Published ahead of print on June 7, 2004, doi:10.1164/rccm.200311-1607OC
© 2004 American Thoracic Society
Heme Oxygenase-1 Expression in Human Lungs with Cystic Fibrosis and Cytoprotective Effects against Pseudomonas Aeruginosa In VitroDepartments of Pediatrics, University of Florida, Gainesville, Florida; and Department of Pathology, University of Texas Medical School at Houston, Houston, Texas Correspondence and requests for reprints should be addressed to Gary A. Visner, D.O., University of Florida, Department of Pediatrics, P.O. Box 100296, Gainesville, FL 32610. E-mail: visnega{at}peds.ufl.edu
Inflammation and oxidative stress play important roles in cystic fibrosis (CF) lung disease. Inflammatory/oxidant-mediated induction of heme oxygenase-1 (HO-1) is believed to be a cytoprotective response. This study examined HO-1 expression in lung samples from patients with CF using immunohistochemistry and quantitative reverse transcription-polymerase chain reaction. In addition, we evaluated myeloperoxidase staining as a marker of acute inflammation and potentially an increase in oxidant stress and Prussian blue and ferritin staining to assess iron status of the lung. Macrophage HO-1 staining was increased in diseased lungs as compared with normal control subjects and correlated with myeloperoxidase staining. Quantitative reverse transcription-polymerase chain reaction further supported an increase in HO-1 expression in CF lung disease. Although iron staining was minimal, ferritin staining was increased in diseased lungs in concert with HO-1 staining. To determine whether HO-1 induction was cytoprotective, we evaluated a CF airway epithelial cell line, IB3.1, in response to Pseudomonas aeruginosa–induced injury/apoptosis in cells overexpressing HO-1 by either transient or stable transfection of pcDNA3.1/HO-1 construct. Overexpression of HO-1 resulted in protection against P. aeruginosa–induced injury/apoptosis. This suggests that the induction of HO-1 in patients with CF is a cytoprotective event and that augmenting its expression is a potential therapy against bacterial injury.
Key Words: chronic airway disease • heme oxygenase-1 • inflammation • oxidative stress • Pseudomonas aeruginosa Cystic fibrosis (CF) is an autosomal recessive disorder that is caused by mutations of the CF transmembrane regulator gene. The defect leads to a dysfunction in epithelial chloride and water transport, increased mucus viscosity, decreased clearance of secretions, and airway defense mechanism resulting in chronic lung infection (1). Consequently, inflammatory cells are recruited in response to the infection leading to a state of chronic airway inflammation (2). Increased production of oxidants is a feature of chronic inflammatory lung diseases and is believed to contribute to the genesis of lung injury in CF (3–5). The cellular effects of oxidant stress are modulated by synthesis of proteins that have protective properties. Stress response protein-32, also known as heme oxygenase-1 (HO-1) (6), is one such protein that is induced by cytokines and oxidant stress (7), and its functions appear to ameliorate oxidative cellular injury (8). The biological importance of HO stems from its function as the initial and rate-limiting enzyme in heme catabolism, in which it catalyzes the oxidative cleavage of heme molecules to equal moles of biliverdin, free iron, and carbon monoxide (CO) (6). There are three HO isoforms: HO-1, HO-2, and HO-3. HO-2 and HO-3, which is very similar to HO-2, are constitutively expressed, whereas HO-1 is inducible in response to reactive oxygen species, endotoxin, and proinflammatory cytokines (6, 9, 10). HO-1 and HO-2 are antigenically distinct, and their respective antisera do not cross-react, whereas less is known about recently described HO-3 (11). Knowledge about the functions of HO-1 has come from investigations using animal models of inflammation, cell cultures, and genetically deficient mice and humans. Mice with a targeted disruption of HO-1 appear to exist in a state of chronic inflammation (12, 13). A well studied HO-1–deficient human demonstrated multiple abnormalities, including growth retardation, anemia, abnormal iron metabolism with increased iron deposition, and enhanced vulnerability to oxidant damage during stress (14). In a model of pleural space inflammation, inhibition of HO-1 worsened inflammation, whereas increasing HO-1 activity suppressed the inflammatory response (15). HO-1 inhibition also resulted in a loss of endotoxin-induced protection against oxidant injury in vascular smooth muscle cells (10). CF is a disease in which airway inflammation is believed to play a crucial role in the pathophysiology of airway injury (1, 2). Because HO-1 appears to mitigate against inflammatory/oxidant-mediated injury, this study was performed to assess the relationship of HO-1 to the airway injury characteristic of this disease. Cellular HO-1 expression was evaluated by immunohistochemistry in human lungs with CF. Myeloperoxidase staining was used to measure neutrophil infiltration. Because cellular iron status is influenced by HO-1 activity and iron availability contributes to the oxidative environment, Prussian blue staining for iron and immunohistochemistry for ferritin were performed to assess the abundance of these substances. Pseudomonas aeruginosa (P.a.) is a gram-negative facultative opportunistic pathogen that chronically infects airways of patients with CF, leading to bronchiectasis that eventually becomes incompatible with life. Infection of mammalian cells with P.a. has been shown to result in the internalization of P.a. and triggers apoptotic or necrotic death of the infected cells (16–19). HO-1 overexpression protects airway epithelial cells, endothelial cells, and pancreatic ß cells from injury and apoptosis (8, 20, 21). To determine whether HO-1 induction in CF lung disease is potentially cytoprotective, a CF airway epithelial cell line, IB3.1, with enhanced HO-1 expression, was evaluated for apoptosis in response to P.a. exposure.
Samples, Staining, and Morphologic Evaluation Samples included sections from pneumonectomies, wedge resections, and transbronchial biopsies of lung from patients with CF (n = 12) and normal lung tissue from lung resections for cancer (n = 6) (Table 1). Lung samples were obtained from patients with CF at a time that they were relatively stable or after antibiotic treatment. Immunohistochemical staining was performed on a Ventana (Ventana Medical Systems, Inc., Tucson, AZ) immunostainer as previously described using a polyclonal rabbit anti–HO-1 antibody (SPA-895; StressGen Biotechnologies Corp., Victoria, BC, Canada) (22), polyclonal rabbit anti-human myeloperoxidase antibody (23) (Ventana Medical Systems, Inc.), and polyclonal rabbit anti-human ferritin antibody (24) (Dako, Carpinteria, CA). Positive and negative control slides were included with each staining run.
Cell types expressing HO-1 were assessed morphologically. In addition, to confirm that cells in alveoli were macrophages and not pneumocytes, double immunohistochemical staining using anti–HO-1 and mouse anti-human CD68 (clone KP1; Dako) was performed in 10 samples. The number of HO-1 or ferritin-stained alveolar macrophages was counted in 20 high-power fields or the maximum number available for examination (if less than 20). Staining of respiratory epithelium, type II pneumocytes, endothelium, myofibroblasts, and smooth muscle cells was graded as none, focal, patchy, or diffuse, and myeloperoxidase staining was assessed as none, intravascular only (+1), and extravascular focal, patchy, or diffuse (+2 to +4).
Quantitative Reverse Transcription-Polymerase Chain Reaction Protocol
Cell Cultures, HO-1 Overexpression, and Western Analysis
Bacterial Infection Experiments and Measurement of Apoptosis
Statistical Analysis
Information about patient demographic characteristics and FEV1 is provided in Table 1. The majority of the human lung samples came from patients with advanced CF who were undergoing lung transplantation, accounting for the relatively low FEV1 values seen in many of these patients. Eight samples were obtained from native lungs with advanced CF, excised at the time of lung transplantation. Four patients with CF had wedge or transbronchial biopsies because of a progressive decline in pulmonary function despite several weeks of antibiotic therapy appropriate for bacterial culture sensitivities. All patients had positive cultures for P.a. except one, and there were no positive cultures for Burkholderia cepacia isolated in these patients. Lung biopsy in two of these patients showed organizing pneumonia, probably representing the residuum of acute bronchopneumonia, a third biopsy demonstrated Nocardia colonizing a bronchus, and the last biopsy was consistent with CF, although there were fewer inflammatory cells than expected. The last four patients' pulmonary function status was better (FEV1 > 45%) than the lung transplant recipients (FEV1 < 25%). Immunohistochemical staining for HO-1 performed on these lung samples demonstrated cytoplasmic staining of macrophages, respiratory epithelial cells, type II pneumocytes, myofibroblasts, endothelial cells, and airway smooth muscle cells (Figures 1 and 2). Staining intensity in the macrophages substantially exceeded that observed in other cell types. The number of HO-1–stained alveolar macrophages in the advanced CF group significantly exceeded that of the normal control group (p < 0.05) and was higher than that observed in the less advanced CF group. Double staining of the macrophages with KP-1 and HO-1 (macrophage marker) confirmed that nearly all macrophages stained for HO-1, and therefore, the increase in the number of HO-1–stained macrophages reflects an increase in the number of macrophages present. The more severe CF samples also showed more intense macrophage HO-1 staining (Figure 1E), suggesting that HO-1 expression in individual cells was also increased.
In addition to the macrophage results, evaluation of other cell types revealed a noticeable difference in type II pneumocytes in CF (p < 0.05) (Figure 2A). Alveolar extension of the inflammatory process is probably the trigger for the increased type II pneumocyte HO-1 expression in CF. Conversely, we observed no differences in the extensiveness of respiratory epithelial cells, smooth muscle cells, and endothelial cells HO-1 staining between the two groups. A semiquantitative RT-PCR technique was used as another measure of HO-1 expression. RNA was isolated from paraffin-embedded formalin-fixed tissue, and HO-1 mRNA and 18s rRNA relative values were determined using Taqman RT-PCR. Assay of 12 samples (five control subjects and seven samples with CF) resulted in values in which the HO-1 and 18s rRNA RT-PCR products were within the linear range of their respective standard curves, and therefore, a ratio of HO-1 to 18s rRNA could be determined. HO-1 mRNA expression was increased in the disease state (CF) as compared with normal (p < 0.05) (Figure 2B). Myeloperoxidase staining was used as a marker for neutrophil infiltration. Some weak macrophage staining was also observed, probably representing phagocytized myeloperoxidase-containing material (29). Not surprisingly, in CF, myeloperoxidase staining was heaviest in areas of severe airway inflammation (Figure 3), and in general, these regions had the most intense HO-1 staining. In control samples, there was relatively little myeloperoxidase staining (mostly +1 [intravascular only]) as compared with CF samples. Myeloperoxidase staining also correlated with HO-1 staining in macrophages (p < 0.05) (Figure 3C).
Cellular iron stores and ferritin were also evaluated. Using the Prussian blue stain, cellular iron stores appeared minimal in both diagnostic groups, with few macrophages staining for iron. Ferritin staining was mainly in macrophages (Figure 4A). The number of macrophages staining for the iron storage protein, ferritin, was increased in patients with CF as compared with normal subjects (p < 0.05) (Figure 4B), and the number of ferritin-stained macrophages correlated with the amount of HO-1 staining (Figure 4C).
To determine whether HO-1 may be a cytoprotective protein in CF lung disease, cell culture studies using airway epithelial cells derived from a patient with CF, IB3.1, were performed. To determine whether the increase in HO-1 expression was in response to bacterial exposure or from one of its products such as lipopolysaccharide (LPS), cells were exposed to either LPS or P.a. HO-1 was induced by hemin, a known stimulus of HO-1, but neither LPS nor P.a. exposure resulted in an increase in HO-1 protein levels (Figure 5A). Both LPS and P.a. increased HO-1 protein expression in RAW264.7 (mice monocyte-like cell line) (Figure 5B) as other authors reported (30).
Survival of IB3.1 cells overexpressing HO-1 was next evaluated in response to P.a. The initial studies used a transient transfection protocol using a pcDNA3.1 empty vector or pcDNA3.1/HO-1 vector cotransfected with pcDNA3.1 vector containing ß-gal as a marker. Evaluation was based on the number of ß-gal–expressing cells. Treatment with the HO-1 vector nearly doubled the ß-gal expression (63%) as compared with control cells treated with ß-gal vector only (38%) or those with ß-gal and the empty pcDNA3.1 vector (37%) (Table 2), indicating improved survival of IB3.1 cells with HO-1 overexpression.
To alleviate the problems inherent to transient transfection, stable IB3.1 cell lines containing the empty vector or pcDNA-HO-1 (HO-1/IB3.1) were generated and evaluated in response to P.a. (PAK). Figure 5A illustrates a Western analysis showing an increase in HO-1 protein levels in HO-1/IB3.1 cells. Both control and HO-1/IB3.1 cells exposed to PAK for 1 or 2 hours showed increased apoptosis/injury, as measured by annexin V staining alone for early apoptosis, annexin V plus propidium iodide for late apoptosis/necrosis, or caspase 3 activity as another marker for apoptosis. PAK exposure resulted in an increase in apoptosis/injury for both time intervals, with more cell injury after 2 hours of exposure. HO-1 overexpression was associated with a significant reduction in apoptosis/injury (Figures 5C–5F). Previous studies have demonstrated that HO-1 overexpression in endothelial cells result in a decrease in COX-2 expression/activity (31, 32). To clarify whether HO-1 antiapoptosis involved COX-2, IB3.1 cells were evaluated for COX-2 expression by Western analysis in response to P.a. and/or the COX inhibitor, indomethacin, under both basal and increased HO-1 conditions. COX-2 was induced in IB3.1 cells exposed to PAK (Figure 6A) or LPS (data not shown). Cells overexpressing HO-1 (HO-1/IB3.1) had lower COX-2 protein levels in response to PAK, similar to what was observed with the COX inhibitor indomethacin (Figure 6A). To determine whether COX inhibition by HO-1 mediates its cytoprotective effects against P.a.-dependent apoptosis, we evaluated whether the COX inhibitor indomethacin resulted in a similar protective effect against P.a.-dependent injury. Indomethacin treatment failed to show a similar protection as observed with HO-1 overexpression (Figure 6B), suggesting that the cytoprotective effects of this model are not through HO-1 inhibition of COX-2.
CF is a chronic airway disease in which inflammation and oxidative stress are critical in the pathophysiology of disease progression. In many models of inflammatory and/or oxidant-mediated injury, HO-1 induction is a cytoprotective response (6, 8, 10, 13–15, 33, 34). Therefore, we evaluated HO-1 expression in lung tissue from patients with CF and evaluated its cytoprotective role in a CF airway epithelial cell line, IB3.1. HO-1 staining was present in macrophages, respiratory epithelial cells, endothelial cells, type II pneumocytes, and myofibroblasts, and increased in macrophages and type II pneumocytes in CF samples. In general, macrophage HO-1 staining was most extensive and intense among patients with more advanced lung disease. The diverse cellular distribution of HO-1 is interesting because of its correlation with the site(s) of inflammation. Lungs from patients with CF typically show bronchopneumonia in addition to inflammation in the large airways. Parenchymal involvement by the inflammatory process is probably the trigger for type II pneumocyte HO-1 expression. Surprisingly, we did not observe an increase in other cell types, especially airway epithelial cells. This may be because we observed a relatively high amount of HO-1 staining in our control samples, which were from patients undergoing lung resection for cancer and receiving mechanical ventilation. In lungs isolated from mice or rats under normal conditions, we normally see little if any staining of airway epithelial cells, whereas mice given an intratracheal injection of agarose beads coated with P.a. show an increase in airway epithelial staining (data not shown). The increased HO-1 expression that we observed correlates well with previous investigations demonstrating elevated exhaled CO levels in patients with CF and bronchiectasis (3, 35–37). CO production is dependent on HO activity. Because HO-1 is the inducible form of HO, the elevated CO levels in these airway diseases would most likely stem from increased HO-1 activity. Our results support this conclusion. Elevated CO levels are thought to be a marker of inflammation or oxidant stress (3, 35–37), a hypothesis that is supported by the finding of increased myeloperoxidase staining in the CF lungs. The role of HO-1 in the host response to inflammation appears to be protective. HO-1 is induced by a variety of agents, including heme, cytokines, endotoxin, and agents causing oxidative stress that may also be responsible for its increase in CF lung disease. Upregulation of HO-1 has been shown to be protective against oxidant-induced kidney failure, endotoxin shock, hyperoxia-induced lung injury, LPS-induced acute lung injury (6, 10, 27, 38), and interestingly, overexpression of HO-1 protects pancreatic ß cells and endothelial cells from apoptosis (20, 21). Over 80% of individuals with CF suffer from considerable morbidity caused by chronic lung infection with P.a., and chronic P.a. infection also is the cause of death in over three-fourths of CF patients. P.a. infection has been shown to damage cells and induced apoptosis in human conjunctiva epithelial Chang cells, respiratory epithelial cells, and endothelial cells (16–18). Interestingly, P.a. exposure and LPS did not induce HO-1 in IB3.1 or the corrected CF airway epithelial cell line, S9 (data not shown), but induced HO-1 in RAW264.7 (Figure 5B). Our experiments demonstrate that HO-1 overexpression is cytoprotective against P.a.-induced apoptosis/injury in CF airway epithelial cells. COX-1 and COX-2 are heme proteins that catalyze the oxygenation and peroxidation of arachidonic acid to prostaglandin endoperoxide H2. COX-1 is constitutively expressed, whereas COX-2 is inducible. Upregulated HO-1 associated with decreased COX-2 has been shown in microvessel endothelial cells and human femoral endothelial cells (31, 32), and we found this to be true in airway epithelial cells (IB3.1 cells). Interestingly, the antiinflammatory drug, indomethacin, a COX inhibitor, is used to slow the progression of CF lung disease (39), leading to the possibility that the protective effects of HO-1 in CF being related to COX-2 inhibition. However, the COX inhibitor indomethacin did not result in a similar protection against P.a.-induced apoptosis of airway epithelial as HO-1 overexpression, suggesting that this is not the protective mechanism for epithelial cells. Although inhibition of COX-2 by HO-1 does not appear to be the mechanism of its protection in epithelial cells, this is still a potential protective antiinflammatory response of HO-1 in CF lung disease. There is evidence that CO accounts for at least some of the protective effects of HO-1 antiapoptosis/injury (21, 40). The mechanism by which CO might provide cytoprotection has not been entirely elucidated; it appears that the interaction between IL-10 and HO-1 plays a role (9, 41). IL-10 resistance to LPS-induced injury in RAW264.7 and lung injury involves HO-1 (9). However, we did not observed increased IL-10 protein level in overexpression HO-1 IB3.1 cells (data not shown), suggesting that HO-1 antiapoptosis/injury is not related to IL-10. Recently, P.a.-induced neutrophilic lung inflammation was shown to be attenuated by adenovirus-mediated transfer of the HO-1 cDNA in mice, and this protected against airway epithelial cell apoptosis by overexpressing Bcl-2 (42). In airways, neutrophils comprise the main inflammatory response to infection by P.a. Myeloperoxidase, an enzyme found in neutrophils, is an accepted marker for this cell type (29). Myeloperoxidase activity is increased in disease states characterized by neutrophilic airway inflammation, such as CF (43), and lung function has been inversely correlated with sputum myeloperoxidase levels (44). The product of myeloperoxidase, hypochlorous acid, is released by neutrophils along with other oxidants as a consequence of the respiratory burst. In CF, myeloperoxidase-derived oxidants have been shown to be harmful to epithelial cells (45). In this study, diffuse and intense myeloperoxidase staining was found in areas of active disease in lungs with CF, and more HO-1 staining, especially in regions of active inflammation, was observed in patients who had greater amounts of myeloperoxidase staining. Oxidative stress and inflammatory cytokines are stimuli for HO-1, and the increased expression of HO-1 observed in CF may be a response to these factors. A potentially beneficial effect of HO-1 activity against oxidant injury is related to its role in iron sequestration. Ferris and coworker correlated protection of cells by HO-1 with a decrease in intracellular iron and reproduced the protective effect by iron chelation (33). Heme is a ubiquitous iron-containing compound, and much of the cellular damage by activated oxygen species involves the collaboration of intracellular iron (46). Recent studies have demonstrated that HO-1 is important for cellular iron efflux and that this may be related to HO-1's cytoprotective effect against oxidant injury (32, 33). The iron released by HO activity is normally sequestered by ferritin, whose induction helps to prevent oxidant injury caused by iron-catalyzed reaction (46). A functional iron deficiency is seen in adults with CF, possibly due to long-standing inflammation, although ferritin levels are not low in the majority of patients with CF, indicating that ferritin levels were not reflective of iron stores (47). In sputa and/or bronchoalveolar lavage fluid from patients with CF, there is an increase in the amount of extracellular iron and ferritin (48, 49). In this study, increased intracellular ferritin expression was found in lungs with CF; however, there was little intracellular iron staining. The differences may be related to the fact that we evaluated intracellular iron and ferritin levels rather than extracellular levels and that similar to the serum levels in patients with CF iron deficiency is not reflected by low serum ferritin levels (47). HO-mediated release of iron has been shown to influence ferritin expression, such that an increase in HO-1 may lead to increased ferritin expression (33). HO-1 and ferritin immunostaining was greatest in macrophages and may be directly related. However, we observed very little iron staining, which suggests that increased ferritin expression was not due to iron. Similar to HO-1, ferritin is also induced by oxidative stress and inflammation, and the similar pattern of expression may reflect a response to the same stimulus (50, 51). In conclusion, CF is a chronic disease marked by chronic necrotizing inflammation of large airways. Neutrophils contribute to the airway damage via release of oxidants, leading to lung damage. HO-1 expression is increased in CF and likely represents an adaptive response to the oxidative stress and inflammation present in the lungs of patients with CF. Our cell culture data, showing HO-1–linked reduction in P.a.-mediated cell apoptosis/injury, suggest that HO-1 plays a cytoprotective role in CF lung disease. Future studies using an animal model of CF will further evaluate the role of HO-1 in this condition.
Supported by the Cystic Fibrosis Foundation (Gary A. Visner, D.O.) and the American Heart Association of Florida (Fuhua Lu, Ph.D.). Conflict of Interest Statement: H.Z. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; F.L. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; C.L. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; D.S.Z. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; G.A.V. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form November 25, 2003; accepted in final form June 4, 2004
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
F508/ W1282X) was a generous gift from Dr. Terence Flotte, University of Florida. IB3.1 cells were cultured in Lechner and LaVeck (LHC)-8 (Biofluids, Rockville, MD) medium supplemented with 10% fetal bovine serum (Invitrogen Corp., Carlsbad, CA). Human HO-1 cDNA 1.0 kb was cloned into pcDNA3.1/zeo(+) (Invitrogen) as previously described (
