Carbon Monoxide Reversibly Alters Iron Homeostasis ... - ATS Journals

Carbon Monoxide Reversibly Alters Iron Homeostasis and Respiratory Epithelial Cell Function Andrew J. Ghio1, Jacqueline G. Stonehuerner1, Lisa A. Dailey1, Judy H. Richards1, Michael D. Madden1, Zhongping Deng2, N.-B. Nguyen3, Kimberly D. Callaghan3, Funmei Yang3, and Claude A. Piantadosi4 1

National Health and Environmental Effects Research Laboratory, Office of Research and Development, United States Environmental Protection Agency, Research Triangle Park, North Carolina; 2Center for Environmental Medicine, Asthma and Lung Biology, University of North Carolina, Chapel Hill, North Carolina; 3Department of Cellular and Structural Biology, University of Texas Health Science Center, San Antonio, Texas; and 4Department of Medicine, Duke University Medical Center, Durham, North Carolina

The dissociation of iron from heme is a major factor in iron metabolism and the cellular concentrations of the metal correlate with heme degradation. We tested the hypotheses that (1) exposure to a product of heme catabolism, carbon monoxide (CO), alters iron homeostasis in the lung and in cultured respiratory epithelial cells; (2) this response includes both decreased uptake and increased release of cell metal; and (3) the effects of CO on cell function track changes in metal homeostasis. In rats exposed to 50 ppm CO for 24 hours, nonheme iron concentrations decreased in the lung and increased in the liver. In respiratory epithelial cells cultured at air–liquid interface, CO exposure decreased cell non-heme iron and ferritin concentrations within 2 hours and the effect was fully reversible. CO significantly depressed iron uptake by epithelial cells, despite increased expression of divalent metal transporter-1, while iron release was elevated. The loss of non-heme iron after CO reduced cellular oxidative stress, blocked the release of the proinflammatory mediator (interleukin-8), and interfered with cell cycle protein expression. We conclude that CO reduces the iron content of the lung through both the metal uptake and release mechanisms. This loss of cellular iron after CO is in line with certain biological effects of the gas that have been implicated in the protection of cell viability. Keywords: oxidants; oxidative stress; inflammation; cell proliferation; ferritin

Iron is an essential micronutrient for normal cell and tissue function but its complex biochemistry leads to intricacy in maintaining cellular metal homeostasis. Under aerobic aqueous conditions, iron forms oxyhydroxides that are often biologically inaccessible. In addition, iron-catalyzed oxidant production creates the potential for cell damage. As a result, cells have acquired strategies to procure adequate iron for homeostasis and function. The dietary requirement for iron (z 1 mg daily) is small in relation to the body’s iron stores, approximately 4 g, and most of the iron used by mammalian cells (70%) is found within heme moieties (1). Therefore, the dissociation of iron from heme and its release by the cell are major factors in heme–protein homeostasis. The mechanisms for reutilization of iron bound in heme are not fully understood. Heme degradation is the responsibility of the heme oxygenases (HO), which convert it to biliverdin, iron, and carbon monoxide. For instance, the induction of HO-1 alters iron stores; the metal is released in a labile form within minutes to hours, suggesting that its egress is associated with

(Received in original form May 18, 2007 and in final form November 20, 2007) Correspondence and requests for reprints should be addressed to Andrew J. Ghio, MD, Campus Box 7315, Human Studies Division, US EPA, 104 Mason Farm Road, Chapel Hill, NC 27599-7315. E-mail: [email protected] Am J Respir Cell Mol Biol Vol 38. pp 715–723, 2008 Originally Published in Press as DOI: 10.1165/rcmb.2007-0179OC on January 18, 2008 Internet address: www.atsjournals.org

CLINICAL RELEVANCE This research impacts clinical medicine by providing a mechanism of biological effect exerted by carbon monoxide.

either HO activity or a product of the enzyme (2). Cells transfected with HO demonstrate not only a decrease in iron uptake but also an increase in iron release (3). HO knockout mice develop iron-deficiency anemia, but also accumulate significant amounts of iron in several tissues, including the liver and kidneys (4). Similarly, anemia and elevated tissue iron have been reported in a 6-year-old male with HO deficiency (5). In cells, tissues, and organs, increased HO activity is associated with iron release while loss in HO function causes the metal to accumulate. Carbon monoxide (CO) is one of the three products of HO. This gas has a high affinity for the ferrous ion and acts as a monodentate ligand (forms only one bond with the central atom) to complete the coordination shell of iron. CO binding of iron affects the reactivity of the metal and its homeostasis (6). The physiological and pathogenic effects of numerous biological agents correspond to their capacity to influence iron homeostasis. Moreover, another diatomic gas like CO, nitric oxide (NO), plays an important role in iron metabolism, and certain effector functions depend on its ability to bind this metal (6). Similarly, CO alters iron metabolism by tight coordination of free ferrous iron as the carbonyl or by binding iron-containing macromolecules, such as ferrous heme proteins, thereby interfering with or activating reactions central to cell function. In this work, we tested the hypothesis that (1) CO alters iron homeostasis in the intact lung and in cultured respiratory epithelial cells, (2) this alteration includes both decreased uptake and increased release of the metal from lung cells, and (3) the cell functional effects of CO correspond to changes in iron homeostasis consistent with the apparent antioxidant and antiproliferative actions of the gas.

MATERIALS AND METHODS Animal Exposures The experimental protocol was reviewed and approved by the Institutional Animal Care and Use Committee at Duke University. Adult male Sprague-Dawley rats (300–400 g body weight) from Charles River Laboratories (Wilmington, MA) were used for the experiments. For the exposure, 1% CO in air was bled into Plexiglas exposure cages to achieve the desired concentrations and steady state. The CO concentration was monitored continuously using a calibrated infrared CO detector (Snifit Model 50; Bacharach, Pittsburgh, PA). The CO concentration was maintained within 6 5 ppm of the desired level. The rats were exposed to either filtered air or 50 ppm CO for 24 hours. The rats were anesthetized with halothane and an aliquot of blood collected by cardiac puncture. The animal was killed by transecting the abdominal aorta and the trachea was cannulated for lung

716

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 38 2008

lavage with 0.9% NaCl (saline). The volume of saline instilled was equal to 90% of the total lung capacity (35 ml/kg of body weight) based on published allometric equations. Saline was withdrawn after a 3second pause, re-instilled twice more in identical fashion, and centrifuged (600 3 g for 10 min) to remove cells. The lungs and liver were resected en bloc.

Measures of Inflammation and Injury in the Animal Model Carboxyhemoglobin (COHb) concentrations in the blood were measured using a CO-oximeter calibrated with an algorithm for the rat (Model 482; Instrumentation Laboratories, Waltham, MA). The cell differential in the lavage fluid was determined employing a modified Wright’s stain (Diff-Quick stain; American Scientific Products, McGaw Park, IL). After centrifugation to remove cells, lavage protein, lactate dehydogenase (LDH), and albumin concentrations were measured using the Pierce Coomassie Plus Protein Assay Reagent (Pierce, Rockford, IL) and commercially available kits from Sigma (St. Louis, MO) and Thermo DMA (Pittsburgh, PA), respectively. These were modified for use on the Konelab 30 automated clinical chemistry analyzer (Thermo Clinical Labsystems, Espoo, Finland). Macrophage inflammatory protein (MIP)2 in the lavage fluid was measured using a commercially available enzyme-linked immunosorbent assay (ELISA) kit (R&D, Minneapolis, MN). Ferritin and transferrin protein concentrations in serum and lavage fluid were analyzed using commercially available kits (enzyme immunoassay and immunoprecipitation analysis, respectively), controls, and standards from Microgenics Corporation (Concord, CA) and INCSTAR Corporation (Stillwater, MN).

Non-Heme Iron Concentrations in the Serum, Lavage Fluid, and Tissues Concentrations of non-heme iron in serum, lavage fluid, and lung and liver tissues were quantified by inductively coupled plasma optical emission spectroscopy (ICPOES, Model Optima 4,300 DV; Perkin Elmer, Norwalk, CT) operated at a wavelength of 238.204 nm. To 1.0 ml of serum and lavage fluid, 1.0 ml 6 N HCl/20% trichloroacetic acid was added and the specimen hydrolyzed by heating to 708C for 18 hours. After centrifugation at 20,000 3 g for 10 minutes, iron concentrations in the supernatant were measured. In lung and liver tissue, non-heme iron was determined at constant volume by the addition of 10.0 ml 3 N HCl/ 10% trichloroacetic acid/g tissue and heating to 708C for 18 hours (7). After centrifugation at 10,000 3 g for 10 minutes, the supernatant was recovered and the iron concentrations were measured. Single element standards were employed. Using this method, the variability in the measurement of non-heme iron concentration was less than 5%.

Western Blot Analysis for DMT1 BEAS-2B cells were lysed with buffer containing 1% NP40, 0.5% deoxycholate, and 0.1% SDS and protease inhibitors (Cocktail Set III; Calbiochem, La Jolla, CA), and sheared through a 22-gauge needle. Protein content was determined using the Bradford assay (Bio-Rad, Hercules, CA). The remaining sample was mixed with an equal volume of 43 loading buffer (0.5 M Tris-HCl, pH 6.8, 10% glycerol, 2% SDS, 0.7 mM b-mercaptoethanol, 0.05% bromphenol blue). Protein samples (50 mg) were separated by electrophoresis on a 4 to 15% SDS acrylamide gel and transferred to nitrocellulose membranes (Bio-Rad), blocked with 3% nonfat milk in PBS, and incubated with an antibody against either DMT1 or metal transport protein 1 (MTP1). The membrane was stained with a horseradish peroxidase–conjugated goat anti-rabbit IgG antibody (1:1,000; Santa Cruz Biotechnology, Santa Cruz, CA) and developed using enhanced chemiluminescence (ECL kit; Amersham Pharmacia Biotech, Piscataway, NJ).

Oxidative Stress in BEAS-2B Cells and Proinflammatory Mediator Release Cells were exposed to either filtered air for 24 hours, or to CO for 2 hours, then removed to filtered air for 22 hours, or to CO for 24 hours. Supernatant was analyzed by ELISA (R&D Systems) for the release of interleukin (IL)-8, a proinflammatory mediator. The cells were washed with HBSS, scrapped into 1.0 ml 2,4-dinitrophenylhydrazine (DNPH) solution (0.125% in acetonitrile), and vortexed. NaCl (0.9%) added to DNPH was used as a blank. Acetaldehyde concentrations were quantified as an index of oxidative stress (9).

Cell Proliferation Cells were exposed to filtered air for 24 hours, 50 ppm CO for 2 hours with removal to filtered air for 22 hours, or 50 ppm CO for 24 hours. Cell proliferation was measured using a modification of an ELISA kit (Roche Diagnostics Corp., Indianapolis, IN) requiring the final 2 hours of 24-hour experiments. Optical densities were read on a Molecular Devices Spectra Max 340PC plate reader (Molecular Devices, Sunnyvale, CA) at 450 and 690 nm. Reported values are averages of A450 to A690 for 6 wells with each set of conditions.

Quantitative RT-PCR for cdk2, cdk4, and p21CIP1/WAF1

BEAS-2B cells, immortalized normal human bronchial epithelium derived by transfection of primary cells with SV40 early-region genes, were used throughout the study. Cells were plated on collagen-coated filters with a 0.4-mm pore size (Trans-CLR; Costar, Cambridge, MA) at a density of 1 3 105 cells/filter and inserted into 12-well culture plates. Keratinocyte growth media (KGM; Clonetics, San Diego, CA) was added to the basolateral chamber, allowing an air–liquid interface to occur in the apical chamber. Fresh medium was provided every 48 hours. BEAS-2B cells were exposed to filtered air or CO using in vitro exposure chambers (8). Each gas was provided at 20 L/minute, balanced with 5% CO2, and kept at 88% relative humidity. Cells were exposed to filtered air or CO (500 ppm) for periods of 2 and 24 hours. Cells were also exposed to varying concentrations of CO (0, 10, 50, 100, and 500 ppm) for 24 hours.

Real-time, quantitative RT-PCR was performed to quantify expression of cdk2, cdk4, and p21CIP1/WAF1 (p21). Cells were washed twice with PBS (Life Technologies, Grand Island, NY) and lysed with 4 M guanidine thiocyanate (Boehringer Mannheim, Indianapolis, IN), 50 mM sodium citrate, 0.5% sarkosyl, and 0.01 M dithiothreitol. The cells were dislodged from wells with scrapers (Costar), lysates were sheared with four passes through a 22-gauge syringe, and total RNA isolated and 100 ng reverse transcribed (Maloney’s murine leukemia virus reverse transcriptase; Life Technologies). Quantitative fluorogenic amplification of cDNA was performed using an ABI Prism 7700v sequence detector (Applied Biosystems, Foster City, CA). cdk2, cdk4, and p21 were quantified using SYBR Green Universal PCR Master Mix (Applied Biosystems) and oligonucleotide primer pairs designed using a primer design program (Primer Express; Applied Biosystems). The mRNA levels were measured using Taqman Universal PCR Master Mix (Applied Biosystems), with primer/probe sets obtained as preoptimized mixes (‘‘assays-on demand’’ from Applied Biosystems) and normalized to b-actin expression as a reference gene using the same detection reagent as the gene of interest.

Non-Heme Iron Concentrations in Cell Lysates

Western Blotting for p21

At the end of the exposure, the BEAS-2B cells were washed with Hanks’ buffered salt solution (HBSS), scraped into 1.0 ml 3 N HCl/10% trichloroacetic acid, and digested overnight at 708C. This was centrifuged and iron concentrations were quantified in the supernatant.

Cells were extracted with RIPA buffer consisting of 10 mM Tris-HCl (pH 7.4) containing 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 20 mM Na2P2O7, 2 mM Na3VO4, 1% Triton, 10% glycerin, 0.5% deoxycholate, 0.1%SDS, phosphatase inhibitor cocktail sets I and II and protease inhibitor cocktail set III purchased (Calbiochem). Each sample was normalized for a protein content of 60 mg and then mixed with one volume of SDS-PAGE loading buffer containing 0.125 M Tris (pH 6.8), 4% SDS, 20% glycerol,10% b-mercaptoethanol, and 0.05% bromophenol blue. The samples were heated for 5 minutes at 958C and run on adjacent lanes of 12% SDS-PAGE with pre-stained molecular

Cultured Cell Exposures

Ferritin and Transferrin Concentrations in Cell Lysates Cells were washed with HBSS, scraped into 0.5 ml HBSS, and disrupted using a 25-gauge needle. Ferritin and transferrin protein concentrations in the lysates were measured using commercial kits from Microgenics Corporation and INCSTAR Corporation, respectively.

Ghio, Stonehuerner, Dailey, et al.: Carbon Monoxide and Iron

weight markers (Bio-Rad) in Tris-glycine-SDS buffer (Bio-Rad). Electrophoresed proteins were electroblotted onto nitrocellulose membranes (Bio-Rad). The blots were blocked with 5% nonfat milk, washed briefly, and incubated overnight with a mouse antibody against human p21(Calbiochem) at 48C followed by incubating with a horseradish peroxidase–conjugated goat anti-mouse antibody (Santa Cruz Biotechnology) for 1 hour at room temperature. Protein bands on the membrane were detected by chemiluminescence with reagents (Amersham Biosciences) used according to the manufacturer’s instructions and high-performance film (Denville Scientific, Metuchen, NJ).

Statistics Data are expressed as mean 6 SE. Differences between multiple groups were compared using one-way ANOVA. The post hoc test was Scheffe’s test, and two-tailed tests of significance were employed. Significance was assumed at P , 0.05.

717

Iron status in the lower respiratory tract was altered significantly by exposure of rats to 50 ppm CO for 24 hours. Lavage iron concentrations increased more than 4-fold (Figure 2A). Levels of the metal storage protein ferritin in lavage also increased significantly (Figure 2B); however, lavage transferrin concentrations did not change. In contrast to lavage iron, serum iron concentration decreased after CO inhalation (695 6 69 and 552 6 22 mg/dl for filtered air and CO, respectively). No significant differences in either serum ferritin or transferrin were observed after CO exposure. Fresh lung tissue was used for the analysis of non-heme iron, which was found to be significantly decreased after CO inhalation (Figure 2C). In comparison, the non-heme iron concentration of liver tissue was increased (Figure 2D). Cell Exposures to CO

RESULTS Animal Exposures to CO

The rats exposed to 50 ppm CO for 24 hours significantly increased their COHb concentrations to 6.9 6 0.6% (relative to 0.6 6 0.2% after filtered air). CO inhalation was associated with mild neutrophilic influx and slight elevations of lavage protein and LDH. Lavage neutrophils increased significantly after CO exposures, but were still less than 10% of total cells (Figure 1A). CO exposure increased the lavage concentrations of the proinflammatory mediator MIP-2 (Figure 1B). There were small but significant elevations of both total protein (Figure 1C), and LDH (Figure 1D) in the lavage fluid of rats exposed to CO for 24 hours. However, lavage albumin concentrations were not significantly increased by CO inhalation (39.3 6 5.7 micrograms/ml) compared with filtered air (35.8 6 4.8 micrograms/ml).

After 24 hours of exposure to as much as 500 ppm CO, there was no change in LDH release or cellular trypan blue exclusion by respiratory epithelial cells, indicating a lack of cytotoxicity in the exposures. However, CO caused BEAS-2B cells to lose nonheme iron, in amounts that depended on both time and dose (Figures 3A and 3B). After CO, cell non-heme iron concentrations could be diminished to approximately 20% of initial values. BEAS-2B cells exposed to 50 ppm CO for 2 hours showed non-heme iron values approximating 50% of those of cells in incubated in filtered air; however, removal of the cells from CO to filtered air at 2 hours for another 22 hours corrected the metal content (Figure 3C). This demonstrates reversibility of the changes in iron homeostasis associated with CO exposure. Comparable to non-heme iron concentrations, cell levels of ferritin decreased with 24 hours of exposure to CO (Figure 4A). Ferritin concentrations were increased in the media of the

Figure 1. Carbon monoxide (CO) exposure is associated with mild neutrophil accumulation in the rat lung in vivo. Rats were exposed to either filtered air or 50 ppm CO for 24 hours. Lung lavage was performed with normal saline and aliquots used for cell counting and after centrifugation to measure total protein, lactate dehydrogenase (LDH), and albumin. CO exposure was associated with neutrophil influx (A) and elevation of the pro-inflammatory MIP-2 mediator (B). The inflammation was minimal but small significant increases were found in both total protein (C ) and LDH (D). *Significantly different relative to air control; P , 0.05.

718


Figure 2. Iron status is altered in vivo after CO exposure. After 24 hours of exposure to either filtered air or 50 ppm CO, rats underwent lung lavage as well as resection of nonlavaged lung and liver tissue. Lavage iron (A) and ferritin (B) concentrations were significantly increased after CO exposure. While non-heme iron concentrations were diminished in the lung tissue (C ), those in liver tissue (D) were increased. *Significantly different relative to air exposure; P , 0.05.

basolateral compartment after exposure to 50 ppm CO for 24 hours, consistent with release of this storage protein (Figure 4B). Cell concentrations of the transferrin transport protein showed no significant change with CO exposure. Because respiratory epithelial cells cycle iron, a decrease in metal concentration can reflect changes in uptake, release, or both (10). Therefore, BEAS-2B cells were removed from media and placed in buffer at known concentrations of iron without transferrin. When these cells were exposed to CO, iron uptake decreased relative to control cells in filtered air. After 24 hours

of exposure to 50 ppm CO, cells demonstrated less iron uptake compared with those in filtered air (Figure 5A). One protein that regulates intracellular iron importation in BEAS-2B cells is DMT1 (11). Western blots for DMT1 revealed increased protein expression after 24 hours of exposure to 50 ppm CO (Figure 5B). Removal of the cells from the CO to filtered air after 2 hours actually led to an increased uptake of iron at 24 hours (Figure 5C). To quantify iron release, BEAS-2B cells were removed from media and placed in HBSS without iron. Cells incubated in

Figure 3. CO exposures decrease non-heme iron concentrations in BEAS-2B cells in vitro. (A) Time-dependent decrements in cell non-heme iron after exposure to 500 ppm CO demonstrate a very rapid effect of CO. (B) Dose-related decreases in cell non-heme iron at 24 hours. (C ) After 2 hours of exposure to 50 ppm CO, removing the cells to air restored non-heme iron concentrations at 24 hours. *Significantly different relative to air exposure; P , 0.05.


719

Figure 4. Ferritin concentrations decrease in BEAS-2B cells after CO. (A) Using an ELISA assay to measure the iron-storage protein ferritin, significant decrements were found in total ferritin in cells exposed for 24 hours to 50, 100, and 500 ppm CO relative to filtered air. (B) Significantly higher ferritin concentrations were present in the media after exposure to 50 ppm CO for 24 hours relative to filtered air.

filtered air showed slight iron release into the apical chamber after 2 hours (Figure 6A). However, cells exposed to 50 ppm CO for 2 hours released a large amount of iron into the apical chamber (Figure 6A). Iron release by respiratory epithelial cells is not well understood but, MTP1 (ferroportin 1) is involved (12). Western blots for MTP1 showed no differences in protein expression after 24 hours of exposure to 50 ppm CO (Figure 6B). The relationship between biological responses to CO and altered iron homeostasis was evaluated with assays for oxidative stress, mediator release, and cell proliferation. Using aldehyde formation as an index of oxidative stress, we found that 50 ppm CO for 24 hours decreased acetaldehyde in cell extracts relative to cells incubated in filtered air (Figure 7A). In the same BEAS2B cells exposed to 50 ppm CO for 24 hours, the levels of IL-8 were diminished in the cell supernatant (Figure 7B). Cell proliferation was evaluated by following 5-bromo-29deoxyuridine (BrdU) incorporation in BEAS-2B cells exposed to either 50 ppm CO or filtered air. CO exposure decreased cell proliferation compared with cells incubated in filtered air (Figure 7C). Removing the cells from CO after 2 hours to filtered air

eliminated the differences in oxidative stress, supernatant IL-8 concentrations, and cell proliferation at 24 hours (Figures 7A, 7B, and 7C). This functional inhibition after CO exposure corresponded to decrements in cell iron that corrected at 24 hours. An important mechanism of cell cycle arrest is inhibition of cyclin-dependent kinases (cdks). In mammalian cells, cyclin– CDK complexes are negatively regulated by CDK inhibitors, including p21. Levels of cdk2, cdk4, and p21 mRNA after air and CO exposures and iron depletion with 50 mM deferoxamine were measured using RT-PCR. We found small but significant decrements in RNA for cdk2 and cdk4 after 4 hours of exposure to 50 ppm CO (Figures 8A and 8B). These changes in cdk2 and cdk4 were similar to those following 4 h incubation of the cells in 50 mM deferoxamine to effect iron depletion (Figures 8A and 8B). RNA for p21 also decreased after 4 hours of exposure to CO and iron depletion using 50 mM deferoxamine (Figure 8C). Finally, Western blotting confirmed loss of protein expression for p21 after both CO exposure and 50 mM deferoxamine for 24 hours (Figure 9A). After 4 hours, CO and deferoxamine exposure and 20 hours air, there were no differences in p21

Figure 5. CO reversibly blocks iron uptake by BEAS-2B cells. Cells were exposed to either filtered air or 50 ppm CO for 20 hours, then 50 ml of either HBSS or 500 mM ferric ammonium citrate (FAC) was added to the apical chamber and the cells returned to CO or filtered air for another 4 hours (total 24 h). The cells were then washed with HBSS, and cell non-heme iron measured. (A) Control cells (filtered air) showed significantly more cell iron after incubation with FAC while those incubated with 50 ppm CO did not. (B) Western blot for DMT1, an iron transporter, demonstrated an increase in protein expression after 24 hours of CO exposure at 50 ppm. (C ) When BEAS-2B cells were exposed to 50 ppm CO for 2 hours, incubated in filtered air for another 18 hours, and iron added apically for 4 hours, cell non-heme iron increased to values even significantly greater than those of the cells exposed to filtered air (*significantly different relative to air exposure with no iron; **significantly different relative to air exposure; P , 0.05).

720

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 38 2008 Figure 6. CO increases iron release from respiratory epithelial cells. Media were removed from the BEAS-2B cells via the basolateral chamber and replaced with 1.0 ml HBSS. Cells were then exposed to either filtered air or 50 ppm CO. After 2 hours, the exposure was stopped and 1.0 ml HBSS was added to the apical chamber, the cells rinsed, and HBSS removed. (A) Iron in the apical chamber increased after CO compared with filtered air exposure. (B) Western blotting indicated no change in expression of the MTP1 iron exporter (ferroportin 1). Bands other than that at 90 kD are commonly present with MTP1 blots and may reflect differences in glycosylation or other posttranslational modifications (*significantly different relative to air exposure; P , 0.05).

expression (Figure 9B). This indicates the reversibility of the biological effect by CO corresponding to the change in iron homeostasis.

DISCUSSION CO is a respiratory poison at one extreme and an endogenous regulatory molecule at the other largely through its binding effects on ferrous heme. In this study, low-level CO exposure altered lung iron homeostasis both in vivo in rats and in vitro in respiratory epithelial cells grown at air–liquid interface. In the rat, iron was mobilized from the lung accompanied by decrements in non-heme concentration. Correspondingly, elevations in non-heme iron concentrations were found in the alveolar and airway compartments by bronchoalveolar lavage. A portion of this increased iron in the lavage fluid was associated with the storage protein ferritin because extracellular ferritin concentrations were also elevated after CO exposure. Levels of non-heme

iron changed in the opposite direction in the liver from those in the lungs, implying that iron homeostasis is regulated systemically during the CO exposure. Although serum iron and ferritin concentrations were not measurably affected by CO in this study, this finding probably reflects late timing and unitary sampling. In mammals, the liver is the preferred site for storing iron in a controlled state to minimize the catalytic capacity of the metal. Thus, physiologically, iron moved from the lungs, where the CO exposure was most intense, to the liver, its main storage site. Exposure of BEAS-2B respiratory epithelial cells to CO led to a rapid, significant, decrease in non-heme iron concentrations in cells grown at an air–liquid interface. Cells exposed to CO at 50 ppm for only 2 hours lost almost half of their non-heme iron compared with those exposed to filtered air. This loss of iron could result from an increased egress, a decreased uptake, or both. It has been proposed that CO can alter iron homeostasis by coordination of Fe21 and release of metal from transferrin in the endosome (13); this would subsequently inhibit metal entry into

Figure 7. CO decreases acetaldehyde production, IL 8 release, and cell proliferation in BEAS-2B cells in vitro. Cells were scraped into 2,4-dinitrophenylhydrazine after 24 hours of CO exposure at 50 ppm for determination of acetaldehyde by GC-HPLC. (A) Acetaldehyde in the cell extract was decreased by CO exposure. (B) IL-8 concentration in the media (measured by ELISA) was decreased after 24 hours of exposure to 50 ppm CO. (C ) Cell proliferation by BrdU incorporation was diminished in cells exposed to 50 ppm CO. Removing the cells after 2 hours of CO exposure to filtered air (CO/Air) eliminated significant differences in oxidative stress, supernatant IL-8 concentrations, and cell proliferation (*significantly different relative to air exposure; P , 0.05).


721

Figure 8. Cell proliferation indices after exposure of BEAS-2B cells to CO or deferoxamine in vitro. After 4 hours of exposure to air, 50 ppm CO, and 50 mM deferoxamine, cell extracts were collected for either RT-PCR or Western blot analysis. Relative to filtered air, respiratory epithelial cells exposed to 50 ppm CO for 4 hours demonstrated decreased mRNA for cdk2, cdk4, and p21 (A, B, and C, respectively). This was comparable to the effect of iron depletion of the cells by deferrioxamine (*significantly different relative to air control; P , 0.05).

the cell. Our data in BEAS-2B cells showed significant decrements in iron uptake during CO exposure in agreement with that earlier study (13). However, removal of the BEAS-2B cells from CO revealed that its effect on intracellular iron was rapidly reversible. Moreover, the DMT1 protein, which participates in the intracellular uptake of Fe21, actually increased with CO

Figure 9. Western blot analysis for p21 after exposure of BEAS-2B cells to CO or deferoxamine in vitro. (A) Protein expression of p21 was decreased by 24-hour exposures to CO (50 ppm) or deferoxamine (DEF, 50 mM). (B) The differences in p21 expression after a 4-hour exposure to CO or deferoxamine were abolished after another 20 hours of exposure to filtered air. The band for p21 is observed between molecular weight markers 20 and 25 kD.

exposure—possibly reflecting a state of intracellular iron depletion. The removal of respiratory epithelial cells from CO to filtered air not only eliminated the decrements in iron uptake but significantly increased intracellular metal transport over 24 hours. This increased cellular iron transport is consistent with the increased expression of DMT1 and perhaps other transporters. The cellular iron egress after CO exposure found here contrasts with an earlier investigation that detected no significant effect of CO on iron release (14). In filtered air, human bronchial epithelial cells normally lose a small amount of iron, which (as in the previous investigation) reflects continuous iron uptake and release by these cells (12). Although CO greatly accelerated the iron loss, expression of the major protein responsible for iron release in these cells, MTP1, was unchanged by the CO exposures. Consequently, the cause of decreased uptake and increased release of iron after CO exposure in respiratory epithelia likely involves another mechanism(s). For instance, after CO exposure, ferritin concentration in the media was elevated, consistent with its release by the airway cells. Investigation in other systems suggests that such ferritin release may be a defense against high intracellular iron concentrations (15). An important factor in the cellular iron depletion by CO could be that the iron carbonyl has limited utility to the cell, in which case, the cell may be forced to release the iron-carbonyl nonspecifically. Alternatively, if CO increases heme release (e.g., from mitochondrial sources) and stimulates HO activity in the cell, cytosolic iron concentration will rise unless the cell exports it (16). Under such circumstances, a reduction in the uptake and/or increase in the release of the metal would reduce iron and oxidative stress in the cell. Thus, CO could signal the need to adjust iron homeostasis to avoid iron accumulation and oxidative stress. Exposure to CO for 24 hours decreased acetaldehyde levels in respiratory epithelial cells consistent with such an antioxidant

722


effect. By coordinating iron, CO can oppose oxidative stress by decreasing or abolishing iron’s potential to support electron transfer to active intermediates, especially reactive oxygen species. In addition, the loss of cellular non-heme iron after CO would block oxidant generation by Fenton-based reactions. These results are consistent with previous studies showing that CO can protect against oxidative stress (17–19); however, removing the cells from CO after 2 hours returned non-heme iron concentrations and acetaldehyde production to control levels within 24 hours. CO-dependent changes in non-heme iron concentration also corresponded to lower production of the proinflammatory mediator IL-8. Once again, removing the epithelial cells from CO after 2 hours returned IL-8 levels to those of control cells in filtered air. This could be interpreted as an anti-inflammatory effect, similar to that reported for HO in various models of inflammatory injury including hyperoxia, sepsis, ischemia, and so on (20–22); however, dosing and timing of CO administration is a variable that could confound the interpretation of such effects (23, 24). The data also demonstrate that changes in iron homeostasis by CO are congruent with apparent antiproliferative effects of CO reported in other cell types (25). Adequate cellular iron is vital for cell cycle progression, and the evidence so far indicates that most of this iron is required for DNA synthesis (26), especially during late G1 and S phase. It is also now known that iron depletion results in a compound response that affects ribonucleotide reductase and multiple cell cycle control points (27). Low iron concentrations block cellular proliferation by impinging on pathways that control cell division (e.g., levels of cyclin D, cyclin-dependent kinases) (28). Iron chelation arrests cell cycle progression in late G1, before the G1/S border. Similarly, CO inhibited cell proliferation in accordance with its capacity to deplete iron. Because the cyclins coordinate cell proliferation and their activities are modulated by both cdks and cdk inhibitors, we checked mRNA levels for cdk2 and cdk4 after CO and found them to be decreased relative to air controls. The response was comparable to that of the BEAS-2B cells after iron depletion with deferoxamine, both here and in other investigations (26, 29). Similarly, the mRNA for p21 was decreased by CO exposure. The effect of iron depletion on p21 mRNA appears to be cell specific: it increases in some (30) and decreases in other cell types (31). In BEAS-2B cells, the effect of CO was identical to that of deferoxamine (both decreased the mRNA equally) and fits with the known transcriptional and post-transcriptional influence of iron on p21 protein expression (27). Diminished p21 protein expression after CO or deferoxamine exposure was confirmed by Western blot analysis, and for both protein and message, the effect of CO on BEAS-2B cells was equivalent to that of deferoxamine and consistent with the influence of iron homeostasis on the cell cycle. The results of this study also reinforce the idea that physiological CO and NO share certain cell functions (e.g., as reported for guanylate cyclase)—as both molecules demonstrate common coordination chemistry, a high affinity for iron (6), and now changes in iron homeostasis. However, NO reacts with both Fe (III) and Fe (II), while CO binds only the latter. Comparable to CO, NO also mobilizes cell metal and inhibits iron uptake from transferrin and its incorporation into ferritin (32, 33). Although our findings contrast with those of a previous study (14) by demonstrating significant iron efflux from BEAS-2B cells after CO exposure, iron loss is consistent with observations that cells transfected with HO show less iron uptake from transferrin and elevated iron efflux and that HO knockout mice show not only iron deficiency anemia but iron accumulation in liver and kidneys (4). Differences in iron efflux rates among studies

may relate to the cell type, HO expression, culture conditions, and so on, but exposure of the lungs or airway epithelial cells to CO occurs without cell hypoxia, and the use of an air–liquid interface is relevant to pulmonary gas exchange. That CO clearly alters lung iron homeostasis possibly reflects interactions with the control of the oxidative potential of the metal by the host. CO is produced endogenously by heme protein degradation (e.g., hemoglobin, cytochromes, etc.) by the heme oxygenases. To limit the propensity of free iron to generate oxidative stress, the cell can increase the release of the iron and decrease its uptake, and our data support the idea that CO, as a product of enzymatic heme breakdown, participates in just such a response. Moreover, the loss of cellular iron is rapidly reversible and reflected in a reversible change in the growth characteristics of respiratory epithelial cells, which could have positive or negative effects on the lung. In conclusion, exogenous CO alters iron homeostasis in the intact lung and in lung epithelial cells by enhancing net iron release. Some of the biological effects reported for CO correspond to these effects on cellular iron status and are comparable to those of the potent metal chelator, deferrioxamine. Analogous to use of chelators, CO as a deliberate intervention will be complicated by a lack of focused delivery to specific cell targets and a lack of capacity of the gas to distinguish between excess/ inappropriate iron which catalyzes harmful oxidative stress and metal required for homeostasis. Perhaps the most interesting new finding reported here, however, is the rapidly reversible but poorly understood effect of CO on iron handling by human respiratory epithelial cells that tracks with changes in cell cycle protein expression. Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript.

References 1. Schapira G. Iron metabolism. Berlin: Springer-Verlag; 1964. pp. 1–8. 2. Moura E, Noordermeer MA, Verhoeven N, Verheul AF, Marx JJ. Iron release from human monocytes after erythrophagocytosis in vitro: an investigation in normal subjects and hereditary hemochromatosis patients. Blood 1998;92:2511–2519. 3. Ferris CD, Jaffrey SR, Sawa A, Takahashi M, Brady SD, Barrow RK, Tysoe SA, Wolosker H, Baranano DE, Dore S, et al. Haem oxygenase-1 prevents cell death by regulating cellular iron. Nat Cell Biol 1999;1: 152–157. 4. Poss KD, Tonegawa S. Heme oxygenase 1 is required for mammalian iron reutilization. Proc Natl Acad Sci USA 1997;94:10919–10924. 5. Yachie A, Niida Y, Wada T, Igarashi N, Kaneda H, Toma T, Ohta K, Kasahara Y, Koizumi S. Oxidative stress causes enhanced endothelial cell injury in human heme oxygenase-1 deficiency. J Clin Invest 1999; 103:129–135. 6. Watts RN, Ponka P, Richardson DR. Effects of nitrogen monoxide and carbon monoxide on molecular and cellular iron metabolism: mirrorimage effector molecules that target iron. Biochem J 2003;369: 429–440. 7. Torrance JD, Bothwell TH. Tissue iron stores. In: Cook JD, editor. Iron. New York: Churchill Livingstone; 1980. pp. 90–115. 8. Ray BM. Performance auditing of an in vitro air pollution exposure system. Inhal Toxicol 1998;10:615–624. 9. Madden MC, Thomas MJ, Ghio AJ. Acetaldehyde (CH3CHO) production in rodent lung after exposure to metal-rich particles. Free Radic Biol Med 1999;26:1569–1577. 10. Turi JL, Yang F, Garrick MD, Piantadosi CA, Ghio AJ. The iron cycle and oxidative stress in the lung. Free Radic Biol Med 2004;36:850–857. 11. Wang X, Ghio AJ, Yang F, Dolan KG, Garrick MD, Piantadosi CA. Iron uptake and Nramp-2/DMT1/DCT1 in human bronchial epithelial cells. Am J Physiol Lung Cell Mol Physiol 2002;282:L987–L995. 12. Yang F, Wang X, Haile DJ, Piantadosi CA, Ghio AJ. Iron increases expression of iron-export protein MTP1 in lung cells. Am J Physiol Lung Cell Mol Physiol 2002;283:L932–L939.

Ghio, Stonehuerner, Dailey, et al.: Carbon Monoxide and Iron 13. Richardson DR, Ponka P. The molecular mechanisms of the metabolism and transport of iron in normal and neoplastic cells. Biochim Biophys Acta 1997;1331:1–40. 14. Watts RN, Richardson DR. Differential effects on cellular iron metabolism of the physiologically relevant diatomic effector molecules, NO and CO, that bind iron. Biochim Biophys Acta 2004;1692: 1–15. 15. Geiser DL, Zhang D, Winzerling JJ. Secreted ferritin: mosquito defense against iron overload? Insect Biochem Mol Biol 2006;36:177–187. 16. Cronje FJ, Carraway MS, Freiberger JJ, Suliman HB, Piantadosi CA. Carbon monoxide actuates O(2)-limited heme degradation in the rat brain. Free Radic Biol Med 2004;37:1802–1812. 17. Otterbein LE, Otterbein SL, Ifedigbo E, Liu F, Morse DE, Fearns C, Ulevitch RJ, Knickelbein R, Flavell RA, Choi AM. MKK3 mitogenactivated protein kinase pathway mediates carbon monoxide-induced protection against oxidant-induced lung injury. Am J Pathol 2003;163: 2555–2563. 18. Otterbein LE, Zuckerbraun BS, Haga M, Liu F, Song R, Usheva A, Stachulak C, Bodyak N, Smith RN, Csizmadia E, et al. Carbon monoxide suppresses arteriosclerotic lesions associated with chronic graft rejection and with balloon injury. Nat Med 2003;9:183–190. 19. Morse D, Pischke SE, Zhou Z, Davis RR, Flavell RA, Loop T, Otterbein SL, Otterbein LE, Choi AMK. Suppression of inflammatory cytokine production by carbon monoxide involves the JNK pathway and AP-1. J Biol Chem 2003;278:36993–36998. 20. Lee PJ, Alam J, Wiegand GW, Choi AM. Overexpression of heme oxygenase-1 in human pulmonary epithelial cells results in cell growth arrest and increased resistance to hyperoxia. Proc Natl Acad Sci USA 1996;93:10393–10398. 21. Otterbein L, Sylvester SL, Choi AM. Hemoglobin provides protection against lethal endotoxemia in rats: the role of heme oxygenase-1. Am J Respir Cell Mol Biol 1995;13:595–601. 22. Li L, Hamilton RF Jr, Holian A. Protection against ozone-induced pulmonary inflammation and cell death by endotoxin pretreatment in mice: role of HO-1. Inhal Toxicol 2000;12:1225–1238. 23. Clayton CE, Carraway MS, Suliman HB, Thalmann ED, Thalmann KN, Schmechel DE, Piantadosi CA. Inhaled carbon monoxide and hyperoxic lung injury in rats. Am J Physiol Lung Cell Mol Physiol 2001;281: L949–L957.

723 24. Otterbein LE, Kolls JK, Mantell LL, Cook JL, Alam J, Choi AMK. Exogenous administration of heme oxygenase-1 by gene transfer provides protection against hyperoxia-induced lung injury. J Clin Invest 1999;103:1047–1054. 25. Song R, Mahidhara RS, Zhou Z, Hoffman RA, Seol DW, Flavell RA, Billiar TR, Otterbein LE, Choi AM. Carbon monoxide inhibits T lymphocyte proliferation via caspase-dependent pathway. J Immunol 2004;172:1220–1226. 26. Kulp KS, Green SL, Vulliet PR. Iron deprivation inhibits cyclin-dependent kinase activity and decreases cyclin D/CDK4 protein levels in asynchronous MDA-MB-453 human breast cancer cells. Exp Cell Res 1996;229:60–68. 27. Le NTV, Richardson DR. Potent iron chelators increase the mRNA levels of the universal cyclin-dependent kinase inhibitor p21C1P1/ WAF1, but paradoxically inhibit its translation: a potential mechanism of cell cycle dysregulation. Carcinogenesis 2003;24:1045–1058. 28. Lucas JJ, Szepesi A, Domenico J, Takase K, Tordai A, Terada N, Gelfand EW. Effects of iron-depletion on cell cycle progression in normal human T lymphocytes: selective inhibition of the appearance of the cyclin A-associated component of the p33cdk2 kinase. Blood 1995;86:2268–2280. 29. Wang G, Miskimins R, Miskimins WK. Regulation of p27(Kip1) by intracellular iron levels. Biometals 2004;17:15–24. 30. Kim H-J, Hida H, Jung C-G, Miura Y, Nishino H. Treatment with deferoxamine increases neurons from neural stem/progenitor cells. Brain Res 2006;1092:1–15. 31. Gazitt Y, Alcantara RO, Yang J, Boldt DH. A new molecular role for iron in regulation of cell cycling and differentiation of HL-60 human leukemia cells: iron is required for transcription of p21(WAF1/CIP1) in cells induced by phorbol myristate acetate. J Cell Physiol 2001;187: 124–135. 32. Watts RN, Richardson DR. Nitrogen monoxide (no) and glucose: unexpected links between energy metabolism and no-mediated iron mobilization from cells. J Biol Chem 2001;276:4724–4732. 33. Watts RN, Richardson DR. The mechanism of nitrogen monoxide (NO)-mediated iron mobilization from cells. NO intercepts iron before incorporation into ferritin and indirectly mobilizes iron from ferritin in a glutathione-dependent manner. Eur J Biochem 2002;269: 3383–3392.