N-Acetyl-Lysyltyrosylcysteine Amide, a Novel Systems Pharmacology Agent, Reduces Bronchopulmonary Dysplasia in Hyperoxic Neonatal Rat Pups

Abstract

Bronchopulmonary dysplasia (BPD) is caused primarily by oxidative stress and inflammation. To induce BPD, neonatal rat pups were raised in hyperoxic (>90% O₂) environments from day one (P1) until day ten (P10) and treated with N-acetyl-lysyltyrosylcysteine amide (KYC). In vivo studies showed that KYC improved lung complexity, reduced myeloperoxidase (MPO) positive (+) myeloid cell counts, MPO protein, chlorotyrosine formation, increased endothelial cell CD31 expression, decreased 8-OH-dG and Cox-1/Cox-2, HMGB1, RAGE, TLR4, increased weight gain and improved survival in hyperoxic pups. EPR studies confirmed that MPO reaction mixtures oxidized KYC to a KYC thiyl radical. Adding recombinant HMGB1 to the MPO reaction mixture containing KYC resulted in KYC thiylation of HMGB1. In rat lung microvascular endothelial cell (RLMVEC) cultures, KYC thiylation of RLMVEC proteins was increased the most in RLMVEC cultures treated with MPO+H₂O₂, followed by H₂O₂, and then KYC alone. KYC treatment of hyperoxic pups decreased total HMGB1 in lung lysates, increased KYC thiylation of HMGB1, terminal HMGB1 thiol oxidation, decreased HMGB1 association with TLR4 and RAGE, and shifted HMGB1 in lung lysates from a non-acetylated to a lysyl-acetylated isoform, suggesting that KYC reduced lung cell death and that recruited immune cells had become the primary source of HMGB1 released into the hyperoxic lungs. MPO-dependent and independent KYC-thiylation of Keap1 were both increased in RLMVEC cultures. Treating hyperoxic pups with KYC increased KYC thiylation and S-glutathionylation of Keap1, and Nrf2 activation. These data suggest that KYC is a novel system pharmacological agent that exploits MPO to inhibit toxic oxidant production and is oxidized into a thiyl radical that inactivates HMGB1, activates Nrf2, and increases antioxidant enzyme expression to improve lung complexity and reduce BPD in hyperoxic rat pups.

Graphical Abstract

Introduction

Bronchopulmonary dysplasia (BPD) affects more than 10,000 infants annually [1], making it the most common pulmonary morbidity of premature infants in the United States [2]. BPD is caused by complications from respiratory distress syndrome (RDS) of the newborn. Current treatments for RDS include surfactant, caffeine, supplemental oxygen, and gentle mechanical ventilation [2]. Although supplemental oxygen is essential for preventing hypoxemic organ injury in premature neonates, supplemental oxygen always increases oxidative stress and inflammation. One way hyperoxia increases inflammation is by activating resident myeloid cells and vascular endothelial cell NADPH oxidase (NOX) dependent superoxide anion production [3–5]. Activated resident myeloid cells release MPO and also generate superoxide anion, which dismutates into hydrogen peroxide (H₂O₂) thereby providing the substrate for MPO to oxidize chloride or nitrite anions into hypochlorous acid (HOCl) [6] or nitrogen dioxide radical (NO₂^•) [7], respectively.

A consequence of chronic increases in oxidative stress and inflammation in the lung is pulmonary cell injury and cell death. Dead and dying cells release high mobility group box1 (HMGB1). Under normal conditions, HMGB1 serves as a nuclear DNA binding protein that performs many vital roles in maintaining DNA structure and regulating gene expression [8]. However, after cells die or immune cells and platelets become activated, HMGB1 is released into extracellular spaces, where it turns into a danger-associated molecular pattern (DAMP) molecule that possesses potent cytokine- and chemokine-like properties [9, 10]. HMGB1’s ability to recruit myeloid cells to the lung begins to explain how hyperoxia and excessive mechanical ventilation increase myeloid cell recruitment and MPO-dependent oxidative damage to the premature neonate’s lung to induce a form of lung injury [11] that is far greater than that which is initially induced by hyperoxia. Support for HMGB1 playing a causal role in inflammation in BPD comes from clinical studies showing that the concentration of HMGB1 in tracheal aspirates directly correlates with BPD severity and death in ventilated premature neonates [12].

One of the reasons premature neonates are at increased risk of BPD is that they often lack the antioxidant enzymes that protect against the oxidative stress induced by supplemental oxygen [13, 14]. Exactly why premature neonates lack antioxidant enzymes is unclear. A potential explanation however is that hypoxic in utero environments do not require a robust antioxidant system and some premature neonates do not respond to hyperoxia appropriately, possibly because of genetic predisposition [14]. Reports by Cho and Kleeberger [15–17] have defined the relationship between Keap1 and Nrf2, linking hyperoxic lung injury to Nrf2 [18], and confirmed that Nrf2 is the primary regulatory pathway for modulating antioxidant enzyme expression in the lung [19]. Although developing antioxidant agents that scavenge multiple oxidants is attractive [14], effective targeting of oxidative stress to reduce lung injury in neonates has not been achieved and optimizing scavengers to target specific oxidants and free radicals is extremely difficult. As hyperoxia has been linked to Nrf2 [18], and premature neonates who are at risk of hyperoxic lung injury lack the antioxidant enzymes required for protection [13, 14], it should be possible to exploit the Keap1/Nrf2 signaling pathway to protect premature neonates against supplemental oxygen.

The above reports, taken together, suggest that BPD is a complex, multifactorial disease process that is induced primarily by oxidative stress and inflammation. Such a statement implies that a systems biology approach should be useful for treating BPD. However, to our knowledge, no such agent exists. If a therapeutic agent could be developed to inhibit MPO, inactivate HMGB1, or active Nrf2, then such an agent might be useful for decreasing oxidative stress and inflammation in the lungs of neonates such that lung development will proceed even in high-risk premature neonates treated with supplemental oxygen.

In the present study, we report on the effects of KYC on BPD in hyperoxic neonatal rat pups. KYC is a tripeptide that was initially designed to inhibit MPO toxic oxidant (HOCl) production and increase MPO-dependent catalase activity [20, 21]. Our studies suggest that KYC exploits MPO peroxidase activity to be oxidized into a KYC thiyl radical that inactivates HMGB1 and activates Nrf2. When hyperoxic neonatal rat pups are treated with KYC, oxidative stress and inflammation decrease, and antioxidant enzymes increase. Studies here show that KYC prevents oxidative lung injury and improves lung development in neonatal rat pups raised in hyperoxic environments.

Material and Methods

Peptide Synthesis:

KYC and biotin-aminohexanoyl-N-[Ahx]-KYC-amide or biotinylated-KYC (B-KYC) were synthesized using Fmoc [N-(9-fluorenyl)methoxy-carbonyl] chemistry, prepared and purified as an acetate salt by Biomatik USA, LLC (Wilmington, DE) as previously described [22–24]. Trifluoroacetic acid (TFA) in the tripeptide preparations was removed by dissolving KYC in distilled water containing 6 mM HCl followed by lyophilization (2–3X). TFA in KYC and B-KYC was quantified by NMR in the Biomolecular NMR Facility at MCW [25]. TFA was reduced to < 0.01% before being used for experiments.

Rats and Experimental Protocols:

Time-dated pregnant Sprague-Dawley rats were obtained from Envigo (Madison, WI) and acclimated in the animal facility for one week. Animal protocols were approved by MCW’s Institutional Animal Care and Use Committee and conformed to NIH Guide for the Care and Use of Laboratory Animals. Animals were housed in barrier cages with a 12-h dark-light cycle and were given free access to chow and water. Pups from two dams were mixed and randomly distributed to each nursing dam by balancing sex and size of the pups. The sex of the pups was determined from the relative distance between the genitalia and anus. The dam and pups were placed in cages in either room air (normoxia) or in >90% oxygen (hyperoxia) in an enclosed chamber from postnatal day 1 to day 10 (P1-P10) to induce BPD as previously reported [26, 27]. Oxygen concentrations were continuously monitored with an oxygen sensor (Reming Bioinstruments Co., Redfield, NY).

Since the number of neonatal rat pups per pregnant dam limits the number of experimental conditions that can be compared, we modified our standard experimental protocol. The first study was designed to determine if hyperoxia increased BPD and caused a change in proteins of interest. As numerous studies have shown that hyperoxia increases experimental BPD in neonatal rats and mice, these findings can be found in on-line supplemental data. The second study was designed to determine if inhibiting MPO toxic oxidant production reduced BPD. At least three sets of neonatal pups from three different dams were used for each study. Pups were fed ad libitum from nursing dams. Pups were inspected daily starting on P1 and weighed starting on P3 in room air for less than ten minutes. Nursing dams were switched daily to avoid the impact of oxygen toxicity. Experience has taught us that severity of hyperoxia-induced BPD is litter dependent. To minimize litter differences in both types of experiments, pups from different litters were mixed and randomly reallocated to the nursing dams.

Previously we reported that KYC was highly effective at reducing MPO-dependent oxidative injury to cultured endothelial cells and that KYC had little to no effect on endothelial cell viability, apoptosis, necrosis, or mitochondrial function when added to culture media up to 4000 μM [20]. In other studies, we observed that injection of increasing single doses (IP) of KYC into control mice did not induce observable adverse effects until the dose reached 800 mg/kg (unpublished observations). KYC has been used in a variety of animal models and has been shown to reduce chemically-induced tumor formation [28], vasculopathy in sickle cell mice [22], secondary brain injury in middle cerebral artery occlusion mice [23, 29], neurological disease scores in EAE [24, 30], and improve vasculogenesis and reduce neutrophil infiltration in hindlimb ischemia in diabetic mice [31]. With such an abundance of murine models of vascular disease and forms of injury showing that KYC effectively reduced myeloid cell recruitment and oxidative damage, with little to no evidence of toxicity, we focused our efforts on determining if KYC reduced BPD in hyperoxic neonatal rat pups.

No KYC dose-response studies were performed in neonatal rat pups in preparation for the current study. Instead, we decided to treat neonatal rat pups empirically with KYC at 5 mg/kg twice per day based on published and unpublished data in adult mice. A review of doses in previous studies in adult mice suggests that the therapeutic window for KYC is from 0.3 mg/kg (once per day) [24] and up to 15 mg/kg (twice per day) [30]. The higher dose is slightly more than 26 times less than the concentration of KYC that was observed to induce adverse effects. The dose of KYC used in the current study is based on our experience in treating chronic inflammation in other disease states [22, 24, 30] and is within the range that we consider appropriate for neonatal rat pups and adult rats.

To determine the effects of KYC on BPD in hyperoxic neonatal rat pups we randomly assigned pups to the PBS injection control group, or the KYC treatment group. KYC was injected (intraperitoneally, IP), starting at P2 to half of the randomly assigned pups in each litter using a sterile insulin syringe fitted with a 30G needle (Beckon Dickinson, New York, NY) until P10. To determine the effects of KYC on KYC thiylation of HMGB1 and Keap-1, pups were injected IP twice daily with KYC at P2-P9, and then once with B-KYC (5 mg/kg), or an equal volume of phosphate buffer solution (PBS). Pups were euthanized at P10 with carbon dioxide and lungs removed en bloc. A small cut was created on the left atrium and ice-cold normal saline was gently infused through the right ventricle to flush blood from the lungs before inflation for histology or snap-frozen in liquid nitrogen for protein studies. For weight-gain and survival studies, comparisons were made between the effects of normoxia and hyperoxia and between the effects of KYC and PBS on hyperoxic pups. Weight was recorded starting on P3 and death was recorded each day from P1 to P10.

Cells, Materials, and Antibodies:

Rat lung microvascular endothelial cells (RLMVEC) from neonatal Sprague-Dawley rats (RN-6011) and rat endothelial growth medium (EGM, M1266SF) were from Cell Biologics (Chicago, IL). Amicon®Ultra-4 centrifugal concentrators with 3K cut-off (UFC800324) were from Merck Millipore Ltd (Carrigtwohill, Ireland). Pierce™ Streptavidin (#88816), Protein G (#88848), Protein A Magnetic Beads (#88845), Pierce™ ECL Western Blotting Substrate (#32106), and SuperSignal™ West Femto Maximum Sensitivity Substrate (#34095) were from Thermo Fisher Scientific (Waltham, MA). Rabbit antibodies for myeloperoxidase heavy chain (MPO, sc-33596) and TLR4 (sc-10741) were from Santa Cruz Biotechnology (Dallas, Texas). Rabbit antibody for Cl-Tyr (HP5002) was from Hycult Biotech (Plymouth Meeting, PA). The chicken antibody for HMGB1 (326052233) was from SHINO-TEST Corporation (Kanagawa, Japan). Rabbit antibody for RAGE (GTX23611) was from GeneTex (Irvine, CA). Mouse antibodies for PECAM-1/CD31 (ab24590), for 8-OH-dG (ab48508), and for rat endothelial cell antigen-1 (RECA-1, ab9774) were from Abcam (Cambridge, MA). Rabbit antibody for Cox-1 (sc-7950) and Cox-2 (sc-7951) were from Santa Cruz Biotechnology. Mouse (ab190377) and rabbit (ab77302b) antibodies for HMGB1, and streptavidin-HRP (ab7403) were from Abcam (Cambridge, MA). Rabbit antibody for acetylated-HMGB1 (OASG03545) was from Aviva System Biology (San Diego, CA). Rabbit antibodies for Nrf2 (sc-722) and TLR4 (sc-30002) were from Santa Cruz Biotechnology (Dallas, TX). Rabbit antibody against Keap1 (ABS97) and mouse antibody for β-actin (A2228) were from Millipore-Sigma (St. Louis, MO). Rabbit antibody for cysteine sulfonyl (ADI-OSA-820) was from Enzo Life Sciences (Farmingdale, NY). Mouse antibody for glutathionylated proteins (D8) and all other chemicals were from Sigma-Aldrich (St. Louis, MO). All antibodies were certified either by the manufacturer or were used previously and certified by the authors, as was the case for the anti-HMGB1 antibody [32].

Immunohistochemistry and Immunocytochemistry:

After euthanasia, the trachea was cannulated with an Instech Solomon (20G) stainless steel feeding tube (Plymouth Meeting, PA) and the lungs were inflated with 10% neutral buffered formalin at 25 cm-H₂O (2.4 kPa) for one hour. After the trachea was securely tied to the stainless feeding tube to maintain pressure with surgical silk, the lungs were perfusion fixed with an additional aliquot of 10% neutral buffered formalin, surgically removed and then stored in 10% buffered formalin for 24 h before paraffin embedding. Lung sections (5 μm) were mounted on SuperFrost plus-coated slides (Denville Scientific, Metuchen, NJ). Slides were deparaffinized, and sections stained with hematoxylin and eosin (H&E). Histology images were captured with a mounted digital camera using an Olympus IX 51 microscope and a 10× objective. Inflammatory cells, myeloid cells, including neutrophils, monocytes, and macrophages, were stained with MPO antibody (1:200) overnight at 4°C and HPR-conjugated anti-rabbit antibody (1:1000) at room temperature for one hour then visualized by diaminobenzidine to generate a dark-brown color. Blood vessels were stained with RECA-1 and visualized with horseradish-conjugated secondary antibody and diaminobenzidine. Immunofluorescence of 8-OH-dG was used as a biomarker for oxidative DNA damage. Lung sections were stained with the 8-OH-dG antibody (1:100) for overnight at 4°C then treated with AlexaFluor488-conjugated secondary antibody for one hour in room temperature and counterstained with DAPI before imaging with a fluorescent microscope. The average of three sections per pup, and five counts per section (15 counts/pup) was used for statistical analysis. Quantified data were obtained and entered into the record using predetermined codes by one of the coauthors using double-blind protocols.

Morphometric Analysis:

The mean linear intercept (MLI), or chord length, was used as a method to estimate the volume-to-surface ratio of acinar airspaces whereas radial alveolar count (RAC) and secondary septa were investigated to study the complexity of lung structure [33]. Ten equally spaced horizontal lines were drawn on each picture, and the number of intercepts through the alveolar wall was counted. MLI was obtained by multiplying the number of times the traverses are placed on the lung times the length of the traverse and dividing the result by the sum of all the intercepts. For RAC, a line from the center of the respiratory tract perpendicular to the nearest connective tissue septum was drawn, and alveoli intercepting with the line were counted. For measurement of secondary septa, elastin was stained with resorcin fuchsin and Van Gieson’s solution.

Immunoblot Analysis:

Whole lung lysates were obtained by homogenizing in MOPS buffer (20 mM 3-N-morpholino-propane sulfonic acid, 2 mM EGTA, 5 mM EDTA, 30 mM NaF, 10 mM β-glycerophosphate, 10 mM Na pyrophosphate, 2 mM Na orthovanadate, 1 mM PMSF, 0.5% NP-40, 1% protease inhibitor cocktail, and 1% phosphatase inhibitor cocktails 2 and 3, pH 7.0) by Bullet Blender (Next Advance, Inc., Averill Park, NY). For protein immunoblots, 30 μg of protein lysate was separated by SDS-PAGE, transferred to nitrocellulose membranes (0.22 μm), and then probed with the appropriate primary antibodies overnight at 4°C on a continuously rocking platform. Signals were generated after incubation with horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse IgG (1:8,000) using Pierce™ ECL Western Blotting Substrate or SuperSignal™ West Femto Maximum Sensitivity Substrate. Integrated optical density (IOD) was calculated using ImageJ software and normalized to β-actin, a loading control.

Statistical Analysis:

Representative images of lung sections and immunoblots are presented in the figures. Data from lung sections and immunoblots from the samples were analyzed for statistical differences using GraphPad Prism version 8.4.3, for Windows (GraphPad Software, La Jolla, CA). Quantitative morphometric and cell count data are expressed as mean ± SD scatter-plots adjacent to the representative images. Student t-test, or Mann-Whitney U test, was used for comparing two groups wherever appropriate. One-way analysis of variance with post-hoc Student-Newman-Keuls test was used when more than two groups of data were analyzed. Significant differences between groups were determined by either unpaired student’s t-test or Mann-Whitney U test, depending on the distribution. Weight gain and survival data were analyzed using MedCalc Statistical Software version 15.2.1 (MedCalc Software bvba, Ostend, Belgium; http://www.medcalc.org). Kaplan-Meier survival curves were plotted using Kaplan-Meier tables constructed in GraphPad Prism. A p-value <0.05 was considered statistically significant.

Results

Effects of KYC on Hyperoxia-induced Myeloid Cell Recruitment, MPO, and Oxidative Stress:

KYC reduced the number of MPO+ myeloid cells in hyperoxic lungs by 50% (brown cells in Figure 1A). MPO immunoblots show that KYC reduced MPO protein by 50% (Figure 1B) and Cl-Tyr formation by 27% (Figure 1C). These data demonstrate that KYC reduces oxidative stress and myeloid cell recruitment to the lungs of hyperoxic pups. Data demonstrating that chronic hyperoxia increases BPD in neonatal rat pups by P10 as defined by lung morphometrics and mechanism-based biomarkers of oxidative stress and inflammation are provided in on-line supplemental data (Figures S1 and S2).

Figure 1.

KYC Decreased BPD in Hyperoxic Pups. (A) Representative immunohistochemistry images of lung sections from PBS- (Left panel) and KYC-treated, hyperoxic neonatal rat pups. Images show that KYC decreased MPO+ myeloid cell counts relative to MPO+ cell counts in lungs of hyperoxic pups (n=9). The bar chart shows the mean ± SD of MPO+ cell counts in lung sections from hyperoxic pups treated with PBS and KYC, respectively. (B) Immunoblots show that lung lysates from KYC-treated hyperoxic pups contain less MPO protein than lung lysates from PBS-treated hyperoxic pups (n=9). The bar chart shows the mean ± SD of the band densities for MPO relative to actin in lung lysates from hyperoxic pups treated with PBS and KYC. (C) Immunoblots show that lung lysates from KYC-treated hyperoxic pups have less Cl-Tyr immunoreactive proteins than lung lysates from PBS-treated hyperoxic pups (n=9). Quantification of immunoreactive Cl-Try protein densities was by scanning the entire lane and normalizing the integrated scanned profile to actin, an internal control (Hyperoxia = O₂ > 90%; * = p<0.05).

Effects of KYC on Hyperoxia-induced BPD:

If MPO plays a causal role in the mechanisms by which hyperoxia induces BPD, then KYC, which inhibits MPO toxic oxidant production but not peroxidase activity [20], should reduce hyperoxia-induced changes in lung architecture. KYC treatment of hyperoxic pups increased RAC, secondary septa, and blood vessel counts by 35%, 55%, and 92%, respectively, while decreasing MLI by 22% (Figure 2A–2D). Immunoblots for CD31 confirm that KYC increased the number of endothelial cells in hyperoxic lungs (23% more, Figure 2E). The increase in CD31 expression supports immunohistochemistry data showing that KYC increased microvascular blood vessels and alveolar complexity that are both essential for improving lung architecture and decreasing BPD in hyperoxic pups.

Figure 2.

KYC Prevented Alveolar Simplification in Lungs of Hyperoxic Pups. (A) Representative H&E images of lung sections showing alveolar and vascular simplification resulting from chronic hyperoxia that were reversed with KYC treatment. RAC are increased in lung sections from KYC-treated hyperoxic pups (left, n=9) compared to RAC levels in lung sections from PBS-treated hyperoxic pups (right, n=12). The bar chart shows the mean ± SD for RAC from lung sections from hyperoxic pups treated with PBS and KYC. (B) Representative H&E images of lung sections from hyperoxic pups treated with PBS (left) vs. KYC (right). Arrows indicate lungs structures counted as secondary septa. Bar charts summarize secondary septation data and show that KYC treatment of hyperoxic pups increased secondary septation (right, n=9) compared to secondary septation in PBS-treated hyperoxic pups (left, n=12). The bar chart shows the mean ± SD for secondary septa in lung sections from hyperoxic pups treated with PBS and KYC. (C) Representative H&E images showing capillary structures in lung sections from hyperoxic pups treated with PBS (left) vs. KYC (right). Arrows indicate lung structures counted as capillaries. The bar chart shows capillary counts are increased in KYC-treated hyperoxic pups (right, n=9) compared with counts in PBS-treated hyperoxic pups (left, n=12). (D) Representative H&E images of lung sections from PBS-treated (left) and KYC-treated (KYC) hyperoxic pups. The bar chart shows MLI are decreased in lung sections in KYC-treated hyperoxic pups (right, n=9) compared to MLI in lungs of PBS-treated hyperoxic neonatal rat pups (left, n=12). (E) Representative immunoblots for CD31 relative to actin, an internal loading control from lung lysates prepared from KYC-treated hyperoxic pups (right, n=9) and PBS-treated hyperoxic pups. (left, n-9). The bar chart shows the mean ± SD of relative differences in CD31 band densities normalized to actin in lung lysates from PBS and KYC treated hyperoxic pups. (Hyperoxia = O₂ > 90%, * = p<0.05).

Effects of KYC on 8-OH-dG in Hyperoxic Lungs:

Hyperoxia increased 8-OH-dG, a biomarker of DNA oxidative damage in the lungs of neonatal rat pups (see supplemental Figure S3). Immunofluorescence staining for 8-OH-dG showed that KYC treatment reduced 8-OH-dG in the nuclei of lung cells in hyperoxic rat pups (Figure 3).

Figure 3.

KYC Decreased Oxidative DNA Damage in Lungs of Hyperoxic Pups. Representative images of lung sections stained for nuclei (top, DAPI, Blue), immunostained for 8-OH-dG (middle, Red) and combined (bottom) in hyperoxic pups treated with PBS (left) and KYC (right). The bar chart shows the mean ± SD of the relative changes in fluorescent densities for 8-OH-dG in lung sections from hyperoxic pups treated with PBS and KYC (PBS, n = 7; KYC, n = 8, * = p<0.05).

Effects of KYC on Cox-1/Cox-2 in Hyperoxic Lungs:

Hyperoxia increased Cox-1 slightly and Cox-2 markedly in the lungs of neonatal rat pups (see supplemental Figure S4). KYC treatment of hyperoxic neonatal rat pups reduced pulmonary Cox-1 by 26.5% and Cox-2 by 58.1% (Figure 4).

Figure 4.

KYC Decreased Cyclooxygenase-1 (Cox-1) and −2 (Cox-2) Expression in Lungs of Hyperoxic Pups. Representative immunoblots showing that KYC decreased hyperoxia-induced increases in Cox-1 (n=6) and Cox-2 (n=6) in lungs of hyperoxic pups. The bar chart shows relative changes in mean ± SD of Cox-1 and Cox-2 band densities relative to β-Actin showing that KYC treatment reduced Cox-1 and Cox-2 expression in hyperoxic pups. (■: Hyperoxia + PBS; ▲: Hyperoxia + KYC; * = p<0.05).

Effects of KYC on HMGB1 in Hyperoxic Lungs:

Hyperoxia increased HMGB1 levels in lung lysates of neonatal rat pups (see supplemental Figure S5). KYC treatment of hyperoxia pups decreased HMGB1 levels in lung lysates from hyperoxic pups by 38% (Figure 5).

Figure 5.

KYC Decreased HMGB1 Release in Lungs of Hyperoxic Pups. Representative immunoblots for HMGB1 and Actin in lung lysates from hyperoxic pups treated with PBS (left) or KYC (right). These immunoblots show that KYC treatment decreased HMGB1 release in lung lysates isolated from hyperoxic pups relative to differences in Actin, an internal loading control. The bar chart shows relative changes in the mean ± SD of HMGB1 band densities relative to the band densities of Actin in lung lysates from hyperoxic pups treated with PBS or KYC (■: Hyperoxia + PBS, n=10; ▲: Hyperoxia + KYC; n=11, * = p<0.05).

Effects of KYC on RAGE and TLR4 in Hyperoxic Lungs:

Hyperoxia increased RAGE and TLR4 in lungs of neonatal rat pups (see supplemental Figure S6). KYC treatment reduced the expression of RAGE and TLR4 in lung lysates from hyperoxic pups by 60% and 22%, respectively (Figure 6A and 6B).

Figure 6.

KYC Decreased RAGE and TLR4 in Lungs of Hyperoxic Pups. (A) Representative immunoblots for RAGE in lung lysates from hyperoxic pups treated with PBS and KYC (n=16, * = p<0.05). The bar chart presents the mean ± SD of RAGE band densities normalized to actin and show that KYC decreased RAGE expression in lung lysates of hyperoxic pups. (B) Representative immunoblots for TLR4 expression in lungs lysates from hyperoxic pups treated with PBS and KYC (n=10). The bar chart presents the mean ± SD of TLR4 band densities normalized to actin and show that KYC treatment decreased TLR4 expression in lung lysates of hyperoxic pups. (n=12, * = p<0.05).

Effects of Oxidative Stress on KYC Thiylation on Endothelial Cell Proteins in RLMVEC Cultures:

To determine the effects of oxidative stress on KYC thiylation, RLMVEC cultures were treated with B-KYC to follow KYC thiylation with streptavidin affinity-blotting as outlined in Methods. Three different levels of oxidative stress were used to determine the effects of oxidative stress on KYC thiylation of endothelial cell proteins: media alone (baseline); media containing MPO+H₂O₂ (MPO-dependent); and media containing H₂O₂ alone (H₂O₂-dependent) (Figure 7A). KYC thiylation of RLMVEC proteins at baseline was low (Figure 7A, first two lanes). MPO-dependent KYC thiylation of RLMVEC proteins was markedly increased (Figure 7A, middle two lanes). H₂O₂-dependent KYC thiylation of RLMVEC proteins was more than observed at baseline, but not as high as MPO-dependent B-KYC thiylation. These studies show that MPO-dependent thiylation induces the greatest level of KYC thiylated RLMVEC proteins (Figure 7F, first three bars).

Figure 7.

Effects of Oxidative Stress on KYC Thiylation of RLMVEC Proteins. (A) Streptavidin affinity-blot of KYC thiylated proteins from cell lysates prepared from RLMVEC cultures treated with B-KYC (baseline), B-KYC+MPO+H₂O₂ (MPO-dependent), and B-KYC+H₂O₂ (H₂O₂-dependent). (B) Immunoblot for Nrf2 in RLMVEC cultures treated as described in A. (C) Immunoblot for Keap1 in RLMVEC cultures treated as described in A. (D) Immunoblot for HMGB1 in RLMVEC cultures treated as described in A. (E) Immunoblot for β-Actin in RLMVEC cultures treated as described in A. (F) Bar chart showing relative levels of KYC thiylation, Nrf2, Keap1, and HMGB1 as a function of β-Actin (n=2, * = p<0.05, statistical analysis by ANOVA) These data show that MPO-dependent oxidation of KYC increases KYC thiylation in RLMVEC proteins cells that are proximal to MPO, while H₂O₂-dependent oxidation induces a slight, if any, increase in KYC thiylation of RLMVEC proteins.

Immunoblot band densities for Nrf2 (7B), Keap1 (7C), and HMGB1 (7D) relative to β-actin (7E) reveal that protein expression for the three proteins is differentially modulated by oxidative stress. Nrf2 expression is increased in RLMVEC cultures subjected to MPO-dependent KYC thiylation more than in RLMVEC cultures subjected to KYC thiylation under baseline or H₂O₂-dependent oxidation (Figure 7B and 7F). Keap1 expression followed a pattern similar to Nrf2 (Figure 7C and 7F). In contrast, HMGB1 expression increased in RLMVEC cultures subjected to H₂O₂-dependent KYC thiylation more than in RLMVEC cultures subjected to KYC thiylation at baseline or in RLMVEC cultures subjected to MPO-dependent KYC thiylation (Figure 7D and 7F). These data are consistent with the idea that MPO-dependent KYC thiylation is protective, while H₂O₂-dependent KYC thiylation may be injurious even when KYC is present because MPO isn’t available to degrade H₂O₂ [20], which would reduce oxidative stress.

MPO Oxidizes KYC to a Thiyl Radical that Thiylates HMGB1:

Figure 8A shows a four-line ESR spectrum corresponding to DMPO-S•-KYC. This spectrum is similar to the spectrum of DMPO-S•-YC generated when the dipeptide YC was added to an MPO reaction mixture containing DMPO [34]. Figure 8B shows an autoradiogram of the bands corresponding to MPO and HMGB1 from a fluorescent streptavidin affinity-blot. The fluorescent density of the bands for B-KYC thiylated HMGB1 and MPO are decreased in the sample treated with dithiothreitol (DTT, 100 mM, final concentration, second lane). As DTT is a potent disulfide reducing agent, the decrease in band density confirms that the bonds between KYC and HMGB1 and MPO are disulfides.

Figure 8.

MPO Oxidizes KYC to a KYC Thiyl Radical That Thiylates HMGB1. (A) ESR spectrum of KYC thiyl S•-DMPO generated by an MPO reaction mixture containing MPO (120 nM), H₂O₂ (50 μM) and KYC (30 μM). (B) Streptavidin affinity-blot of KYC thiylated HMGB1 and MPO in a split sample, one half treated neat and other half treated with DTT (100 mM) to reduce disulfide bonds. The affinity-blot shows that DTT reduced B-KYC-thiylated HMGB1 and B-KYC-thiylated MPO fluorescent band densities, confirming that the bond between KYC and HMGB1 and MPO were both disulfides.

MPO Oxidation of KYC Results in KYC Thiylation of HMGB1 Released from RLMVEC Cultures:

To determine if MPO oxidizes KYC to a thiyl radical that thiylates HMGB1 released from RLMVEC cultures, proteins in the media and the cell lysates from RLMVEC cultures at baseline with no KYC, with KYC and with MPO+H₂O₂+KYC were streptavidin affinity-blotted for KYC-thiylated proteins and immunoblotted for HMGB1. HMGB1 immunoblots showed that all RLMVEC cultures released low amounts of HMGB1 (Figure 9A, lower immunoblot, all nine lanes) into the conditioned media. However, KYC thiylation of HMGB1 was greater in RLMVEC cultures incubated with MPO+H₂O₂+KYC (Figure 9A, upper affinity-blot) than the other two conditions. Similar differences across the three test groups can be seen in the streptavidin affinity-blots of cell lysates (Figure 9B, upper affinity-blot) compared with the HMGB1 immunoblot (Figure 9B, lower blot).

Figure 9.

MPO Oxidation of KYC Increases KYC Thiylation of Extracellular HMGB1. (A) Streptavidin affinity-blot of conditioned media from RLMVEC cultures treated with media, media + B-KYC, and media + MPO reaction system (MPO = 120 nM; H₂O₂ = 50 μM) + B-KYC. The affinity-blot shows that the MPO reaction system + B-KYC predominately thiylates HMGB1 released into the conditioned media relative to the immunoblot levels of HMGB1 protein released (immunoblot for HMGB1). (B) Streptavidin affinity-blot of cell lysates of RLMVEC cultures treated with media, media + B-KYC, and media + MPO reaction system (MPO = 120 nM; H₂O₂ = 50 μM) + B-KYC. This affinity-blot shows that the MPO reaction system + B-KYC thiylates low levels of HMGB1 in cell lysates. This conclusion is based on the relative levels of KYC thiylated HMGB1 in panel A vs. KYC thiylated HMGB1 as a function of HMGB1 protein in conditioned media vs. RLMVEC cell lysates.

Although no efforts were made to quantify B-KYC thiylated HMGB1 in the conditioned media and cell lysates (intra-cellular HMGB1), it should be possible to appreciate relative differences in B-KYC thiylation of HMGB1 from the ratio of B-KYC-thiylated HMGB1 (upper affinity blot) to HMGB1 (lower immunoblot) in Figures 9A and 9B. The heavy bands for B-KYC-HMGB1 (upper affinity blot) relative to the bands for HMGB1 (lower immunoblot) in Figure 9A suggest that HMGB1 in the conditioned media is greater than the light bands for B-KYC-HMGB1 (upper affinity blot) relative to the bands for HMGB1 (lower immunoblot) in Figure 9B. The affinity blot and immunoblot data are consistent with the idea that MPO-generated KYC thiyl radicals have greater access to extracellular HMGB1 than to intracellular HMGB1 in the nucleus.

Effects of KYC Treatment on HMGB1 Association with TLR4 and RAGE in Hyperoxic Lungs:

Although KYC treatment reduced total HMGB1 in the lung lysates of hyperoxic neonatal rat pups (Figure 5), precisely how KYC thiylation of HMGB1 alters the interaction of HMGB1 with TLR4 and RAGE is unknown. HMGB1 pulldown assays revealed that KYC treatment decreased HMGB1 association with TLR4 in lung lysates from hyperoxic neonatal rat pups (Figure 10A and 10C). Please note, even though the binding affinity of HMGB1 for TLR4 is reported to be weak [35], the band corresponding to HMGB1 was still visible in all lanes (Figure 10A). KYC treatment also reduced HMGB1’s association with RAGE in the lungs of hyperoxic neonatal rat pups (Figure 10B and 10C).

Figure 10.

Effects of KYC Treatment on HMGB1 Association with TLR4 and RAGE in Lung Lysates from Hyperoxic Pups. (A) Immunoblots of TLR4 associated with HMGB1 in lung lysates from PBS- and KYC-treated hyperoxic pups. A non-depleting concentration of anti-HMGB1 antibody was used to pull-down HMGB1 from lung lysates. The pull-down was immunoblotted for TLR4 and HMGB1. These immunoblots show that the levels of TLR4 association with HMGB1 decreased in lung lysates from KYC-treated hyperoxic pups relative to the levels of TLR4 associated with HMGB1 in PBS-treated hyperoxic pups. The lower immunoblot shows that non-depleting levels of anti-HMGB1 antibody pull-down essentially equal levels of HMGB1 from each sample irrespective of how treatments modulated lung TLR4 expression (Figure 6, supplemental data Figure S6). (B) The immunoblots show how much RAGE is associated with HMGB1 in lung lysates from PBS- and KYC-treated hyperoxic pups. A non-depleting concentration of anti-HMGB1 antibody was used to pull-down HMGB1 from lung lysates. The pull-down was immunoblotted for RAGE and HMGB1. These immunoblots show that the levels of RAGE association with HMGB1 decreased in lung lysates prepared from KYC-treated hyperoxic pups relative to the levels of RAGE associated with HMGB1 in PBS-treated hyperoxic pups. The lower immunoblot shows that non-depleting levels of anti-HMGB1 antibody pull-down essentially equal levels of HMGB1 from each sample irrespective of how treatments modulate lung RAGE expression (Figure 6, supplemental data Figure S6). (C) The bar charts show the mean ± SD of relative levels of TLR4 and RAGE associated with HMGB1 in lung lysates from PBS- and KYC-treated hyperoxic pups (n=4, * = p<0.05).

Effects of KYC Thiylation on HMGB1, Terminal HMGB1 Thiol Oxidation, and Shifts in Cell Sources for HMGB1 in Hyperoxic Lungs:

Immunoblots of normoxic and hyperoxic lung lysates revealed that lungs from normoxic pups had higher levels of sulfonyl HMGB1 than lungs from hyperoxic pups (Figure 11A). Furthermore, KYC treatment of hyperoxic pups increased sulfonyl HMGB1 levels relative to sulfonyl HMGB1 levels in PBS-treated hyperoxic pups (Figure 11B). These data suggest that KYC thiylation increased terminal HMGB1 thiol oxidation, which is well-known to inactivate HMGB1 [36]. These data support the idea that KYC thiylation broke the destructive cycle in BPD by promoting HMGB1 terminal thiol oxidation. In addition to decreasing total HMGB1 levels in lung lysates (Figure 5), non-depleting HMGB1 immuno-pulldown studies showed that KYC treatment shifted the HMGB1 in lung lysates from the non-acetylated (HMGB1) isoform to the lysyl-acetylated (Ac-HMGB1) isoform (Figure 11C). These blots show that lung lysates from PBS-treated hyperoxic neonatal rat pups contained primarily non-acetylated-HMGB1, while lung lysates from KYC-treated hyperoxic neonatal rat pups contained predominantly lysyl-acetylated-HMGB1. As Ac-HMGB1 is secreted by activated immune cells, data in Figure 11C are consistent with the idea that KYC treatment reduced pulmonary cell death, which is the most likely cellular source of non-acetylated HMGB1 in the hyperoxic lungs.

Figure 11.

Effects of KYC on HMGB1 Thiol Oxidation State and Lysyl-Acetylation. (A) Representative immunoblots for cysteine sulfonyl on HMGB1 in lung lysates from normoxic and hyperoxic pups. Bar chart presents mean ± SD of relative levels of cysteine sulfonyl on HMGB1 in lung lysates from normoxic and hyperoxic pups. These data show that hyperoxia decreased the levels of cysteine sulfonyl on HMGB1 in lung lysates. (B). Representative immunoblots for cysteine sulfonyl on HMGB1 in lung lysates from hyperoxic pups treated with PBS or KYC. The bar chart presents mean ± SD of relative levels of cysteine sulfonyl on HMGB1 in lung lysates from hyperoxic pups treated with PBS or KYC. These data show that KYC treatment increased the levels of cysteine sulfonyl on HMGB1 in lung lysates of hyperoxic pups. (C) These data suggest that activated immune cells are the predominant source of lysyl-acetylated HMGB1. Dead and dying cells are the predominant source of non-acetylated HMGB1. Representative immunoblots for lysyl-acetylated residues on HMGB1 immunoprecipitated from lung lysates from hyperoxic pups. HMGB1 was immunoprecipitated with non-depleting concentrations of anti-HMGB1 antibody. The immunoblots show difference in lysyl-acetylated HMGB1 in lung lysates from hyperoxic pups with PBS and KYC treatments. The hyperoxic pups treated with PBS have predominantly non-acetylated HMGB1, which is released by dead and dying cells. In contrast, the predominant HMGB1 isoform in lung lysates from KYC treated hyperoxic pups is lysyl-acetylated HMGB1. These data demonstrate that HMGB1 release in lungs of hyperoxic pups is shifted from dead and dying lung cells in PBS-treated hyperoxic pups to activated immune cells in KYC-treated hyperoxic pups.

Effects of KYC on Keap1 KYC Thiylation and S-Glutathionylation, and Modulation of Nrf2 in Normoxic and Hyperoxic Lungs:

Band density analysis of streptavidin affinity blots and immunoblots revealed that KYC treatment of hyperoxic pups increased B-KYC thiylation of Keap1 and S-glutathionylation of Keap1 in lung lysates from hyperoxic pups (Figure 12A and 12B). As KYC increased B-KYC thiylation and S-glutathionylation of Keap1 in RLMVEC cultures even in the absence of MPO (Figure 9), we examined KYC-dependent activation of Nrf2 in lung lysates from normoxic and hyperoxic pups. KYC treatment of normoxic pups increased Nrf2 levels in lung lysates (Figure 12C) as well as in the lung lysates from hyperoxic neonatal rat pups (Figure 12D). Data in Figure 12D showing that KYC treatment increased Nrf2 levels in lung lysates from hyperoxic pups are in stark contrast with supplement data in Figure S7 showing that Nrf2 levels were markedly reduced in lung lysates from hyperoxic pups when compared to Nrf2 levels in lung lysates from normoxic pups.

Figure 12.

Effects of KYC Treatment on KYC Thiylation of Keap1, Keap1 S-Glutathionylation and Nrf2 Activation in Lungs of Hyperoxic Pups. (A) The streptavidin affinity-blot for B-KYC-thiylated Keap1 and the immunoblot for Keap1 show that KYC treatment increased B-KYC-thiylation of Keap1 in lung lysates from hyperoxic pups. The bar chart shows the mean ± SD of the relative levels of B-KYC-thiylated Keap-1 as a function to Keap1 in lung lysates of hyperoxic pups treated with PBS and KYC. These data show that KYC treatment increased KYC thiylation of Keap1 in the lungs of hyperoxic pups. (B) Immunoblot for S-glutathionylated Keap1 shows that KYC treatment increased Keap1 S-glutathionylation in lung lysates from hyperoxic pups. The bar chart shows the mean ± SD of the relative levels of GS-thiylated Keap-1 as a function to Keap1 in lung lysates of hyperoxic pups treated with PBS and KYC. These data show that KYC treatment increased S-glutathionylated Keap1 (GS-Keap1) in the lungs of hyperoxic pups. (C) The immunoblot shows the relative levels of Nrf2 in lung lysates from normoxic pups treated with PBS and KYC. The bar chart shows the mean ± SD of the relative levels of Nrf2 as a function to actin in lung lysates of normoxic pups treated with PBS and KYC. These data show that KYC treatment increased Nrf2 activation in lungs of normoxic pups. (D) Representative immunoblots of relative levels of Nrf2 activation in lung lysates from hyperoxic pups treated with PBS and KYC. The bar chart shows the mean ± SD of the relative levels of Nrf2 as a function of actin in lung lysates of hyperoxic pups treated with PBS and KYC. These data show that KYC treatment increased Nrf2 activation in lungs of hyperoxic pups.

Effects of KYC Treatment on HO-1, GST, and Trx Expression in Hyperoxic Lungs.

The immunoblots show that treating hyperoxic pups with KYC increased the expression of HO-1, GST, and Trx, which are all members of the antioxidant enzyme system known to be upregulated by Nrf2 (Figure 13). These findings demonstrate that KYC effectively increased Nrf2-dependent antioxidant enzyme expression in the lungs of hyperoxic pups.

Figure 13.

Effects of KYC on Antioxidant Enzyme Expression in Lungs from Hyperoxic Pups. Representative immunoblots show that KYC treatment increased HO-1, GST and Trx expression in lungs of hyperoxic pups. The bar charts show mean ± SD of the relative levels of HO-1, GST and Trx as a function of actin in lung lysates of hyperoxic pups treated with PBS and KYC (n = 9, * = p<0.05).

Effects of KYC on Weight Gain and Survival of Hyperoxic Pups:

KYC-treated normoxic pups gained more weight than untreated normoxic pups based on curve analysis (Figure 14A). Although KYC-treated hyperoxic pups gained significantly more weight from P3 to P10 than PBS-treated hyperoxic pups (inset bar graph in Figure 14C), daily time-dependent weight gain was not significantly different between the two test groups until P10, the last day of the study. These data are in contrast to the effects of KYC on normoxic pups that gained significantly more weight daily than untreated normoxic pups (Figure 14A). It is unclear why KYC did not improve the weight gain in the hyperoxic test group like it did in the normoxic test group. It could be argued however, that the improvements in alveolar formation without notable increases in weight gain support the idea that KYC attenuated hyperoxic lung injury by decreasing toxic oxidant-mediated lung injury, not by improving nutritional intake. Concerning survival analysis, although more normoxic pups treated with KYC survived to P10 than untreated normoxic pups, no significant differences in survival were detected between these two groups. (Figure 14B). In contrast, KYC-treated hyperoxic pups had a higher probability of surviving to P10 than PBS-treated hyperoxic pups (Figure 14D).

Figure 14.

Effects of KYC on Weight Gain and Survival. (A) Line plot showing means of weight gain of normoxic pups treated with KYC. Statistical analysis reveals that KYC increased weight gain in normoxic pups from P3 to P10 and that corrected multiple t-test analysis shows significant differences between means for different days in addition to significant differences between curves. (21% O₂, n = 13, 21% O₂ + KYC, n = 14, * = p<0.05, ** = p<0.01). (B) Survival analysis plots the probability of “survival” versus time postpartum for normoxic pups and normoxic pups treated with KYC finds no significant differences in survival. (C) Line plot showing means of weight gain of hyperoxic pups treated with PBS or KYC. Statistical analysis reveals that KYC increased weight gain in normoxic pups from P3 to P10 and that corrected multiple t-test analysis shows a significant difference between means only for P10 in addition to a significant difference between curves. (>90% O₂, n = 22, >90% O₂ + KYC, n = 27, * = p<0.05, ** = p<0.01). (D) Survival analysis plots the probability of survival versus time postpartum for hyperoxic pups treated with PBS or KYC. Analysis reveals that KYC significantly improved the survival of hyperoxic pups. (>90% O₂, n = 22, >90% O₂ + KYC, n = 27, * = p<0.05).

Discussion

Our findings suggest that MPO, HMGB1, and Nrf2 play essential roles in BPD. MPO, released from resident myeloid cells, plays an initiating role by amplifying the oxidative stress induced by hyperoxia. HMGB1, released from dead and dying lung cells, plays a propagating role by recruiting the myeloid cells that enter the interstitium of the lung and release MPO to increase inflammation. The arrival and activation of the recruited myeloid cells induces a second wave of oxidative injury and cell death that is always greater than that which is induced initially by hyperoxia. Reduced Nrf2 activation plays a permissive role in BPD by preventing premature neonates who lack antioxidant enzymes from defending themselves against the oxidative stress mediated by chronic hyperoxia. In this way, blunted Nrf2 responses to the oxidative stress induced by hyperoxia allows hyperoxia to induce greater pulmonary cell injury and death [14], which result in proportionately greater release of HMGB1 [12]. Viewed in this fashion, failure to target adequately all three mediators permits the cycle to continue, resulting in varying degrees of BPD severity. Our findings suggest that BPD is caused by a destructive cycle that is mediated by oxidative stress and inflammation, where activation of one mediator results in the activation of a downstream second mediator to worsen BPD. As myeloid cells are capable of releasing various mediators of oxidative stress and inflammation, and HMGB1 increases vascular inflammation, it is likely that additional mediators of inflammation and oxidative stress will be added to this list over time.

The mechanisms by which hyperoxia impairs lung development and increases BPD are complex. To begin to understand the cellular mechanisms mediating BPD, we first determined the effects of hyperoxia on the lungs of neonatal rat pups. After establishing baseline differences for our biomarkers of interest, we determined KYC’s effects on preventing oxidative lung damage in the hyperoxic pups. Our studies showed that chronic hyperoxia decreased lung complexity and CD31 expression, a cellular biomarker for endothelial cell content. These data, taken together, suggest hyperoxia impaired angiogenesis. Hyperoxia increased lung MPO+ myeloid cell counts, MPO protein, and Cl-Tyr formation (a biomarker for HOCl-dependent oxidative damage). These data suggest that hyperoxia increased oxidative stress. In this context, it is significant that the increase in Cl-Tyr was paralleled by a second biomarker, 8-OH-dG, a biomarker of DNA oxidative damage, and that both biomarkers increased as lung development decreased and cell death increased [37]. Inflammation was assessed by immunoblotting, which showed that hyperoxia increased HMGB1, Cox-1/Cox-2, RAGE and TLR4. HMGB1 is a biomarker for lung cell death and injury, myeloid cell recruitment and vascular inflammation. Cox-1/Cox-2, RAGE and TLR4 are all biomarkers of inflammation. Finally, other immunoblots revealed that hyperoxia decreased Nrf2 activation (see supplemental data, Figure S7). The inability of the lungs of neonatal rat pups to activate Nrf2 makes the pups more susceptible to oxidative stress.

Treating hyperoxic pups with KYC improved lung complexity, reduced oxidative stress and inflammation, and reversed the biomarkers for oxidative stress, DNA damage and inflammation in the lungs of hyperoxic pups. Thus, KYC treatment effectively reduced oxidative stress and inflammation, and restored lung development in the hyperoxic pups, which likely played a critical role in increasing pup survival.

The fact that KYC inhibits multiple mechanisms suggests that KYC is a systems pharmacology agent. In 2013, we reported that KYC not only inhibited MPO-dependent HOCl and nitrogen dioxide (NO₂^•) production, but also increased MPO-dependent H₂O₂ consumption [20], which, all things being equal, should also reduce oxidative stress. Our idea that KYC increases MPO-dependent H₂O₂ consumption is consistent with the report by Kettle and Winterbourn [21], who showed that tyrosine increases MPO-dependent catalase activity. In 2013, we also reported that MPO oxidized KYC to a KYC thiyl radical that prevented further oxidative damage by autoscavenging to form either a homodimer with a second KYC or a heterodimer with GSH [20]. As all thiols require activation before they can thiylate a thiol group or a protein cysteine, we reasoned that KYC should reduce oxidative stress in the very places where myeloid cells release MPO. Such a model implies that the tissues benefiting the most from KYC should be proximal to MPO. Although our intent in developing this model was to address non-specificity, embracing such a model demanded that we take a closer look at KYC thiylation.

To better understand thiylation, biotin-labeled KYC (B-KYC) was added to RLMVEC cultures under baseline, H₂O₂-dependent, and MPO-dependent conditions of oxidative stress. At baseline, KYC thiylation of RLMVEC proteins was low. However, under H₂O₂-dependent oxidative stress conditions, KYC thiylation of RLMVEC proteins was only slightly increased, the majority of RLMVEC proteins that were KYC thiylated were not heavily thiylated and were scattered throughout the full range of molecular weights. Under MPO-dependent oxidative stress conditions, the level of KYC thiylation of RLMVEC proteins was markedly increased throughout a wide range of molecular weights. These data support our model that KYC thiylation occurs predominantly in cells proximal to MPO. Although oxidative stress may increase KYC thiylation in a few cell proteins, the KYC thiylation that occurs is diffuse and not much more than that which occurs at baseline. Protein expression data show that Keap1 and Nrf2 were increased in RLMVEC cultures treated MPO+H₂O₂+B-KYC, while HMGB1, a likely injury mechanism, was increased in the RLMVEC cultures treated with H₂O₂-dependent oxidative stress + B-KYC but not in RLMVEC cultures treated with MPO+H₂O₂+B-KYC. These data support our idea that B-KYC is an MPO substrate that increases MPO-dependent catalase activity [21] and reduces oxidative injury and damage to RLMVEC cultures.

Previously, we reported that KYC reduced MPO-dependent oxidative damage because KYC displaces native substrates, such as chloride and nitrite, that are oxidized to “toxic” oxidants. In 2013, we argued that KYC reduced MPO-dependent oxidative damage because the resultant KYC thiyl radical would autoscavenge to prevent radical propagation [20]. HMGB1 expression data in Figure 7 suggests that MPO-dependent KYC thiylation protects RLMVEC against oxidative injury. To determine the radical mechanisms mediating MPO oxidation of KYC and KYC thiylation of HMGB1, we performed two experiments. In the first experiment, we trapped KYC thiyl radicals generated by MPO with DMPO (Figure 8A), which confirms our 2013 report [20]. In the second experiment, we added recombinant HMGB1 to PBS containing MPO+H₂O₂+B-KYC and determined KYC thiylation by streptavidin affinity blotting. The affinity blot showed that MPO oxidizes KYC into a new product that thiylates MPO and HMGB1. Evidence for KYC binding MPO and HMGB1 via a disulfide bond comes from DTT reductions in band density for MPO and HMGB1 in the split sample. These two in vitro studies confirm that MPO oxidizes KYC into a thiyl radical (Figure 8A) and that MPO-generated KYC thiyl radicals thiylate HMGB1 and MPO (Figure 8B).

Although the exact mechanisms by which KYC thiylation of HMGB1 promotes terminal HMGB1 thiol oxidation remain unclear, some insight may be gained into the reactions promoting terminal thiol oxidation from in vitro studies. In vitro incubations show that GST reduces KYC disulfide to KYC monomers only in the presence of excess GSH (see supplemental data, Figure S8). The importance of these observations is they confirm that GSH metabolizing enzymes do not metabolize heterodisulfides made with tripeptides other than GSH. Accordingly, in extracellular spaces, where GSH and other reducing agents are in limited supply, KYC thiylation of HMGB1 will prevent HMGB1 from productive binding to RAGE and TLR4. If KYC-thiylated HMGB1 cannot productively bind to its receptors, then the KYC-thiylated HMGB1 will remain in the interstitium or circulation until its cysteines become fully oxidized, which is a form of HMGB1 that is completely inactive [38]. In contrast, when KYC thiyl radical thiylates intracellular Keap-1, where GSH is abundant, KYC may be removed from the thiylated Keap-1 via GSH exchange. The exact mechanisms by which KYC thiylation of intracellular Keap-1 occurs and increases S-glutathionylation have not been fully examined. However, it is interesting to speculate that if KYC is removed from Keap-1 cysteines by GSH exchange, then the KYC monomer will be released immediately adjacent to the other reduced cysteines in Keap-1’s cysteine clamp. Thus, the released monomeric KYC would have an opportunity to thiylate a second cysteine in the clamp. Theoretically, such a KYC thiylation/GSH exchange mechanism could be repeated until Keap-1 is fully glutathionylated, a process that would certainly ensure Nrf2 activation. Additional studies are required to identify and define the mechanisms by which KYC treatment leads to KYC thiylation and, subsequently, S-glutathionylation of Keap-1 to activate Nrf2 and increase antioxidant gene expression. Data showing that KYC treatment of hyperoxic pups increases lung HO-1, GST, and Trx expression confirm that KYC increases Nrf2 activation, which increases antioxidant enzyme expression in lungs of hyperoxic pups. These data are in stark contrast to the effects of hyperoxia, which decreased Nrf2 activation, and increased 8-OH-dG in the lungs of hyperoxic pups (see supplemental Figures 7 and 3, respectively).

Relative rates of weight gain in neonatal rat pups are commonly used as an independent measure of development. Slower rates of weight gain suggest that the pups are under-developed or suffering from excess stress. Survival curves reveal which experimental test groups are under stress or lack the nutrition required to survive. In this context, it is important to note that KYC treatment of normoxic rat pups significantly increased weight gain (Figure 14A). Although more KYC treated normoxic pups survived to P10 than untreated pups, the differences were not significant. In contrast, although the curves for hyperoxic pups treated with PBS and KYC are closer to each other, statistical analysis clearly shows that the KYC-treated hyperoxic pups gained weight at a higher rate than the PBS-treated hyperoxic pups even though differences between daily means did not achieved statistical significance until P10. KYC treatment significantly improved survival of the hyperoxic pups. These data are consistent with the marked increases in lung development, and reductions in pulmonary oxidative stress and inflammation mediated by KYC treatment.

In conclusion, BPD is caused by a destructive cycle mediated by oxidative stress and inflammation. The major mediators are the recruited myeloid cells that release MPO and generate H₂O₂ that activates MPO to generate toxic oxidants, HMGB1 released from dead and dying pulmonary cells, and inadequate Nrf2 activation. KYC is a novel systems pharmacology agent that we previously reported inhibits MPO toxic oxidant production and enhances MPO-dependent catalase activity [20]; and, that we now show here exploits MPO peroxidase activity to be oxidized into a KYC thiyl radical that thiylates and inactivates HMGB1, and thiylates Keap-1 to increase Nrf2 activation. As Nrf2 mediates antioxidant enzyme expression, KYC reduces oxidative stress and inflammation and increases the lung’s ability to protect itself against the chronic oxidative stress induced by hyperoxia. These observations begin to explain how KYC treatment improves lung development in neonatal rat pups raised from birth under conditions of chronic hyperoxia.

Manuscript Highlights.

• The mechanisms mediating the on-set and progression of BPD are poorly defined.

• Hyperoxia, myeloid cells and MPO initiate, while HMGB1 propagates BPD.

• KYC exploits MPO peroxidase activity to be oxidized to a KYC thiyl radical (KYC•).

• KYC• thiylates HMGB1, which inhibits HMGB1 cytokine and chemokine activity.

• • KYC• thiylates Keap1, which activates Nrf2 and increases antioxidant enzymes.