Migrating leukocytes are the source of Peroxiredoxin V during inflammation in the airways

Background We characterized changes in expression of the antioxidant protein Peroxiredoxin V (PRXV) during airway inflammation. Methods Studies in anesthetized rats and mice; PRXV expression determined by Western blot analyses and immunohistochemistry; PRXV m-RNA expression determined by Taq-Man RT-PCR. Results Bacterial lung inflammation did not change expression of PRXV in murine epithelia but produced massive influx of leukocytes highly expressing PRXV. Endotoxin and f-MLP induced leukocyte migration in rat trachea but did not change mRNA levels and PRXV protein expression in tracheal epithelial cells. In primary airway cell culture (cow), alveolar epithelial cells A549, or co-culture of A549 with murine macrophages RAW264.7, exposure to live bacteria increased expression of PRXV, which required serum. PRXV was secreted in vitro by epithelial and immune cells. Conclusion Inflammation increased expression of PRXV in airways by at least 2 mechanisms: cell population shift by massive influx of leukocytes expressing PRXV, and moderate post-transcriptional up-regulation of PRXV in epithelial cells.


Background
To ensure adequate protection against oxidative stress during states of pulmonary disease, several antioxidant systems have evolved in the epithelial cells of mammalian airways [1][2][3][4]. Peroxiredoxins I-VI (PRX I-VI) are a group of potent antioxidant proteins that are the subject of much research [5][6][7][8][9][10]. PRXs neutralize reactive oxygen by transferring electrons from thioredoxins or cyclophilins. The six PRXs differ in their intracellular distribution and are thought to serve different functions and be regulated by different mechanisms. PRXV is one of the key enzymes of cellular antioxidant defense, as it is a potent protector against DNA damage and also has other functions [11][12][13][14].
Toxic insults to the respiratory tract down-regulate synthesis of the PRXV protein. We have recently demonstrated in vivo in rat tracheal epithelial cells that cigarette smoke extract (CSE) directly down-regulated expression of PRXV, which is one mechanism of cigarette smoke toxicity [15].
Exposure of isolated tracheal segment to CSE significantly reduced mRNA levels for PRXV and the amount of PRXV protein in the epithelium. In cultures of the tracheal epithelial cell lines, primary airway cell culture, and the alveolar epithelial cells A549, CSE significantly decreased transepithelial electrical resistance, expression of PRXV protein, and significantly induced glutathione and protein oxidation. Similarly, when respiratory tract toxicity was induced in mice with naphthalene, the loss of the Clara cell population was associated with a significant decrease in PRXV expression [16]. In contrast, previous reports had indicated that PRXV was over-expressed in the lung during inflammation induced by endotoxin [17]. However, experiments in vitro in which pro-inflammatory cytokines were added to human alveolar or bronchial epithelial cells did not result in an up-regulated expression of PRXV [18]. Neither the mechanism by which PRXV is upregulated during inflammation in tissues of the lung nor the identity of the cells that are the source of PRXV production are known.
We therefore investigated the effects of gram-negative bacterial inflammation on expression of PRXV in lung, lung epithelial cells, and immune cells in vivo and in vitro. Our first aim was to determine whether inflammation in vivo influences expression of PRXV in the bronchial epithelium and alveoli. Our second aim was to use an in vivo model of inflammation to investigate whether changes of transcription or translation of PRXV in the tracheal epithelium, if they occurred, were a direct response to bacterial pathogen lipopolysaccharide (LPS) by these cells or whether the increased level of PRXV was induced by leukocyte migration. Our third aim was to determine in vitro whether exposure of the airway and alveolar epithelial cells to live bacteria, either alone or in co-culture with murine macrophages RAW264.7 changes the level of PRXV mRNA as well as protein expression and secretion.
We found that both in vivo and in vitro inflammation induced by bacteria resulted in an increased expression of PRXV in the airway epithelium by at least 2 different mechanisms: massive influx of activated leukocytes, which highly express PRXV, and moderate translational up-regulation of PRXV in the epithelial cells.

In vivo studies
Experiments in animals were performed according to protocols approved by the Institutional Animal Use Committee of the Children's Hospital Oakland Research Institute and Institute of Cytology, RAS.

Experiments in mice
Bone-marrow transplantation Recipient mice (n = 12) were given a sub-lethal dose of whole-body irradiation (5.05 Gy) the day before transplantation. While under general anesthesia (Pentobarbital, 25 mg/kg IP), the mice were infused with 10 6 whole bone-marrow cells in 0.2 ml of PBS into the jugular vein.

Bacterial lung injury
In the experimental group, six chimeric mice received intratracheal instillation of PBS (n = 3, control) or 7 × 10 6 cfu of E. coli K12 JM109 in 50 µl of PBS; the chimeric model has been described previously [16]. As a secondary control group for the bacterial inflammation study, 3 non-chimeric C57BL/6 mice received 7 × 10 6 cfu of E. coli, while 3 non-chimeric mice without known lung pathology were used as controls. These mice were euthanized and studied 1-2 weeks after the E. coli instillation.

Experiments in rats Perfusion of rat trachea
An anesthetized Sprague-Dawley rat model of an in situ perfusion of isolated tracheal segment with an intact blood supply was used, as described previously [19]. Experimental groups: Control group (n = 6): In the control group, tracheal segments were filled with PBS and sampled at 2 and 4 hours thereafter. Induced leukocyte migration (n = 4): In this group, 5 × 10 -8 M f-MLP (final concentration) was added to tracheal lumen in PBS and samples were taken at 4 hours. Endotoxin model (n = 4): In this group, LPS E. coli O55:B5 at a concentration of 100 µg/ml was applied to the inner trachea for 4 hours.
At the end of the experiment in all groups, tracheal lumen was thoroughly washed, and samples of the epithelial layer from the tracheas were cut out, frozen in liquid nitrogen, and further used for RT-PCR or immunohistochemical analyses to determine expression of mRNA or PRXV protein.

In vitro cell culture experiments
Cell culture techniques used have been described previously [20]. A549 (ATCC) cells were grown in Hank's F12 K medium with 2 mM L-glutamine, 10% fetal bovine serum (FCS) (Life Technologies, Gaithersburg, MD), and streptomycin/penicillin. Co-culture experiments were performed in DMEM with or without 10% heat-inactivated FCS. P. aeruginosa PAO1 was added for 12-24 hours to the apical surface at a concentration of 5 × 10 7 cfu/ml. Following exposure, cells were washed 3 times with PBS and then either fixed with 4% paraformaldehyde for 24 hours for IHC or collected for Western blot analyses in cell lysis buffer on ice. Experiments were performed in triplicate in 3 different cultures.
Bronchial epithelial cells Calu-3 (ATCC) (gift of Dr. T. Machen, University of California, Berkeley) were grown on the internal surface of polycarbonate membranes (0.3 µm pore size, 6.5 mm diameter) in Transwells (Costar, Cambridge, MA) with an air-liquid interface. These cells were similarly exposed to PAO1 for 12 hours at a concentration of 5 × 10 7 cfu/ml. TER, a measure of tight junctional permeability, was measured with a voltmeter (EVOMX-G, World Precision Instruments, Sarasota, FL). CTE cells were grown and studied similarly. Primary cultures of cow tracheal epithelial (CTE) cells was performed as follows: Surface of the cow tracheas was scored into thin strips and those were separated from the underlying cartilage rings and placed in cold phosphate buffered saline (PBS) + PSFG (Penicillin, Streptomycin, Fungizone, Gentamycin). Strips were placed in 40 ml of Hank's BSS, Ca 2+ /Mg 2+ free + PSFG with 1 mg/ml protease (Sigma Co), and digested overnight at 4°C. Strips were then resuspended in DME H21/F-12 mix + 5% FCS + PSFG, shaken vigorously to pull the cells off. The cell suspension was centrifuged for 10 minutes at 1000 rpm. The cells were plated 10 6 cells/cm 2 on 3 µ pore polycarbonate membranes and grown in DME-H21/F-12 mix with PSFG and a mixture of growth factors consisting of transferrin, insulin, triiodothyronine, hydrocortisone, endothelial cell growth supplement, and epidermal growth factor. As CTE cells were more resistant to PAO1 than were the Calu-3, exposure to 5 × 10 7 cfu/ml of bacteria was extended to 12 hours.

Analyses of PRXV mRNA expression
Taq-Man analyses were performed at the University of California, Davis, Lucy Whittier Molecular and Diagnostic Core Facility, at the Department of Medicine and Epidemiology by using standard techniques. For rat PRXV gene, pre-developed TaqMan PCR assay (Rn00586040-m1) was purchased from Applied BioSystems (Foster City, CA). In order to determine the most stably transcribed housekeeping gene, a housekeeping gene validation experiment was conducted on a representative number of samples. The housekeeping gene with the least standard deviation in all treatment groups (HPRT1 or TFR2) was used to normalize the target gene CT values. All gene transcriptions were expressed and are presented here as an n-fold difference relative to the calibrator.

Statistical analyses
At least six different sections from each lung were used for analyses. Cell counting was performed in 20 different randomly selected visual fields. Numbers of GFP + cells were determined as a percentage of the total number of cells (counted by numbers of PI-stained nuclei). In antibodyspecific staining, numbers of ligand-positive cells were expressed as a percentage of the total numbers of GFP + cells in 20 different visual fields. Fluorescence intensity in cells was determined by built-in Zeiss LSM software options. Western blot analyses was performed in samples from 3 different cultures; results were quantified by photometry. Data are presented as the MEAN ± SE, statistical significance by ANOVA or Student's t-test was established at p < 0.05.

Migrating leukocytes are the source of PRXV in the lung
We used a chimeric model to study the presence of leukocytes in the lung during inflammation. Transplanted mice demonstrated 30-50% bone marrow chimerism 3 months after transplantation. Engraftment of GFP + cells to the lungs of control mice (non-injured lungs) was found to be distinctive, but at a very low level (0.001-0.1%). IHC and confocal microscopy allowed us to readily identify GFP + cells and determine the expression of PRXV (Figure 1).
Half of the mice subjected to LD 50 intratracheal instillation of live E. coli died from pneumonia within 1 week. In the surviving mice, the peak of lung inflammation (7 days after E. coli instillation) was predominantly associated with the influx of GFP + leukocytes, which represented 16 ± 3 % of total lung cells. 95% of GFP + cells in the lung were CD45 + cells.
Using this model, we first determined the level of expression of PRXV in the cells of the murine bronchial epithelium (Figures 1 and 2). PRXV was abundantly expressed in the bronchial epithelium of the lungs of control mice. PRXV expression in the bronchial epithelial cells was several-fold higher than in the cells of alveoli. We did not observe significant changes in the level of PRXV expression in the bronchial epithelial cells during acute inflammation ( Figure 2). Similarly, we did not observe a significant increase in PRXV expression in the cells of alveolar epithelial lining during inflammation. However, during the development of inflammation, multiple leukocytes appeared in the lung parenchyma, most of which highly expressed PRXV ( Figure 2). Therefore, infiltration of the lung parenchyma with leukocytes resulted in an enhanced overall expression of PRXV at sites of inflammation.

PRXV protein expression is up-regulated in rat tracheal epithelium cells by f-MLP
We then used a perfused tracheal segment in vivo rat model to determine whether short-term (4 hours) exposure to f-MLP (induced leukocyte migration) or bacterial (E. coli) LPS would enhance transcription and translation of PRXV in the tracheal epithelium. Following exposure to f-MLP or LPS, the tracheal segment was carefully washed A-C: Bone marrow-derived GFP + cells infiltrated the lung following acute bacterial pneumonia Control staining of paraffin-sectioned lungs with isotype primary antibody, no non-specific green fluorescence can be noted, same panel description as in B. D-E: PRXV was abundantly present in cells of the bronchial epithelium of mice, and acute bacterial inflammation did not further significantly increase it. Confocal microscopy images of the cryosectioned lung, stained for PRXV with red-fluorescent secondary antibody. D: -Non-inflamed control lung (cryosection), original magnification × 40, bar is 50 microns. Note high expression of PRXV in the bronchial epithelium (blue arrow) but not in the alveoli (green arrow). E: Control staining with isotype primary antibody; no non-specific red fluorescence is present. F: GFP + cells, which are present in high numbers in the lung following pneumonia, highly express PRXV. Cryosection of the lung, stained for PRXV with Rhodamine-labeled antibodies (red). Fluorescence intensity of the bronchial epithelium does not differ from control (Panel D). Note the presence of bright green GFP + (or yellow due to superposition of green GFP and red PRXV signals) cells, which also highly express PRXV.
Following bacterial inflammation, GFP + cells in the lung highly expressed PRXV Figure 2 Following bacterial inflammation, GFP + cells in the lung highly expressed PRXV. Animals were transplanted with GFP + bone marrow, and progeny of GFP + cells (green fluorescence) was located to the sites of inflammation in the bronchial epithelium.  [19,20]. We therefore used this time period to assess expression of PRXV in the model of inflammation. In the f-MLP model of inflammation, a 4-hour exposure of the isolated tracheal segment to f-MLP provided a small (32%) yet significant (p < 0.05) increase in the PRXV expression in the cells of tracheal epithelium (from 182 ± 16 relative units in the control to 241 ± 3 relative units in the experimental group), but not in mRNA levels (2.36 ± 0.23 in the control versus 1.51 ± 0.22 in the experimental group). In the LPS model, we also did not observe statistically significant difference in PRXV mRNA levels in the tracheal epithelium (4.71 ± 0.9 in the control versus 2.3 ± 0.7 in the LPS experiment model). There were no significant differences in PRXV protein expression in the epithelium (data not shown).

Live P. aeruginosa bacteria up-regulates expression, but not transcription, of PRXV in cultured airway epithelium in the presence of serum
Experiments were first performed in the alveolar epithelial cell line A549, co-cultured with mouse macrophage cell line RAW264.7, both with and without the presence of serum. Western blot analyses demonstrated that co-culture of A549 with RAW264.7 and stimulation with PAO1 resulted in enhanced expression of PRXV only in the presence of serum, as shown in Figure 3. Results of quantitative IHC are shown in Figure 4. In the presence of serum, the addition of live P. aeruginosa modestly increased PRXV expression in A549 cultures, as well as in co-cultures with RAW264.7. P. aeruginosa bacteria itself were not positive for PRXV staining. The levels of PRXV mRNA did not change significantly in this system (data not shown). As can be seen from Figures 3 and 4, RAW264.7 expressed higher amounts of PRXV than the epithelial cells in culture, which is similar to our findings in vivo. However, small amounts of RAWs in the co-culture (5:1 ratio of epithelial cell/macrophages) did not significantly influence the overall expression of PRXV in the co-culture system, as epithelial cells were the predominant cell type.
Using the Calu-3 bronchial epithelial cell line, which permits electrically resistant cell layers to be obtained, we measured TER, mRNA levels, and the expression of PRXV. We used TER as a measure of tight junctional electrical permeability, a characteristic of the epithelial phenotype. Following exposure to P. aeruginosa, the TER of these cells significantly decreased (p < 0.05), indicating that -in this model -addition of bacteria produced a considerable damaging effect on the epithelial cell layers ( Figure 5). However, neither PRXV protein expression nor PRXV mRNA levels changed after exposure to PAO1; the mean relative PRXV protein expression following exposure to PAO1 was 106 ± 25% in the presence of serum and 76 ± 22% without serum as compared to baseline ( Figure 5A).
Unlike Calu-3, the primary cultures of cow tracheal epithelium showed a pattern of increased PRXV expression after exposure to bacteria which was similar to the pattern shown by the A549 cells ( Figure 5C).
Finally, using Western blot analyses of cell-conditioned medium with and without serum, we studied the presence of PRXV in the cell secretions of all cell lines that we used. Actin was used as a marker of intracellular non-diffusible proteins, and it was not found in the conditioned medium. Calu-3 and THP-1 secreted the monomeric form of PRXV into the medium ( Figure 6A). THP-1, a human acute monocytic leukemia cell line was used here as positive control for inflammatory reaction. In the medium conditioned by the A549 cells, we observed only the PRXV form with approximately 60 kDa weight, which probably reflected polymer formation. We did not observe stimulation of secretion by exposure to PAO1 in the medium with serum (data not shown) or in the serum-free medium (68 ± 21% of control) ( Figure 6B).

Discussion
We investigated both in vivo and in vitro models of the lung bacterial inflammation. In mice, rats, and cultures of human airway epithelium cells, PRXV was abundantly expressed under non-inflammatory control conditions. In rats, neither the presence of endotoxins nor f-MLPinduced migration of leukocytes in the tracheal epithelium changed mRNA levels of PRXV; f-MLP slightly increased expression of PRXV protein in the tracheal epithelium. In mice, bacterial inflammation of the lung resulted in a massive influx of leukocytes, which were the source of the increased PRXV in the lung tissues. In primary airway cell culture (cow) and alveolar epithelial cells A549, or co-culture of the epithelial cells with murine macrophages RAW264.7, exposure to live bacteria mildly, yet significantly, increased expression of PRXV protein.
Transcription of PRXV protein was not increased by exposure to bacteria in the A549 or Calu-3 cells. PRXV was secreted in vitro by both the epithelial and immune cells.
PRXV is a protein abundantly expressed under the baseline conditions in the airway epithelium, and these observations suggest that the major pathophysiological mechanism of its overall up-regulation in the lung during gram-negative bacterial inflammation is a shift in tissue cell populations due to migrating leukocytes. In the in vitro cultured airway epithelia, expression of PRXV protein was only moderately up-regulated in bacterial inflammation, while no transcriptional up-regulation was observed. PRXV was not upregulated in these cells, and its secretion in the medium was not changed. To confirm that PAO1 induced alterations in Calu-3 layers, TER of epithelial layers was measured. B: Summary results of TER following exposure of the Calu-3 epithelial cells to PAO1 with and without FCS. PAO1 induced a significant decrease in TER, which was more pronounced in the presence of FCS. Open bar -initial TER, closed bar -TER after a 4-hour exposure to medium or bacteria. In the "control" condition, cells were exposed only to the medium (or the medium with FCS) without bacteria. * -p < 0.05 compared to the initial value, + -p < 0.05 compared to the control without bacteria, n = 6. complex even in this simplified in vitro model, with multiple loops of feedback regulation, both positive and negative. Likely, PAO1 caused activation of RAWs and possibly apoptosis of these cells. Activated RAWs release an array of pro-inflammatory cytokines, which might initiate apoptosis in epithelial cells and therefore decrease PRXV expression. The fate of RAWs co-cultured with the epithelial cells is difficult to estimate, but very likely RAWs did not have much survival advantage in the medium designed for epithelial cells.
Prior studies of PRX expression showed that PRXI, II, III, V and VI are highly over-expressed in the human lung cancer cells [21]. Allergic inflammation in response to ovalbumin induced overexpression of PRXI [22], which is also well known to be induced by hyperoxia [23]. Stimulation of the A549 cells and BEAS 2 B cells with hydrogen peroxide, menadione, tumor necrosis factor α, or transforming growth factor β did not result in significant changes of PRXV expression [18]. These in vitro results are in agreement with our data.
In studies of secreted PRXV, we observed only a 60 kDa band by Western blot analyses. Peroxiredoxins may form polymers in an oxidized state. It is unlikely that the band of interest was non-specific staining, simply because it was observed only after stimulation, but not in control nonstimulated cells and not in serum. Further investigations are needed to define the mechanisms of PRXV polymerization in extra-cellular fluids.
Some insights into possible mechanisms of PRXV gene regulation can be obtained by analysis of the PRXV gene structure. The PRXV gene is located on human chromosome 11q13, which is a region of genetic linkage for atopic hypersensitivity such as bronchial asthma. A 5' promoter region (4 kb upstream of the first exon) contains 3 potential binding sites (hypoxia-response element HRE, motifs ACGTG for hypoxia-inducible transcription factor HIF-1 and one potential antioxidant/electrophile response element (ARE/EpRE, motif TGACNNNGC). Additional ARE/EpRE is also present within the first intron, along with potential binding sites for transcription factor NF-kappa-B (motif GGRNAKTCCC) and Alu-associated retinoic acid-response element (RARE, motif AGGTSMNNAGWTCR). Therefore, in theory, transcription of this gene can be modulated in response to hypoxia, inflammation, and oxidative stress by intrinsic PRXV was secreted into the medium by the epithelial cells Figure 6 PRXV was secreted into the medium by the epithelial cells. B regulatory elements. It should be noted, however, that the functional activity of these potential transcription elements in the human PRXV promoter region has not been confirmed experimentally.
According to our data from the in vivo studies, PRXV protein is already abundantly up-regulated, and it is not upregulated further by inflammation. One explanation is that the mechanisms regulating PRXV transcription cannot further increase expression in these cells, which are in constant direct contact with pathogens, antigens, and oxidants from the environment that are present in the bronchial tree in vivo.
Cells of the alveolar lining, however, are protected from these stimuli. It is unclear why we did not observe up-regulation of PRXV in the alveolar epithelium in vivo during bacterial inflammation of the airways, though we observed mild up-regulation in the A549 cells in vitro. Very likely, the model of lung bacterial inflammation that we used (instillation of bacteria into the airways) affected primarily the upper and conducting airways without causing massive inflammation in the alveolar spaces.
Adequate antioxidant protection in the lung is required for normal function, especially under conditions of oxidant stress caused by environmental factors [2]. Classic antioxidant enzymes of the lung cells that reduce hydrogen peroxide are catalase and glutathione peroxidase, while airway surface liquids and interstitial fluids are rich in superoxide dismutase and glutathione. Other hydrogen peroxide-reducing enzymes include thioredoxin-thioreductase, peroxiredoxins, and glutaredoxins [3]. Under physiologic conditions, superoxide dismutases and glutathione peroxidase are much more efficient than peroxiredixins in regulating the cell redox state. Under conditions of high oxidative stress, however, enzymes like thioredoxin may become physiologically important [1]. Hoshino and co-workers [24] showed that overexpression of thioredoxin or administration of its recombinant form protected mice against lung injury induced by pro-inflammatory cytokines and bleomycin. Adenovirus-mediated transfer of 1-cys peroxiredoxin gene was shown to protect mice from oxidative injury induced by exposure to oxygen [8]. PRXV has multiple functions: in addition to its antioxidant activity, PRXV is also a transcriptional co-repressor [13,14] and an inhibitor of p53-dependent apoptosis [10]. The anti-apoptotic activity of PRXV was demonstrated in tendon cells [9].
Our results showed that acute inflammation in the lung results in up-regulation of PRXV expression by different mechanisms. In the bronchial epithelium, we observed only a moderate rise in expression of this protein on a post-transcriptional level. Activated leukocytes that move to the lungs when inflammation occurs are a rich source of PRXV. However, this is likely to be a mechanism of leukocyte self-defense against self-induced oxidation rather than the mechanism of tissue protection. Defining the mechanism of PRXV regulation of expression in leukocytes during activation by mitogens was not the aim of present investigation and requires further study. Nonetheless, as PRXV is normally abundantly expressed in airways and as mechanisms of further PRXV up-regulation in the bronchial epithelium are limited, there is a basis for proposing therapeutic administration of this protein in recombinant form. Administration of PRXV in aerosol form may have significant therapeutic potential, especially in conditions where PRXV expression is down-regulated [15]. Inflammatory conditions are not the likely candidates for such an intervention, however, as there is no deficiency of PRXV in the airway epithelium during inflammation.

Conclusion
In vivo and in vitro bacterial inflammation mildly up-regulates expression of PRXV protein in the airway epithelial cells. An increased influx of activated leukocytes to inflamed tissues serves as a source of enhanced expression of PRXV in the lung.

Declaration of competing interests
The author(s) declare that they have no competing interests.
CL carried out Taq-man RT-PCR analyses.
VBS carried out animal experiments, IHC, study design, and drafted the manuscript.