Fluctuations of intracellular iron modulate elastin production.

Production of insoluble elastin, the major component of elastic fibers, can be modulated by numerous intrinsic and exogenous factors. Because patients with hemolytic disorders characterized with fluctuations in iron concentration demonstrate defective elastic fibers, we speculated that iron might also modulate elastogenesis. In the present report we demonstrate that treatment of cultured human skin fibroblasts with low concentration of iron 2-20 microm (ferric ammonium citrate) induced a significant increase in the synthesis of tropoelastin and deposition of insoluble elastin. Northern blot and real-time reverse transcription-PCR analysis revealed that treatment with 20 microm iron led to an increase of approximately 3-fold in elastin mRNA levels. Because treatment with an intracellular iron chelator, desferrioxamine, caused a significant decrease in elastin mRNA level and consequent inhibition of elastin deposition, we conclude that iron facilitates elastin gene expression. Our experimental evidence also demonstrates the existence of an opposite effect, in which higher, but not cytotoxic concentrations of iron (100-400 microm) induced the production of intracellular reactive oxygen species that coincided with a significant decrease in elastin message stability and the disappearance of iron-dependent stimulatory effect on elastogenesis. This stimulatory elastogenic effect was reversed, however, in cultures simultaneously treated with high iron concentration (200 microm) and the intracellular hydroxyl radical scavenger, dimethylthiourea. Thus, presented data, for the first time, demonstrate the existence of two opposite iron-dependent mechanisms that may affect the steady state of elastin message. We speculate that extreme fluctuations in intracellular iron levels result in impaired elastic fiber production as observed in hemolytic diseases.

Production of insoluble elastin, the major component of elastic fibers, can be modulated by numerous intrinsic and exogenous factors. Because patients with hemolytic disorders characterized with fluctuations in iron concentration demonstrate defective elastic fibers, we speculated that iron might also modulate elastogenesis. In the present report we demonstrate that treatment of cultured human skin fibroblasts with low concentration of iron 2-20 M (ferric ammonium citrate) induced a significant increase in the synthesis of tropoelastin and deposition of insoluble elastin. Northern blot and realtime reverse transcription-PCR analysis revealed that treatment with 20 M iron led to an increase of ϳ3-fold in elastin mRNA levels. Because treatment with an intracellular iron chelator, desferrioxamine, caused a significant decrease in elastin mRNA level and consequent inhibition of elastin deposition, we conclude that iron facilitates elastin gene expression. Our experimental evidence also demonstrates the existence of an opposite effect, in which higher, but not cytotoxic concentrations of iron (100 -400 M) induced the production of intracellular reactive oxygen species that coincided with a significant decrease in elastin message stability and the disappearance of iron-dependent stimulatory effect on elastogenesis. This stimulatory elastogenic effect was reversed, however, in cultures simultaneously treated with high iron concentration (200 M) and the intracellular hydroxyl radical scavenger, dimethylthiourea. Thus, presented data, for the first time, demonstrate the existence of two opposite iron-dependent mechanisms that may affect the steady state of elastin message. We speculate that extreme fluctuations in intracellular iron levels result in impaired elastic fiber production as observed in hemolytic diseases.
Mature elastic fibers and laminae provide extensibility and elastic recoil to vascular walls and ligaments and form a connective tissue framework of lungs, elastic cartilage, and skin (1,2). They are complex structures made of polymeric (insoluble) elastin and 12-nm microfibrils that consist of several glycoproteins, e.g. fibrillins, fibulins, and microfibril-associated glycoproteins (3)(4)(5)(6). Elastic fiber formation (elastogenesis) is a complex process involving several intracellular and extracellular events. Cells (fibroblasts, endothelial cells, chondroblasts, or vascular smooth muscle cells) must first synthesize and secrete numerous glycoproteins to form a microfibrillar scaffold upon which tropoelastin, the precursor peptides, are properly assembled and covalently cross-linked by lysyl oxidase into a resilient polymer, insoluble elastin (7)(8)(9)(10). Production of elastin reaches its highest levels in the third trimester of the fetal life and steadily decreases during early postnatal development (11,12). In undisturbed tissues, elastic fibers may last over the entire human lifespan (13,14).
The net deposition of elastin appears to be controlled on both the transcriptional level (tropoelastin mRNA message expression (15,16)) and-post-transcriptional level (tropoelastin message stability (17)(18)(19)). There are also several other post-transcriptional events, which control secretion of tropoelastin monomers and their proper extracellular assembly (20,21) and regulate the cross-linking of tropoelastin into the polymeric "insoluble" elastin, the most durable element of the extracellular matrix (14).
Diffuse elastic fiber defects, resembling those reported in inherited pseudoxanthoma elasticum, have also been detected in patients with ␤-thalassaemia and sickle cell anemia (30), and in other hemolytic anemias (31,32). The genetic basis for these diseases cannot be directly linked to any structural or regulatory components involved in elastic fiber production (30,33,34). However, it has been suggested that the accumulation of iron in these patients, resulting from hemolysis, increased iron absorption, and multiple blood transfusions may lead to acquired elastic tissue defects (30).
Iron is a physiologically essential nutritional element for all life forms (35). It plays critical roles in electron transport and cellular respiration, oxygen transport by hemoglobin, cell proliferation, and differentiation (36). It has been shown that modulating intracellular iron levels may also affect expression of numerous genes that are not directly involved in iron metabolism, such as protein kinase C-␤, an important component of intracellular signaling pathways (37), or those encoding extracellular matrix (ECM) 1 components (36,38). It has been demonstrated that dietary iron overload in rats resulted in an increase in the steady-state level of pro-␣2(I)-collagen in hepatocytes (39) and that 50 M iron treatment stimulated collagen gene expression in cultured stromal hepatic cells by inducing the synthesis and binding of Sp1 and Sp3 transcription factors to two regulatory elements located in the collagen ␣1 (I) promoter region (40). On the other hand, iron loading in cultured cardiac myocytes and fibroblasts decreased the expression of transforming growth factor-␤, biglycan, and collagen type I mRNA, whereas it facilitated the expression of decorin mRNA. Interestingly, iron deprivation exerted a similar effect, suggesting that the expression of these genes involved in extracellular matrix production is regulated by certain iron-dependent mechanisms (41).
The molecular basis of iron-dependent mechanism(s) regulating the expression of ECM-encoding genes is not well understood. Because raising levels of iron may overwhelm the iron binding capacity of transferrin, resulting in the appearance of non-transferrin bound iron (42), which is capable of catalyzing the formation of the hydroxyl radicals (through the Fenton and Haber-Weiss reactions) (43), it has been suggested that irondependent induction of reactive oxygen species (ROS) may modulate the transcription of these genes (36,44). The possibility of iron-dependent oxidative damage to elastic fibers has also been suggested, but not proven (30).
The present study was designed to elucidate the mechanisms by which iron might modulate elastogenesis. We utilized normal human skin fibroblasts that are capable of elastic fiber production in vitro to explore whether changes in concentrations of iron would affect elastin gene expression and the subsequent deposition of elastic fibers. We also tested the influence of different iron concentrations on generation of intracellular ROS, and attempted to identify on which level these ironinduced ROS affect elastogenesis.
Secondary antibody fluorescein isothiocyanate-conjugated goat antirabbit was purchased from Sigma. DNeasy Tissue system for DNA assay and RNeasy Mini Kit for isolation of total RNA were purchased from Qiagen (Mississauga, Ontario, Canada). A OneStep RT-PCR kit was purchased from Qiagen. A SuperScript First-Strand Synthesis System for RT-PCR was purchased from Invitrogen. TaqMan Universal PCR master mix, TaqMan GAPDH control, and Assays-on-Demand Gene Expression probe for elastin were purchased from Applied Biosystems (Foster City, CA). The radiolabeled reagents, [ 3 H]valine and [ 3 H]thymidine, and Rediprime (II) Random Primer labeling system were purchased from Amersham Biosciences. Hybridization solution Miracle Hyb was purchased from Stratagene (Cedar Creek, TX), and the human GAPDH control was purchased from Clontech (Palo Alto, CA).
Cultures of Normal Human Skin Fibroblasts-Fibroblasts grown from skin biopsy explants of six normal subjects, aged from 2 months to 10 years, were obtained from the cell repository at The Hospital for Sick Children in Toronto with the permission of the Institutional Ethics Committee. Fibroblasts were routinely passaged by trypsinization and maintained in Dulbecco's modified eagle's medium supplemented with 1% antibiotics/antimycotics, and 10% FBS. In all described experiments passages 2-6 were used.
In experiments aimed at assessing ECM production, fibroblasts were initially plated (100,000 cells/dish) and maintained in normal medium until confluency at which point they produce abundant ECM. Confluent cultures were then treated for 72 h with or without FAC producing iron concentrations from 2 to 200 M. The low iron concentration (2 and 20 M) of iron utilized in the present study remained in the range that did not induce any disturbances in cellular metabolism when tested by other investigators (45). The high iron concentration (200 M) was comparable to concentrations used in studies of iron overload (41,46).
In some experiments the membrane permeable ferric iron chelator, DFO, was added 30 min prior to FAC treatment. For the experiments conducted in the presence of various antioxidants, the antioxidants were applied 1 h prior to FAC treatment. For experiments conducted in serumfree conditions, 7-day-old confluent fibroblast cultures were starved for 12 h in serum-free medium and incubated with various concentrations of iron (as FAC) for additional 72 h in serum-free medium.
Immunostaining-At the end of the incubation period confluent cultures were fixed in cold 100% methanol at Ϫ20°C for 30 min and blocked with 1% normal goat serum for 1 h at room temperature. Cultures were then incubated for 1 h with 10 g/ml polyclonal antibody to tropoelastin followed by a 1-h incubation with fluorescein-conjugated goat anti-rabbit. Nuclei were counterstained with propidium iodide. Secondary antibody alone was used as a control. All of the cultures were then mounted in ELVANOL, and examined with an Nikon Eclipse E1000 microscope attached to a cooled charge-coupled device camera (QImaging, Retiga EX) and a computer-generated video analysis system (Image-Pro Plus software, Media Cybernetics, Silver Spring, MD).
Quantitative Assays of Tropoelastin and Insoluble Elastin-Normal human skin fibroblasts were grown to confluency in 35-mm culture dishes (100,000 cells/dish) in quadruplicates. Then, 2 Ci of [ 3 H]valine/ ml of fresh media was added to each dish, along with or without 2, 20, 100, and 200 M FAC. Cultures were incubated for 72 h, and the soluble and insoluble elastin was assessed separately in each dish. The cells were extensively washed with PBS, and the soluble proteins present in the intracellular compartments were extracted overnight at 4°C with 0.1 M acetic acid in the presence of proteinase inhibitors. After centrifugation, the supernatants were pre-cleaned by 30-min incubation with 50 l of 4% protein A-beaded agarose, then 500 l of the supernatant was incubated with 5 g of polyclonal antibody to tropoelastin for 2 h and subsequently with 50 l of 4% protein A-beaded agarose for 3 h at 4°C. The protein A-containing beads were sedimented by centrifugation, washed with immunoprecipitation buffer, mixed with scintillation fluid, and counted (21). The remaining cultures containing cell remnants and deposited insoluble extracellular matrix were scraped and boiled in 500 l of 0.1 N NaOH for 45 min to solubilize all matrix components except elastin. The resulting pellets containing the insoluble elastin were then solubilized by boiling in 200 l of 5.7 N HCl for 1 h, and the aliquots were mixed in scintillation fluid and counted (21). Aliquots taken from each culture were also used for DNA determination according to Ref. 47, using the DNeasy Tissue System from Qiagen. Final results reflecting amounts of metabolically labeled insoluble elastin in individual cultures were normalized per their DNA content and expressed as cpm/g of DNA. In separate experiments, the specified treatment described in the figure legends was added along with 2 Ci of [ 3 H]valine/ml media to confluent cultures of normal human skin fibroblasts in 35-mm culture dishes (100,000 cells/dish) in quadruplicates for 72 h. The conditioned media was then removed, the cell layers were washed, and incorporation of [ 3 H]valine into the insoluble elastin was assessed as described above.
Assessment of Cell Proliferation-Normal human skin fibroblasts were suspended in Dulbecco's modified eagle's medium containing 10% FBS and plated in 35-mm culture dishes (100,000 cells/dish) in quadruplicates. Twenty-four hours later, the cells were transferred to the serum-free medium for synchronization of their cell cycle and then maintained in the presence or absence of FAC (2-200 M) and 2 Ci of [ 3 H]thymidine/ml in media with 10% FBS for 72 h. These cultures were then washed in PBS and treated with cold 5% trichloroacetic acid twice for 10 min at 4°C. One-half milliliter of 0.3 N NaOH was added to all dishes, incubated at room temperature for 30 min, and 200-l aliquots of each culture mixed with scintillation fluid were counted.
Assays of Intracellular ROS Levels-The ROS-sensitive fluorescent probe, CM-H 2 DCFDA has been used to detect oxidative activity in cultured fibroblasts (48). This probe passively diffuses into the cell interior, and only upon oxidation is a fluorescent product released that can be visualized under fluorescent microscope or captured by flow cytophotometry, when it is excited at 480 nm. To measure intracellular ROS production, normal human skin fibroblasts were plated on glass coverslips in 35-mm dishes (50,000 cells/dish) and grown to confluency. The cells were then washed with PBS and incubated with or without 10 M CM-H 2 DCFDA for 30 min in fresh media. The cells were then washed again in PBS and incubated with new media in the presence or absence of FAC (2-400 M) for three additional hours. The cells were then washed twice with PBS before being mounted to the glass slides, and the images were captured using a fluorescence microscope under identical parameters of contrast and brightness.
In addition, the quantification of this reaction was performed by flow cytometry ( excitation, 480 nm; emission, 520 nm). Quadruplicate cultures of fibroblasts were preincubated with CM-H 2 DCFDA and maintained in the presence or absence of FAC as described above. To reduce stress-induced oxidant activation, the cells were cooled and harvested by trypsinization at 4°C. They were then collected by centrifugation (4°C, 1000 rpm for 3 min), washed in cold PBS, and fixed with 4% formaldehyde for 10 min in the dark and analyzed by flow cytometry (FACSCalibur, BD Biosciences).
Northern Blots-Normal human skin fibroblasts were grown to confluency in 100-mm culture dishes. Fresh media was added along with or without 2, 20, and 200 M of FAC for 24 h. Total RNA was isolated using RNeasy Mini Kit according to manufacturer's instructions, and 10 g were resolved by electrophoresis on formaldehyde-1% agarose gels. Recovery of 18 S and 28 S rRNA was analyzed using ethidium-bromide staining and image analysis on a Gel Doc 1000 optical-system (Bio-Rad). RNA was transferred onto Hybond-N membrane (Amersham Pharmacia Biotech) by capillary transfer in 10ϫ SSC and immobilized by UV cross-linking. Human elastin cDNA recombinant probe H-11 (49,50) was radiolabeled with 32 P random primer method and incubated One-step RT-PCR-To further confirm the effect of iron on elastin mRNA levels, confluent normal human skin fibroblast cultures were treated with or without intracellular ferric iron chelator, 20 M DFO in the presence or absence of 20 M FAC for 24 h. Total RNA was extracted using the RNeasy Mini kit, according to the manufacturer's instructions, and 1 g of total RNA was added to each one-step RT-PCR (Qiagen OneStep RT-PCR kit), and reactions were set up according to the manufacturer's instructions in a total volume of 25 l. The reverse transcription step was performed for elastin and ␤-actin reactions at 50°C for 30 min followed by 15 min at 95°C. The elastin PCR reaction (sense primer: 5Ј-GGTGCGGTGGTTCCTCAGCCTGG-3Ј, antisense primer: 5Ј-GGGCCTTGAGATACCCCAGTG-3Ј; designed to produce a 255-bp product) was performed under the following conditions: 25 cycles at 94°C denaturation for 20 s, 63°C annealing for 20 s, and 72°C extension for 1 min, then 1 cycle at 72°C with a final extension for 10 min.
The ␤-actin PCR reaction (sense primer: 5Ј-GGTCAGAAGGATTC-CCTATGTG-3Ј; antisense primer: 5Ј-ATTGCCCAATGGTGATGAC-CTG-3Ј; designed to produce a 615-bp product) was performed under the following conditions: 25 cycles at 94°C denaturation for 60 s, 60°C annealing for 60 s, and 72°C extension for 120 s, the 1 cycle at 72°C with a final extension for 10 min. 5-l samples of the elastin and ␤-actin PCR products from each reaction were run on a 2% agarose gel and post-stained with ethidium bromide. The amount of elastin mRNA was standardized relative to the amount of ␤-actin mRNA.
Data Analysis-In all biochemical studies quadruplicate samples in each experimental group were assayed in two separate experiments. Mean and standard deviations (S.D.) were calculated for each experimental group, and statistical analyses were carried out by analysis of variance, p value of Ͻ0.05 (p Ͻ 0.05) was considered significant.

Low and High Doses of Iron Produce Opposite Effects on
Production of Insoluble Elastin-We first tested low concentrations of iron 2-20 M (supplied as FAC) relevant to physiological concentrations of mammalian serum iron (10 -30 M (52)) and then higher concentrations (100 and 200 M) relevant to iron overload. Immunostaining of confluent fibroblast cultures with anti-elastin antibody revealed that 3-day-long treatment with 2 and 20 M of iron significantly increased the production of elastic fibers over control levels (Fig. 1A). Interestingly, raising iron concentration to 100 M did not induce better elastin deposition than treatment with 20 M, and treatment with 200 M iron dragged elastin deposition back to the control levels. Metabolic labeling of cultured fibroblasts with [ 3 H]valine followed by quantitative assays of insoluble elastin confirmed the results obtained with immunocytochemistry (Fig.  1B). Importantly, the same trend induced by different concentrations of iron was observed in cultures maintained in the presence and absence of serum (Fig. 1B). Results of parallel experiments measuring the incorporation of [ 3 H]thymidine demonstrated that the detected stimulation on elastogenesis in cultures treated with low iron concentrations was not due to increased cellular proliferation rate and that the reverse effect observed at higher concentrations of iron was not due to cellular cytotoxicity (Fig. 1C).
Because the net production of elastic fibers depends on the coordinated expression of multiple factors, we tested the expression of three major factors facilitating elastogenesis by immunofluorescent microscopy after exposure of normal human skin fibroblasts to low (20 M) and high (200 M) iron concentrations. In contrast to elastin, the immunodetectable levels of fibrillin-1, a major component of fibrillar scaffold, the elastin-binding protein, required for normal tropoelastin secretion and extracellular assembly, and lysyl oxidase, the enzyme responsible for elastin cross-linking, were not changed in cultures treated with 20 and 200 M of iron (data not shown).  (Fig. 2B, left panel). We further examined elastin gene expression under same experimental conditions by quantitative real-time RT-PCR analysis. This confirmed a substantial (ϳ3-fold) increase in elastin mRNA levels in cultures treated for 24 h with 20 M iron and a significant reduction in tropoelastin mRNA in cultures maintained in the presence of 200 M iron (Fig. 2B, right panel). after ϳ6-h incubation (Fig. 2C). These observations suggested that the treatment with high iron concentrations induce a decay in elastin mRNA levels.

The Influence of Iron on Elastin mRNA Levels and Message
Intracellular Iron Levels Influence Elastin Production-Because we have demonstrated that addition of low iron concentrations (up to 20 M) to the culture media induced ϳ3-fold increase in elastin mRNA steady-state levels and subsequent increase in elastic fiber formation, we further tested whether this effect is specifically dependent on intracellular iron. To test this we utilized a highly specific membrane-permeable ferric iron chelator, DFO, which have been shown to deplete intracellular pools of free iron (53)(54)(55). Results of immunocytochemistry (Fig. 3A), quantitative assay of newly deposited (metabolically labeled) insoluble elastin (Fig. 3B), and one-step RT-PCR analysis assessing elastin mRNA levels (Fig. 3C) demonstrated that chelating intracellular iron in cultured fibro- blasts with 20 M DFO significantly reduced elastin mRNA levels and consequent elastic fibers deposition, as compared with untreated control. Simultaneous treatment of cultured fibroblasts with equimolar amounts (20 M) of ferric iron and DFO abolished the iron-induced increase in elastin mRNA levels and elastin deposition (Fig. 3). Cumulatively, these data indicate that chelatable intracellular iron facilitates normal expression of elastin gene and the consequent production of insoluble elastin.
The Effect of Iron on the Production of Intracellular ROS-It has been well established that iron has the capacity to generate ROS through the Fenton reaction and that ROS acting as second messengers may induce specific intracellular signaling pathways (56). We tested whether different iron concentrations that we utilized in our study may affect the production of ROS in normal human skin fibroblasts. Both fluorescence microscopy and flow cytometry measuring intracellular levels of ROS with a specific fluorescent probe showed that cells incubated with 2-40 M iron produced the same amount of ROS as untreated controls. In contrast, the addition of higher concentrations of iron (100 -400 M) to the culture medium induced a dose-dependent increase in ROS production (Fig. 4).
Scavenging of Intracellular Hydroxyl Radical Reverts Inhibition of Elastin Production in Cells Treated with High Concentration of Iron-Because the above results indicate that the decrease in elastogenesis in cells treated with high concentrations of iron coincide with an increase in the production of intracellular ROS, we anticipated a pathophysiological link between these two effects. Results of the next series of experiments confirmed this hypothesis. We found that treatment of cultured fibroblasts with 200 M iron and DMTU, the membrane-permeable scavenger of hydroxyl radicals (57), reversed the inhibitory effect of 200 M iron treatment on elastin deposition (Fig. 5). In fact, 200 M iron treatment in the presence of DMTU produced almost a 2-fold increase in elastin production as compared with cultures treated with 200 M iron alone. We could not, however, detect a similar effect in cultures simultaneously treated with 200 M iron, the membrane-impermeable antioxidants, catalase and superoxide dismutase (data not shown), and the membrane-permeable superoxide dismutase mimetic, Tempol (Fig. 5). We also found that pretreatment of cells with any of the four antioxidants prior to the addition of 20 M iron did not change the stimulatory effect on elastin deposition (data not shown). These results further indicate that 20 M iron treatment does not stimulate intracellular ROS production. DISCUSSION Impaired elastic fibers have been detected in patients with ␤-thalassaemia and in other inherited hemolytic disorders characterized with fluctuating iron levels (30,32,34,58), how-ever, no pathophysiological link between these two clinical features have ever been disclosed.
Mammalian cells maintain steady levels of metabolically active iron, also referred as the chelatable iron pool or labile iron pool (LIP), through the regulation of iron uptake and storage (59,60), which is critical to maintaining normal cellular iron requirements (43). It has been shown that cells treated with lower than 25 M iron (supplied as FAC) are able to maintain an equilibrium between LIP and iron bound to ferritin without a disturbance in cellular metabolism (45). Results of the present in vitro study demonstrate, for the first time, that treatment of normal human skin fibroblasts with such concentrations of iron can up-regulate tropoelastin synthesis and its final extracellular deposition into elastin fibers. Importantly, these low iron concentrations did not cause any increase in cellular proliferation rate (Fig. 1A). On the other hand, we have also shown that treatment of fibroblasts with elevated iron concentrations (100 -200 M FAC) slightly stimulated cellular proliferation but failed to further stimulate elastin production and in fact elicited an inhibitory effect.
The role of metal ions in transcription regulation has been documented for zinc (61,62), copper (63), and calcium (64). It is becoming increasingly evident that fluctuations in iron levels FIG. 6. A schematic depicting two parallel iron-dependent mechanisms that may modulate elastin mRNA levels and consequently affect the net production of elastin. The intracellular chelatable iron binds to the transcription factor or cofactor, which stimulates a specific cis element within the elastin promoter region and, in turn, up-regulate transcription of elastin mRNA. On the other hand, an excess of intracellular chelatable iron also induces production of ROS that reduce the stability of newly transcribed elastin mRNA.
can influence the expression of various genes through non-iron responsive element-mediated changes (36,38,65). Because we found that treatment of cultured fibroblasts with low iron concentrations (2-20 M) caused 2-to 3-fold increase in the elastin mRNA level (Fig. 2B) and that the elimination of the LIP by treatment with a highly specific intracellular ferric iron chelator, DFO, led to a significant decrease in elastin mRNA levels and consequent elastin deposition (Fig. 3), we concluded that low intracellular concentrations of chelatable iron may facilitate normal elastogenesis. These data strongly indicate a new level of complexity to the poorly explored area of elastin gene regulation. However, the precise iron-dependent mechanism responsible for up-regulation of elastin gene transcription remains to be elucidated.
Using the analogy to the iron-dependent mechanism suggested for the activation of other genes such as protein kinase C-␤ (37), we speculate that certain iron-responsive transcriptional regulatory elements could be located within the elastin 5Ј-flanking region. However, to date only one true activating sequence has been identified within the elastin promoter, the nuclear factor-1 (NF-1) binding sequence, which, upon the interaction with one of the NF-1 family members, can directly activate elastin gene transcription (16). In separate studies a newly identified nuclear protein, pirin, has been show to bind to NF-1 and was proposed as a functional cofactor for regulating gene transcription at the level of DNA complexes (66). Because pirin has recently been demonstrated to contain an iron-binding domain that is required for its function (67), we could speculate that the iron-induced increase in elastin message level would result from the pirin-dependent activation of NF-1 and consequent up-regulation in elastin gene transcription. However, more studies are needed to confirm this hypothesis.
The present study was designated to elucidate the cellular mechanism by which heightened iron concentrations might affect elastogenesis. Previously published data (45) indicated that treatment with elevated iron concentrations (100 -200 M FAC) causes an expansion in the LIP that cannot be buffered by the sequestration capacity of cellular ferritin. The high iron concentration (100 -200 M) utilized in the present study was comparable to concentrations used in other investigations of iron overload (41,46). Our experimental conditions were also relevant to iron overload in ␤-thalassaemia patients, where the levels of non-transferrin-bound iron rise substantially (68,69). The non-transferrin-bound iron and its intracellular counterpart, the LIP, are able to catalyze the formation of the most toxic hydroxyl radicals through the Fenton and Haber-Weiss reactions (43,70). The link between elastic fiber defects reported in chronic hemolytic syndromes and iron overload-induced oxidative stress have been previously suggested but not documented (30,32,34,71,72). Speculations about the involvement of oxidative stress in the development of elastic fiber pathology present in patients who suffer from inherited pseudoxanthoma elasticum, characterized with genetic deficiency of the MRP6 transporter (also needed for transport of reduced glutathione conjugates, free radical scavengers (73), have been raised by previous investigators (74) but not proven).
Results of our study provide the anticipated experimental evidence that the expansion of the intracellular LIP in cultured fibroblasts, treated with high concentrations of iron, result in a significant rise in intracellular levels of hydroxyl radicals, and the consequent decrease in elastic fiber formation. The fact that scavenging intracellular hydroxyl radicals with DMTU induced by 200 M iron treatment leads to restoration of elastin deposition (ϳ2-fold increase over untreated control, Fig. 5), confirmed our hypothesis that iron overload may impair elastogenesis.
It has been previously documented that ROS may alter the expression of certain genes by interfering with message stability (75)(76)(77)(78). Our present data provide evidence that the irondependent generation of ROS indeed coincided with a decrease in the stability of elastin mRNA (Figs. 2C and 5). Although several mechanisms for regulating the stability of mRNAs have been described, only a few have been well characterized. In general, removal of 5Ј-cap structures or 3Ј-polyadenosine tails are considered to lead to rapid degradation of messages (79,80). Sequence elements in the 3Ј-untranslated region have also been implicated in regulation of the stability of many mRNAs (81)(82)(83)(84). Although the stability of elastin mRNA appears to be an important factor in regulating the expression of this protein, which has been reported to be affected by transforming growth factor-␤ (17,85), phorbol esters (86), and vitamin D (87,88), details of the mechanism of this regulation are still not understood.
Conserved GA-rich sequences present in the 3Ј-untranslated region of elastin have been shown to be an important element in the regulation of elastin mRNA stability (18). The presence of this sequence has been shown to be particularly susceptible to RNase attack when this site is not protected by yet unidentified binding protein factor(s) (18). We speculate that ROS might alter the binding of this putative protein to the GA-rich sequence in the 3Ј-untranslated region of elastin mRNA and consequently allow RNase to attack. Alternatively, elastin mRNA might be directly affected by oxidants or oxidant-dependent signaling molecules stimulating its degradation.
Despite the fact that our current study did not identify the one or more putative iron-dependent factors modulating elastin gene expression or stability of elastin message, the presented data clearly indicate that iron overload-induced oxidative stress interferes with elastogenesis. Importantly, we also demonstrate that the inhibitory effect of free radicals on elastogenesis can be minimized or eliminated by utilization of cell membrane-permeable antioxidants. We therefore suggest that further studies with these reagents would lead to development of novel therapeutic approaches for patients characterized with elevated iron levels and elastinopathy.
In conclusion, the present study, for the first time provides experimental evidence that intracellular chelatable iron can facilitate elastin gene expression resulting in abundant production of elastic fibers. Our data also indicate the existence of a parallel mechanism, in which an excess of intracellular chelatable iron induces the formation of free radicals that, through an unknown molecular manner, down-regulate elastin message stability and consequently decrease elastogenesis (Fig. 6). We suggest that an apparent balance between these two iron-dependent mechanisms may constitute a novel level of complexity regulating normal elastogenesis. We also believe that a disturbance of this balance, caused either by increased levels of free iron or chelation of intracellular iron, may result in impaired elastin production as observed in human hemolytic disorders.