPO and EPO, which is approximately 3 orders of magnitude faster than Cl- and 1-3 orders of magnitude faster than Br- (58, 59), protects cells against exposure to the more reactive and highly oxidizing HOCl and HOBr (HOSCN, pKa=4.85, E’˚=+560 mV; HOCl, pKa=7.53, E’˚=+1280 mV; HOBr, pKa=8.8, E’˚=+1130 mV (68, 69)). SCN- also reacts directly as a
nucleophile with HOCl and HOBr and similarly can scavenge monohalamines (RNHX), a range of oxidizing metabolites generated by HOCl (56, 94). “Trading” HOSCN for HOX, most often HOCl, changes the relationship of the ultimate oxidant with the biological system. For example, HOCl is much less
discriminatory or pH-dependent in reactivity and at least 100-fold more rapid than HOSCN when oxidizing thiols (68, 71, 73, 95, 96) and unlike HOSCN will also target amino acids, nucleotides and poly-unsaturated fatty acids and can also produce hydroxyl radical and singlet oxygen (96-99). By replacing HOX with HOSCN and effectively decreasing the reactivity of the ultimate oxidant in this series of biochemical reactions, SCN- is functioning under the standard definition of an antioxidant as well as blocking a cascade of deleterious reactions.
Decreasing HOX inhibits the formation of RNHX, which contribute to the
Figure 1.2. Chemical and enzymatic pathways driving the host defense and antioxidant properties of SCN-. Haloperoxidases (e.g., MPO, LPO or EPO) utilize H2O2 generated by NADPH oxidases (NOX) and dual oxidases (DUOX) to form HOSCN or HOX. HOX (e.g., HOCl or HOBr) can react with amino acids to form haloamines (RNHX) and directly oxidize other cellular macromolecules resulting in cytotoxicity. SCN- can directly scavenge HOX and repair RNHX and in the process generates HOSCN, which maintains host defense properties of HOX but is better tolerated by host tissue. This figure was originally published in (174).
is potently inhibited by SCN-, although SCN- also participated in MPO-catalyzed LDL oxidation (90). SCN- inhibits urate radical formation by MPO from stimulated neutrophils in plasma, sparing ascorbate, GSH, and nitric oxide (NO･) (101). Similarly, SCN- blocks EPO-catalyzed nitration of Tyr and BSA (58) and can modulate NO･ binding to heme peroxidases and inhibit MPO and LPO-mediated NO･ consumption in physiologic fluids (102, 103). In addition, generating HOSCN is a mechanism used by mammalian cells to restrict the concentration of H2O2 in secretory fluids, demonstrated by the paucity of H2O2 in the oral cavity compared with HOSCN (11, 61), which may be preferable for cells owing to H2O2’s much greater redox potential and cell permeability compared to HOSCN (104).
Evidence for SCN--mediated protection from HOCl, HOBr and H2O2 in whole cells has been reported by many researchers. Accumulation of HOSCN in ex vivo neutrophils, eosinophils and macrophages was reported to be non-toxic (105). SCN- was reported to protect HL-60 cells from H2O2/MPO-mediated apoptosis in the presence of physiologic levels of Cl- by outcompeting Cl- for MPO compound I, in contrast to Br- which did not decrease apoptosis in the presence of Cl- (106). SCN- potently inhibited cell death of ex vivo rat aortic
endothelia exposed to EPO, H2O2 and Br- (107). Cell death associated with MPO, H2O2 and Cl- was inhibited with the addition of 100-400 µM SCN- in lung, nervous, pancreatic and endothelial cells and similarly SCN- protected cells from H2O2 through the activity of LPO (64). Addition of SCN- to endothelial cell cultures inhibited fibronectin oxidation by HOCl and loss of cellular adhesion (108). The
apparent lack of toxicity of SCN-/HOSCN during enzymatic generation may
implicate an evolutionary rationale for the substrate selectivity of haloperoxidases for SCN- over the true halides (vide supra).
However, researchers using acute exposures to bolus doses of HOSCN have reported cytotoxicity in J774A.1 murine peritoneal macrophage-like cells and human umbilical vein and coronary artery endothelial cells (HUVEC and HCAEC respectively) (109-111). These experiments do not agree on the
mechanism of cell death (e.g., apoptosis or necrosis, although caspase inhibition is noted in all studies) but do suggest HOSCN can injure mammalian cells by oxidizing one or more sensitive targets. However, lower doses of HOSCN actually enhanced HUVEC viability (110) and HCAEC generally fared much better when exposed to HOSCN compared with HOCl (111). J774A.1 cells were reported very sensitive to HOSCN-mediated apoptosis while showing relatively little response to HOCl (109), although multiple findings using J774A.1 cells presented in Chapter II of this thesis are in conflict with those results.
Interestingly, HOSCN has not been reported to induce cell injury in peroxidase-driven models without baseline H2O2 injury, which may be the more
physiologically relevant method compared with the bolus dosing approach. In vivo and clinical data from the literature and presented in this thesis (Chapters III-V) also strongly support the hypothesis that SCN- is protective against
Antimicrobial properties of HOSCN
HOSCN is major component of innate immune response to pathogens (Figure 1.3). HOSCN is 10-500-fold more effective than H2O2 in inhibiting the growth and metabolism of oral streptococci with a reported minimum effective concentration of 1-10 µM varying by bacterial strain (46, 50, 112). Antibacterial effects of HOSCN have been demonstrated against multiple gram-negative and gram-positive species (46, 50, 61, 113). The antibacterial effect of HOSCN has been attributed to its ability to cross the bacterial cell wall to oxidize critical
metabolic elements, and because of its predominantly ionized state in physiologic fluids, it may utilize porins to penetrate bacterial cells (114). The reported targets of HOSCN in Streptococcus strains are glycolytic enzymes including hexokinase, glucose-6-phosphate dehydrogenase and aldolase as well as inhibition of oxygen uptake (46). HOSCN blocks uptake and induces cellular leakage of glucose, amino acids and K+ in E. coli and S. lactis, likely by targeting membrane-bound transport proteins for these nutrients (113). Similarly, HOSCN oxidizes glucose uptake transporters preventing glycolytic metabolism in S. agalactiae (115). HOSCN inhibits Helicobacter pylori viability and urease activity, which is required by the bacteria to alkalinize gastric juice in order to colonize the stomach (116). In a complementary mechanism of action, HOSCN also enhances innate immune response by inducing the expression of endothelial cellular adhesion molecules (CAMs) regulated by NF-κB in a mechanism of inflammatory amplification at sites of phagocytic activity (117).
Figure 1.3. Transport of SCN- by secretory epithelium and its protective role in the lumen during infection and inflammation. SCN- enters secretory epithelia from the blood stream by sodium-iodide symporter (NIS, double-triangle) and is exported into the lumen by cystic fibrosis transmembrane conductance regulator (CFTR, triangle) and/or calcium-dependent chloride channel (CaCC) and pendrin (square). Peroxidases arrive in the lumen from both epithelia, which secrete lactoperoxidase (LPO), and inflammatory leukocytes such as neutrophils, which release myeloperoxidase (MPO) upon activation. Epithelia and inflammatory leukocytes also provide H2O2 to drive peroxidase activity via dual oxidase (DUOX) and NAD(P)H oxidase (NOX) depicted as circles. The peroxidases and SCN- generate the thiol-selective oxidant HOSCN or in low concentrations of SCN- the more reactive HOCl is generated by MPO. SCN- directly reduces HOCl to form HOSCN and spares potential toxicity resulting from HOCl while maintaining peroxidase-mediated host defense. This figure was originally published in (174).
Many reports of the antibacterial action of the SCN--peroxidase-H2O2 system in vivo have come from the oral cavity. The oral cavity is heavily concentrated with 670-2500 µM SCN- (10-12, 118), is in a constant state of
inflammation (e.g., H2O2 production) and its mildly acidic conditions optimize LPO activity and HOSCN reactivity (119). Mean HOSCN concentration in human saliva has been reported from 58-65 µM (11, 120), which is greater than the observed 10 µM tolerance of the oral cavity to H2O2 (11). HOSCN from saliva has been reported to inhibit acid production by glucose-stimulated plaque (121) and reduce growth of both aerobic and anaerobic periodontopathic bacteria (122) (123, 124).
Multiple reports exist of nicotinamide adenine dinucleotide
(NADH)-dependent HOSCN resistance in oral and lactic streptococci which is associated with a purifiable inhibition “reversal factor” (46, 125). This adaptation allows the bacteria to resist oxidative stress by HOSCN, which provides a selective
advantage in the SCN--rich oral environment. This may also be the reason for the increased resistance to HOSCN-mediated cellular dysfunction exhibited by
certain strains of lactic streptococci compared to gram-negative pathogens, although differences in the cell wall have also been posited as the reason (113). Similar adaptations to safely metabolize HOSCN by human secretory epithelia are among the subjects of this thesis.
In addition to bacteria, HOSCN has been shown to inhibit the growth of fungi and viruses. Candida albicans cell viability is inhibited by enzymatic
exposure to HOSCN and ablated with acute exposure (126, 127). HOSCN blocked viral infection of human gingival fibroblasts by herpes simplex virus, respiratory syncytial virus, and echovirus with enhanced efficacy at mildly acidic pH as can be found in the oral cavity (128). Similarly, enzymatically-generated HOSCN rapidly ablated HIV replication competency in lymphocytic culture (129). HOSCN may be as potent as HOBr and HOCl in killing parasitic schistomula (52).
Interest in the role of HOSCN on microbicidal activity in the airway has grown since the importance of LPO in airway bacterial clearance was first reported (6). SCN- and LPO were observed to be concentrated and active in airway secretions enough to inhibit airway-localizing bacteria (2). Airway epithelial-localized DUOX2 was later identified as the probable source of H2O2 needed for LPO activity under the regulation of infection-mediated stimuli (130), albeit having its own heme peroxidase activity when expressed on the epithelial surface (63). The importance of this antibacterial activity was later underscored by the discovery that CFTR is a major transporter of SCN-, which may account for the poor bacterial clearance observed in CF subjects (7). This is covered in more detail in the next section.
Cystic Fibrosis
Brief history of CF and current interventions
Cystic fibrosis (CF) was first recognized as a distinct disease of the exocrine ducts in 1938, and in 1989 the causative inheritable mutation in the CF transmembrane conductance regulator (cftr) gene, leading to impaired or absent CFTR protein at the apical surface of cells, was reported (131). CFTR is now known to be a cAMP-regulated transporter of chloride (132), GSH (133, 134), SCN- (7) and bicarbonate (135). CFTR may arise out of one or more faulty copies of the cftr gene, resulting in mutations that impede function (e.g., G551D) or block proper protein maturation (e.g., ΔF508) (136). Treatment has steadily improved in focus and efficacy over the nearly 80 years since the initial
identification of the disease so that median patient life expectancy has improved from a few years of childhood to the mid-30s (131). For patients with G551D mutation (~2% of CF patients), access to the orally bioavailable drug ivacaftor, the first successfully developed CFTR modulator, has greatly improved prognosis (137). Unfortunately, this drug does not correct failed protein maturation in
patients that carry the ΔF508 allele, a much more common and severe form of the disease. For this reason ivacaftor underscores both the potential
breakthroughs and the shortcomings of personalized (i.e., genomically-targeted) medicine (137).
Multiple species of bacteria colonize CF airways that are not readily removed by the impaired host defense system and require antibiotic intervention (138). Infection with Pseudomonas aeruginosa is the most prevalent bacterial airway infection and correlates with lung function decline. P. aeruginosa creates a biofilm in the airway, making it particularly difficult to eradicate (139). In addition, CF patients also suffer from sustained airway inflammation characterized by chronic airway neutrophilia (140). Neutrophils rapidly activate and undergo cell death after entering the airway, contributing to an excess of purulent phlegm and leaving behind a number of damaging enzymes and pro-inflammatory factors (141). The combination of chronic infection and sustained inflammatory response results in progressive destruction of the airway. It is estimated that ultimately 80– 95% of CF patients’ morbidity and mortality results from these processes (138).
A number of lung diseases are treated with therapeutic nebulization of saline and/or drug excipients including CF (142), asthma (143), chronic
obstructive pulmonary disease (COPD) (144) and pneumonia (145). Hypertonic saline (HS) is notable because it improves lung function and decreases
exacerbations in CF patients (146). Furthermore, HS increases thiol antioxidants in the airway epithelial lining fluid (ELF), including SCN-, which, in addition to airway surface hydration, may contribute to its beneficial effects (5). However, HS creates an osmotic stress for airway epithelia (142) and it is possible that years of daily inhalation of a hypertonic solution will have unanticipated consequences.
Molecular implications of CF and biological models
The understanding that SCN- functions in the basic host defense of the airway (2, 6) gained new importance when it was demonstrated that the CFTR regulates the efflux of SCN- into the airway environment (7). Later studies demonstrated the importance of CFTR in SCN--mediated host defense in vitro using primary cultures, suggesting a powerful host defense mechanism was missing in CF (36, 37). However, the discovery that pendrin and CaCC also transport SCN- to the airway ELF under the direction of inducible cytokines (9) and reports of SCN- concentrations similar to control values in CF patients in ELF and NLF (3, 4) has since called this hypothesis into question. Nevertheless, the correlation of improved lung function with SCN- and results from IV administration of SCN- in newborn CF pigs (3) support the idea of potential host defense and antioxidant deficiencies in CF related to dysregulation of SCN-. It is possible that deficiency or dysregulation of SCN- exists in CF that may later be
counterbalanced by disease progression and the changed expression of ion transport proteins (e.g. CaCC and pendrin) (9). In spite of this evidence, SCN -has not been advanced as a CF therapy in humans. Because nebulized HS can result in increased ELF SCN- and has positive effects on lung function in CF (5), there is additional rationale to develop an SCN--delivering therapy for CF. SCN -may have beneficial effects in related illnesses of the airway, particularly involving infection and inflammation.
Developing viable models of CF in non-human species in order to characterize disease mechanisms and test experimental therapies has proven challenging owing to differential vulnerability to the range of consequences of a dysfunctional or absent CFTR protein (147). The CFTR knockout mouse is problematic due to exacerbated complications in the alimentary tract and
relatively little or no spontaneous lung disease compared to human patients (148). Following the discovery that an overactive epithelial sodium channel (ENaC) may play as important a role in CF lung disease as dysfunctional CFTR (136), a βENaC-overexpressing mouse model was developed which in many respects better mimics the airway pathology of CF, including decreased periciliary layer and mucus clearance, goblet cell metaplasia, chronic neutrophilia and pulmonary mortality (149, 150). This model targets the airways using the Clara cell secretory promoter (CCSP) to avoid unwanted effects in other organ systems (150).
βENaC mice also breed well and hemizygosity for the transgene is sufficient to produce the Na+-resorption phenotype, facilitating experimentation (150).
Overview
SCN- is an important molecule in the regulation of inflammatory outcomes of mammalian biology that is produced from the detoxification of cyanide and is ubiquitously found in mammalian physiological fluids. In inflammatory milieu, SCN- is oxidized to HOSCN at rates approaching the diffusion limit by enzymatic
and chemical mechanisms that are dependent on H2O2 reduction by
haloperoxidases. SCN- is the limiting factor in HOSCN formation in inflammatory milieu and displaces HOCl, HOBr and other oxidizing and radical species. Decreasing the reactivity of the final oxidant effectively makes SCN- an antioxidant in this system. HOSCN is nevertheless a reactive species that
oxidizes thiols and selenols and inhibits bacteria, viruses and fungi. The reported tolerance of mammalian cells for HOSCN is poorly understood and has come under scrutiny. In Chapter II of this thesis, the mammalian cellular tolerance for HOSCN is explored and mechanistically described. Further, SCN- is deficient in CF and may be an effective pan-CF therapy that may also be beneficial in other infectious and inflammatory lung diseases. In Chapters III, IV and V of this thesis the results of different in vivo and ex vivo studies to use SCN- as a therapy for lung infection and lung inflammation are presented.
CHAPTER II
MAMMALIAN THIOREDOXIN REDUCTASE CATALYTICALLY REDUCES HYPOTHIOCYANOUS ACID PROMOTING LUNG INNATE IMMUNITY AND
CELL VIABILITY2
Introduction
Multiple studies have shown the protective effect of SCN- against oxidative injury in mammalian cells (64, 105, 106, 108), which is attributed to the reaction of SCN- with compound I of haloperoxidases or the resulting HOX to produce HOSCN as an alternative to more injurious species (48, 56). However, more recent work has called this hypothesis into question, demonstrating changes in cell signaling and cell death following exposure to bolus doses of HOSCN (109-111).
An explanation for this discrepancy in the literature has proven elusive. However, the observation that some bacteria exhibit an NADH-dependent resistance to HOSCN (46, 125) begged the question whether similar
oxidoreductase defense mechanisms were operative in mammals. This would make particular sense considering that, unlike bacteria, mammals regularly
2 Several of the figures and some text in this chapter was originally published in Chandler et al (2013) Selective metabolism of hypothiocyanous acid by
mammalian thioredoxin reductase promotes lung innate immunity and antioxidant defense. J. Biol. Chem. 288 18421-18428; and Figure 2.1 was originally
published in Chandler et al (2013) Nebulized thiocyanate improves lung infection outcomes in mice. Br. J. Pharmacol. 169, 1166-1177.
generate freely soluble HOSCN (11, 120); hence, mammalian cells have probably evolved systems to cope with this oxidant in order to facilitate and survive the process of host defense against invading pathogens. The enhanced reactivity of HOSCN with selenols (74) suggests a selenoprotein in mammals could be responsible for conferring the mammalian tolerance against HOSCN observed in several studies. Further, like any cellular defense system, it could probably be overwhelmed or vary from cell type to cell type leading to the aforementioned discrepancies in the literature, in addition to differences in experimental method.
Thioredoxin reductase (TrxR) is genetically divergent between bacteria and multicellular eukaryotes such as mammals and will be distinguished in this chapter by the prefixes “H-“ and “L-“ referring to heavy (110 kDa) and light (70 kDa) molecular weights (84). H-TrxR is a selenoprotein in mammals responsible for maintaining critical cellular functions mostly through the regulation of the redox state of the low molecular weight dithiol, thioredoxin (Trx), which in turn maintains other cellular functions (85). However, H-TrxR also acts directly to reduce oxidized targets (86, 151). H-TrxR shares high primary, tertiary and quaternary homology with GR but has uniquely evolved an additional redox-active site in its C terminus that in mammals is expressed with a penultimate Sec residue (84). Sec has been proposed to broaden the substrate reactivity of
mammalian H-TrxR and help it resist oxidative inactivation (76). Expression of Sec in mammalian H-TrxR suggests the enzyme may be a sensitive target for
reaction with HOSCN (74). There are two possible outcomes of such a reaction: either the oxidized H-TrxR active site will catalytically regenerate through the donation of electrons by nicotinamide adenine dinucleotide phosphate (NADPH) to its flavin adenine dinucleotide (FAD) moiety (151) or the oxidation of redox residues by HOSCN will be terminal and inhibit H-TrxR’s normal catalytic function. Indeed, one team of researchers reported that HOSCN inhibits H-TrxR activity (74). However, their experimental methods did not include NADPH, the enzyme’s cofactor, which could have created an artifact in the study.
H-TrxR is strikingly different from the low molecular weight L-TrxR found in bacteria, plants and single-celled eukaryotes, sharing only 20% sequence identity (84). Although L-TrxR functions similarly to mammalian H-TrxR to reduce Trx, it differs in many of its other features including absence of Sec expression and a differently structured and located active site. Thus the relationship of L-TrxR with small molecular oxidizing species is likely to differ in one or more ways compared with H-TrxR.
The studies in this chapter were designed with two related goals: First, to assess the differential tolerance of mammalian cells to HOSCN and HOCl in 16HBE and J774A.1 cells: 16HBE serving as a model of normal human
bronchiolar epithelium and J774A.1 cells being of interest as the cell line used in a report describing significantly deleterious effects of HOSCN (109). The cells were exposed to a glucose oxidase (GOX)-LPO/MPO-coupled system to
set as the rate-limiting step and the peroxidases in excess. Both 16HBE and J774A.1 cells were protected from HOCl- (and to a lesser extent, H2O2)-mediated viability loss by HOSCN-generating conditions (i.e., inclusion of SCN-). A lactate dehydrogenase (LDH) release assay was used to measure cell viability, but J774A.1 cells were also assessed using flow cytometry to measure apoptosis and necrosis in an acute exposure model, which directly confirmed the protective result of SCN- against HOCl-mediated cell death.
Second, the role of H- and L-TrxR in mammalian and bacterial sensitivity or tolerance to HOSCN was assessed. H-TrxR function from 16HBE, J774A.1 and RAW cell lysates correlated with their relative tolerance to a high dose of HOSCN but not HOCl. Mammalian and bacterial H-/L-TrxRs were assessed for catalytic reduction of HOSCN and/or inhibition using a variety of in vitro methods. Sec-expressing mammalian H-TrxR rapidly turned over HOSCN with
physiologically relevant Km, which is the first time this metabolic function has been demonstrated. Replacement of the Sec residue with Cys and deletion of the eight final C-terminal amino acids dramatically decreased enzyme activity. In contrast, recombinant Escherichia coli L-TrxR lacked activity and was potently inhibited by HOSCN exposure. HOCl was also assessed and determined to be a poor candidate for catalytic reduction by TrxR due to competing off-target effects (e.g., rapid direct oxidation of NADPH). Lysates of 16HBE, E. coli, and P.
aeruginosa were assayed for HOSCN reductase activity, and rapid turnover of
P. aeruginosa from cystic fibrosis (CF) patients had no detectable or in some
cases very low HOSCN reductase activity an order of magnitude below that of 16HBE cells. In addition, HOSCN exposure was selectively toxic to bacteria but well tolerated by 16HBE cells. Inhibition of H-TrxR with auranofin (AFN) reduced HOSCN metabolism in 16HBE lysates and increased HOSCN-mediated toxicity, coinciding with increased intracellular thiol oxidation. These data suggest that HOSCN and H-TrxR constitute an important mechanism of host defense in mammals that inhibits pathogens while limiting host tissue injury. To the best of my knowledge and the knowledge of other study authors this was the first
published report of enzymatic (pseudo)hypohalous acid metabolism in mammals (152).
The majority of data from this chapter also appear (in some cases with modifications) in the following references: (152, 153). Specific incidences of this will be noted in the figure legends.
Methods
Mammalian cell culture
J774A.1 murine macrophage cells, 16HBE human bronchial epithelial cells and RAW 264.7 murine myelocytic cells were cultured in DMEM containing 1 g L -1
D-glucose supplemented with 10% FBS and penicillin-streptomycin. For studies involving TrxR activity from Figs. 2.4 to 2.6, 16HBE were maintained in DMEM
additionally supplemented with 100 nM L-methylselenocysteine (Sigma). Cells were plated at a density 2 x 105 cells well-1 on 24-well tissue culture plates (Corning) and allowed to adhere overnight. Viability was determined from LDH release into the medium 24 h after the beginning of exposure as described previously (154). Viability is based on percentage of LDH release according to the equation, [(100-% Release [Sample]) / (100-% Release [Mean Control Value])] * 100%.
Flow cytometry for quantification of apoptosis and necrosis
J774A.1 cells were plated in 6-well plates at a density of 2 x 105 cells mL-1 in DMEM with 10% FBS and used for experiments the following day. Cells were washed once with pre-warmed PBS supplemented with 1.26 mM CaCl2 and 812 µM MgCl2, then treated with the indicated concentrations of HOCl and/or SCN- in Ca2+/Mg2+-supplemented PBS for 15 minutes at 37 ˚C. Afterward, treatment buffer was replaced with media for 6 hours and cells were harvested by gentle scraping. Typical yield per well was approximately 106 cells. Annexin V and propidium iodide (PI) staining was used to assess cell death according to manufacturer instructions (BD Pharmingen). Cells were pelleted and washed once with cold PBS, then re-suspended in 1 mL Annexin V binding buffer (10 mM HEPES-NaOH, pH 7.4, with 140 mM NaCl and 2.5 mM CaCl2). An aliquot of 105 cells was mixed with Annexin V and PI for 15 minutes, then diluted to a final volume of 300 µL with binding buffer and kept at 4 ˚C until analysis using a FACSCalibur flow cytometer (BD Biosciences).
Preparation of cell lysates
Cells were pelleted, resuspended in PBS, sonicated for 5 s, and placed at -20 ˚C overnight. The solution was thawed, sonicated for 5 s, vortexed, and centrifuged at 2,000 g to pellet debris. The supernatant was centrifuged at 14,000
g in a 10 kDa cutoff filter. Protein was measured by Coomassie Blue stain, and
the preparation was used immediately.
TrxR assay
75 nM purified H-TrxR or L-TrxR or 500 µg ml-1 lysate protein was added to 100 µM NADPH in 100 mM potassium-phosphate buffer pH 7.5 with 1 mM EDTA at room temperature. Reaction was initiated with the addition of 20 µM oxidized E. coli Trx (Cayman-IMCO) and 1 mM DTNB and followed based on change in A412 (ε=14,150 M-1 cm-1) (155). If HOSCN had been added to the system, 50 µM reduced glutathione (GSH) was added to the DTNB solution beforehand (forming equimolar NTB) to quench any remaining HOSCN. If this step is not performed, the leftover HOSCN will oxidize NTB (68) as it is
generated and will give a false positive of enzyme inhibition.
Sources of purified TrxR and GR
Rat recombinant cytosolic H-TrxR1 was purchased from Cayman-IMCO. Wild type mitochondrial TrxR2, mutant Sec489Cys TrxR2, and mutant H-TrxR2 lacking the eight ultimate C-terminal peptides from mouse were produced
