not just a passive barrier, the airway epithelium is active in airway defense by releasing cytoprotective mucus, defensins, and a host of lipid mediators, prostaglandins, and leukotrienes (9). In addition, it helps to regulate local inflammation and immune responses in part through the generation of numerous cytokines and chemokines that mediate recruitment and activation of inflammatory cells (9). However, the regulation and activation of epithelial cells in inflammatory states are not well defined.

The cellular hallmarks of airway inflammation are the influx of inflammatory cells, the accumulation of inflammatory mediators, disruption of the epithelial layer, and the increased turnover and production of the extracellular matrix. This is often accompanied by mucus gland hypertrophy with excess mucus production and inflammatory cell infiltration in the airway walls and lumen (12). Interestingly, the inflammatory cells present in the airway walls are monocytes, whereas in the lumen the cells are predominantly neutrophils. These neutrophils release numerous chemokines, proteases, and elastases that add to the destruction and turnover of the extracellular matrix.

In fact, various extracellular matrix components, such as hyaluronan (HA), are found in increased concentrations in bronchoalveolar fluid from patients with chronic airway inflammation and are used as a marker of inflammation (35, 42). HA, a negatively charged high-molecular-weight glycosaminoglycan made up of repeating disaccharide units, is ubiquitously distributed in the extracellular matrix (24, 25). It is a main constituent of the basement membranes in normal lungs. In vivo, at sites of inflammation, high-molecular-weight (HMW) HA (mol wt 2–6 × 10⁶) can be depolymerized to lower-molecular-weight (mol wt 0.2 × 10⁶) fragments via oxygen radicals and enzymatic degradation by hyaluronidase, β-glucuronidase, and hexosaminidase (24, 25, 32). Likewise, inflammatory cytokines, such as TNF-α, stimulate pulmonary fibroblasts to produce increased amounts of HA fragments (39).

HA fragments induce the expression of a wide array of inflammatory genes such as chemokines, cytokines, metalloelastase, and nitric oxide synthase (NOS) in alveolar macrophages (11, 14–16, 30). Chemokines play a pivotal role in immune responses by orchestrating the recruitment and activation of specific inflammatory cells to discrete areas of inflammation. Several chemokines, for instance, the CXC chemokines IL-8 and IFN-γ-inducible protein 10 (IP-10), have been noted to be markedly upregulated in airway inflammation (34, 38, 41). Although the exact role of these chemokines in airway inflammation is unclear, both appear to recruit inflammatory cells such as neutrophils (IL-8) and monocytes (IP-10) (34, 38, 41). Potentially, the balance between these chemokines may modify the inflammatory milieu in a more proangiogenic (IL-8) or antiangiogenic (IP-10) direction that could ultimately affect the extent of the inflammation and tissue destruction (21–23).

In this report, we demonstrate the differential induction of CXC chemokines in airway epithelial cells by HA fragments. These data provide an important functional link between the extracellular matrix and the ability of airway epithelial cells to induce inflammatory mediators. We propose that extracellular matrix fragments, generated at sites of inflammation, are not just the result of inflammation but play a key role in the perpetuation and augmentation of the inflammatory response via their interactions with the epithelial cells.

MATERIALS AND METHODS

Cells and cell culture.

NCI-H292 airway epithelium-like cells (derived from a human pulmonary mucoepidermoic carcinoma) and the BEAS-2B bronchial epithelial cell line (derived from adenovirus 12-SV40-transformed normal human bronchial epithelium) were obtained from the American Type Culture Collection and maintained in RPMI with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin (Biofluids, Rockville, MD) at 37°C under 5% CO₂.

Tissue source, preparation, and culture.

Human bronchial tissue was purchased from a nonprofit organization that makes available for medical research organs and tissues from organ donors. Human tissue obtained in this manner and provided without personal identifiers has been determined by our institutional Committee on Human Research to be exempt from human subject review. Bronchial airways (2- to 18-mm diameter) from these sources were dissected free of connective tissue and parenchyma. Epithelial cells were prepared as previously described in detail (37) by digesting the tissues overnight at 4°C in 0.1% protease in Ham's F-12 medium containing antibiotics. FCS (10%) was added to neutralize the protease, and the epithelial cells were freed from the tissue by agitation and isolated by centrifugation. The washed cells were then seeded at a density of 31.6 × 10⁴ cells/cm² and grown on Vitrogen 100-coated P100 plastic dishes in serum-free bronchial epithelial growth medium. When confluent, the cells were frozen at −70°C or immediately reseeded on collagen-coated Falcon filter inserts or plastic dishes and grown to confluence under bronchial epithelial growth medium. When confluent, cells grown on inserts were transferred to the lower chamber to establish the cells at the air-liquid interface (37).

Chemicals and reagents.

Purified HA fragments from human umbilical cords were purchased from ICN Biomedicals (Costa Mesa, CA). The HA-ICN preparation is free of protein (<2%) and other glycosaminoglycans with a peak molecular weight of 200,000 by gel electrophoresis (26). Healon (HMW HA) was purchased from Pharmacia & Upjohn. Chondroitin sulfate A (CSA), chondroitin sulfate B (CSB), heparan sulfate, HA disaccharides, Streptomyces hyaluronidase, and LPS (Pseudomonas aeruginosa serotype 10) were purchased from Sigma (St. Louis, MO). Proteosome inhibitor-1, wortmannin, the p38 inhibitor SB-220025, JNK inhibitor I (L-form, cell permeant), and the MEK inhibitor PD-98059 were purchased from Calbiochem and FuGENE 6 transfection reagent from Roche (Indianapolis, IN). Stock solutions of reagents were tested for LPS contamination by Limulus amoebocyte assay (Sigma), and LPS contamination in the HA-ICN preparation was <10 ng/ml. All studies using HA-ICN fragments were performed in the presence of 10 μg/ml polymixin B, which we previously demonstrated to block up to 1 μg/ml LPS (15, 29).

Northern blot analysis and reverse transcriptase-PCR of mRNA.

RNA was extracted from confluent cell monolayers of cells with TRIzol reagent (Invitrogen) per manufacturer's guidelines. Northern blot analysis was performed as described previously (13). Bands were quantitated with a phosphoimager (Molecular Dynamics, Sunnyvale, CA). Reverse transcription was performed with a Invitrogen SSIII first strand kit per manufacturer's guidelines. Multiplex PCR for IP-10 and GAPDH was performed with a MPCR kit for human chemokine genes set-1 from Maxim Biotech (San Francisco, CA).

ELISA for protein secretion.

ELISAs for IL-8 and IP-10 were performed per manufacturer's guidelines (R & D, Minneapolis, MN). Colorimetric changes were measured in an ELISA plate reader and analyzed with Microplate Manager III (Bio-Rad) software.

Electrophoretic mobility shift assays.

Electrophoretic mobility shift assays (EMSAs) were conducted with 6% polyacrylamide gels, as previously described (13). NF-κB and AP-1 double-strand DNA consensus sequence probes from Santa Cruz Biotechnology were used for analysis. For supershift analysis nuclear extracts were simultaneously incubated with 1 μg of the indicated antibody and the labeled probe before EMSA. The following antibodies were obtained from Santa Cruz Biotechnology: NF-κB p50, NF-κB p65, c-jun, and c-fos.

Transient transfections.

Transient transfections were performed with FuGENE 6 in H292 cells per manufacturer's guidelines. Equivalent numbers (1.5 × 10⁶ cells/well) were transfected with 0.25 μg of reporter construct with FuGENE 6 in a 3-to-1 ratio, rested overnight, and stimulated for 18 h. Each condition was performed in triplicate for each experiment, and the data presented represent the average of three or four different experiments. Cells were lysed in 125 μl of cell lysis buffer, and luciferase assays were performed in triplicate on 25 μl of lysate (13–15). IL-8 and IP-10 reporter constructs were kind gifts from Aristides G. Eliopoulos (Cancer Research UK Institute for Cancer Studies, Birmingham, UK) and Daniel Muruve (University of Calgary, Calgary, AB, Canada), respectively, and were described previously (2, 6). Luciferase expression was measured with a Dual Luciferase Kit (Promega) and a Zylux femtomaster FB-12 luminometer.

Statistical analysis.

Statistical analysis was performed between groups with an ANOVA factorial analysis program from Graph Pad Prism 4 (GraphPad Software). A difference between groups of P < 0.05 was considered significant.

RESULTS

HA fragments-induce IL-8 and IP-10 protein in primary human airway epithelial cells.

In light of the accumulation of low-molecular-weight HA in inflamed airways and our previous findings (14, 29) showing the ability of HA fragments to induce cytokines and chemokines in macrophages, we investigated whether HA fragments are able to induce inflammatory gene expression in airway epithelial cells. Using human airway epithelium isolated from the lungs of healthy, nonsmoking organ donors, we evaluated the ability of HA fragments to induce the CXC chemokines IL-8 and IP-10. Primary human airway epithelial cells were grown either under medium or at an air-liquid interface and stimulated with HA fragments (100 μg/ml) at 37°C, conditioned cell medium was collected, and ELISAs for IL-8 and IP-10 were performed. Figure 1, A and B, shows a significant time-dependent induction of these chemokines in HA fragment-treated cells. Data were similar for primary cells grown under media and at an air-liquid interface. These data clearly establish a role for HA fragments, produced at sites of inflammation, in the upregulation of epithelial cell chemokine production.

Given our finding that HA induces chemokine gene expression in primary airway epithelial cells, we wanted to duplicate the findings in NCI H292 airway epithelium-like cells. These cells were stimulated with HA fragments (100 μg/ml) or LPS (1 μg/ml) for 6 h at 37°C, total RNA was isolated, and Northern blot analysis or reverse transcriptase-PCR was performed for IL-8 and IP-10, respectively. Figure 1C shows a marked induction of steady-state IL-8 mRNA by HA fragments, whereas there was only a faint induction by LPS. Of note, the LPS inhibitor polymixin B was used in all conditions, except for LPS stimulation, to control for potential LPS contamination in the HA, although H292 cells are relatively resistant to the effects of LPS. We found a similar induction of IP-10 mRNA by HA fragments in these epithelial cells with reverse transcriptase-PCR (Fig. 1D). Thus HA fragments induce both IL-8 and IP-10 in NCI H292 epithelial cells at the mRNA level. In additional experiments, we confirmed that HA fragments also induced IL-8 and IP-10 in H292 cells at the protein level and that these cells were also resistant to LPS-induced IL-8 and IP-10 (data not shown). We found similar induction of IL-8 and IP-10 by HA fragments in the epithelial cell line A549 (data not shown).

To further delineate the effects of HA on IL-8 and IP-10 expression by epithelial cells, we stimulated H292 cells with HA fragments (100 μg/ml) for varying time intervals and cell culture supernatants were harvested. Figure 2 shows that little IL-8 or IP-10 protein was present in the conditioned medium from unstimulated cells. We found peak protein secretion at 15–18 h for both IL-8 and IP-10 (Fig. 2, A and B). Additionally, dose-response relationships for HA fragment induction of IL-8 and IP-10 gene expression in H292 cells showed increasing IL-8 and IP-10 protein secretion with increasing doses of HA fragments, with maximal chemokine expression at 1,000 μg/ml HA for 18 h (Fig. 2, C and D). Thus HA fragments induce IL-8 and IP-10 protein in epithelial cells in a time- and dose-dependent fashion. These data clearly establish a role for the extracellular matrix, in the form of HA fragments produced at sites of inflammation, in the upregulation of epithelial cell chemokine production.

Induction of IL-8 and IP-10 in epithelial cells is specific to HA fragments.

Previously we found (14, 15, 29) that only low-molecular-weight HA and not its high-molecular-weight precursor or other glycosaminoglycans induced inflammatory mediator expression in macrophages. To determine whether the induction of chemokines by epithelial cells was specific to low-molecular-weight HA fragments, we stimulated H292 cells with HA fragments as well as other glycosaminoglycans. Figure 3 shows that only HA fragments (0.2 × 10⁶ D) but not HA disaccharides, HMW HA (mol wt 6 × 10⁶), CSA, CSB, or heparan induced IL-8 and IP-10 protein by ELISA. Additionally, HA fragments further digested with hyaluronidase failed to induced chemokine gene expression in H292 cells (data not shown). Thus the induction of IL-8 and IP-10 gene expression in epithelial cells is specific to low-molecular-weight HA fragments.

HA fragments induce CXC chemokines in epithelial cells through NF-κB and AP-1.

Many of the HA-inducible inflammatory genes including IL-8 and IP-10 have both NF-κB and AP-1 transcriptional binding sites in the 5′ regulatory regions of their promoters (3, 36, 40). Indeed, our laboratory has already demonstrated (13, 33) that, in alveolar macrophages, HA fragment stimulation increases NF-κB DNA binding and upregulates inducible NOS (iNOS) and monokine induced by IFN-γ (MIG) gene expression. To begin to identify the trans-acting factors utilized by HA in inducing chemokines in epithelial cells, we examined HA fragment-induced IP-10 induction in the presence of inhibitors of various signaling pathways. As shown in Fig. 4B, HA fragments alone stimulated a marked induction of IP-10 expression in H292 cells; however, this induction was almost completely inhibited by PS-1 (an inhibitor of the NF-κB activation pathway) but not by the MEK inhibitor PD-98059, suggesting a possible role for NF-κB (P values: HA vs. HA + PS-1 = 0.0001, HA vs. HA + PD-98059 = 0.0700, HA vs. HA + wortmannin = 0.6541). In contrast to IP-10, the MEK inhibitor PD-98059 significantly inhibited HA fragment-induced IL-8 expression in H292 cells, suggesting a role for MEK (Fig. 4A) (P values: HA vs. HA + PS-1 = 0.0039, HA vs. HA + PD-98059 = 0.0039, HA vs. HA + wortmannin = 0.124). NF-κB does not appear to be important in HA fragment-induced IL-8 as PS-1 treatment led to the superinduction of IL-8. These data were not due to nonspecific inhibition as the phosphatidylinositol 3-kinase inhibitor wortmannin did not inhibit either HA fragment-induced chemokine expression. Nor was the effect due to cell death, as cell adhesion and viability were similar in all of the conditions. This effect is not exclusive to H292 cells, as a similar pattern of inhibition was demonstrated in BEAS-2B cells (Fig. 4, C and D). Of note, the p38 inhibitor SB-220025 and JNK inhibitor I (L-form, cell permeant) did not inhibit HA induced IL-8 or IP-10 in either cell line (data not shown). Thus the CXC chemokines IP-10 and IP-8 appear to be induced by HA fragments in epithelial cells by different signaling pathways, namely, the NF-κB and ERK-MAP-kinase pathways, respectively.

HA fragments induce c-jun/c-fos/AP-1 and NF-κB p50/p65 heterodimers in airway epithelial cells.

Because inhibiting MAP kinase or NF-κB activity prevented HA fragment-induced IL-8 or IP-10, respectively, we performed EMSA on nuclear extracts from H292 epithelial cells stimulated with HA fragments for 1 h, using ³²P-radiolabeled DNA probes with consensus NF-κB or AP-1 sites. As shown in Fig. 5, HA fragments induce the upregulation of a protein-DNA complex that is competed away by cold consensus NF-κB but not AP-1. Furthermore, supershifts with antibodies to p50 and p65 reveal that this protein contains p50/p65 NF-κB heterodimers. Of note, the p50 antibody decreases binding of the p50 to the DNA probe, resulting in decreased intensity of the p50-DNA band in lieu of shifting.

Similar EMSAs were performed with a consensus AP-1 DNA probe with the same nuclear extracts. As seen in Fig. 5, HA fragments induce binding of a complex to the consensus AP-1 probe that is competed away by cold consensus AP-1 probe but not cold consensus NF-κB. Additionally, supershifts with antibodies to the AP-1 proteins c-jun and c-fos reveal that the protein binding to the probe contains c-jun and c-fos; again, these antibodies inhibit protein-probe binding, resulting in decreased intensity of the DNA-protein bands of interest.

Loss of AP-1 and NF-κB sites on IL-8 and IP-10 promoters, respectively, inhibits HA fragment-induced gene expression.

To determine the functional significance of HA fragment-induced AP-1 and NF-κB proteins and define the cis-acting regions of these promoters important for HA fragment-induced signaling, we preformed transient transfections of H292 epithelial cells with IL-8 and IP-10 promoter constructs, using firefly luciferase reporters. Reporter constructs containing 161 bp of the 5′ IL-8 promoter ± mutations in the AP-1 (−120) or NF-κB (−82) binding sites were transiently transfected into epithelial cells. HA fragments induced a four- to fivefold increase in luciferase expression in the reporter constructs containing −161 bp of the IL-8 promoter (Fig. 6A). However, mutation of the −120 AP-1 site but not the −82 NF-κB site inhibited HA fragment-induced luciferase activity (Fig. 6A) (P values: IL-8 vs. IL-8 NF-κB mutant = 0.5601, IL-8 vs. IL-8 AP-1 mutant = 0.039, IL-8 NF-κB mutant vs. IL-8 AP-1 mutant = 0.007). Thus, consistent with our inhibitor data, HA fragments require AP-1 to induce IL-8 gene expression (Fig. 4A). As a positive control, transfected cells were also stimulated with TNF-α (10 ng/ml). As expected, TNF-α-induced luciferase activity was inhibited only in the IL-8 NF-κB mutant (IL-8 vs. IL-8 NF-κB mutant, P = 0.04; Fig. 6A).

Similarly, we preformed transient transfections of a series of nested deletions of the IP-10 promoter upstream of a luciferase reporter construct to determine the cis-acting sequences required for HA fragment-induced IP-10 expression. Whereas the constructs containing −533 to −190 bp of the IP-10 promoter demonstrated a three- to fivefold induction of luciferase in response to HA fragments, the smaller −161 and −96 bp promoter constructs resulted in loss of HA fragment-induced reporter activity (Fig. 6B) (HA −533 vs. −322, −533 vs. −237, and −533 vs. −190: P > 0.4; HA −533, −322, −237, and −190 vs. −161: P < 0.01; HA −533, −322, −237, and −190 vs. −96: P < 0.005; HA −161 vs. −96: P = 0.366). This is most likely due to the loss of the −169 NF-κB site and not the −113 NF-κB or −73 AP-1 sites. Indeed, in the small −161 and −96 promoter constructs, the distal AP-1 site is still functional as it demonstrates a three- to fourfold induction of activity in response to PMA stimulation (Fig. 6B) (P values: −161 HA vs. −161 PMA = 0.0122, −96 HA vs. −96 PMA = 0.0001). Additionally, HA-induced IP-10 promoter activity is not dependent on the IFN-stimulated responsive element (ISRE), as the −190 construct, which eliminates the ISRE, still responds to HA stimulation (P values: −237 vs. −190 = 0.6720, −190 vs. −161 = 0.0056). Thus IP-10 and IL-8 appear to be differentially induced by HA fragments in epithelial cells via the NF-κB and AP-1 pathways, respectively.

HMW HA inhibits HA fragment induced IL-8 and IP-10 in airway epithelial cells.

Although biologically inert, HMW HA can act as an inhibitor of lower-molecular-weight HA fragment signaling (29). Preincubation of H292 cells for 1 h with HMW HA (500 μg/ml) significantly blocked HA fragment-induced IL-8 and IP-10 expression by 40% (Fig. 7; HA fragments vs. HA fragments + HMW: IL-8 P = 0.0001, IP-10: P = 0.0001). The inhibition of HA fragment-induced IL-8 and IP-10 by HMW HA also held true over a range of concentrations of (data not shown). The lack of complete inhibition of HA fragment-induced chemokine expression by HMW HA may be accounted for, on a mole-to-mole basis, by the excess of HA fragments (HMW HA has mol wt of 6 × 10⁶, whereas HA fragments are 0.2 × 10⁶ in size).

DISCUSSION

In addition to its many homeostatic functions such as barrier protection and mucocilliary clearance, airway epithelium plays an important role in regulating local inflammation and immune responses (9). The airway epithelium participates in inflammatory responses in part through the generation of numerous cytokines and chemokines that mediate recruitment and activation of inflammatory cells. The epithelium is involved in local cytokine networks allowing first response to noxious agents as well as cross talk with structural cells to help provide an effective inflammatory response (9). Although these bidirectional epithelium-driven inflammatory communications are appropriate for normal host defense, they also feature prominently in the pathogenesis of chronic bronchitis and other inflammatory lung disorders. In this report, we demonstrate for the first time that fragments of the extracellular matrix component HA can differentially induce inflammatory chemokine production in primary airway epithelial cells. Although other investigators in a model of ventilator-induced lung injury have demonstrated that HA fragments, especially when combined with mechanical stretch, induced IL-8 in the A549 airway epithelium-like cell line, the molecular mechanism remained undefined (27). We have determined that the molecular pathways responsible for the HA-induced upregulation of the chemokines IL-8 and IP-10 are very different. That is, HA fragments induce IL-8 via AP-1 activation and independent of NF-κB, whereas IP-10 is dependent on the NF-κB pathway and independent of the MAP kinase AP-1 pathway.

We previously demonstrated (14–16, 29, 30) that fragments of HA induce the expression of a wide array of inflammatory genes such as chemokines, cytokines, metalloelastase, and NOS in alveolar macrophages. In this report we extend these findings to airway epithelial cells. In macrophages, HA fragments induce steady-state levels of mRNA for the chemokines macrophage inflammatory protein (MIP)-1α, MIP-1β, RANTES, monocyte chemoattractant protein (MCP)-1, and IP-10 and the cytokine TNF-α (14–16, 29, 30). In addition, fragment size is important in signaling, as fragments larger than 500,000 or smaller than six sugars do not signal (29). As described here, these size restrictions also hold true for HA-induced stimulation of epithelial cells (Fig. 3). Importantly, these results are not due to LPS contamination, as H292 cells do not respond to LPS (Fig. 1).

In terms of signaling, our previous studies (13, 33) showed that HA fragments bind to an as yet undefined cell surface receptor and activates the NF-κB family of transcription factors. Numerous investigators have reported that HA may be signaling via one or more of the Toll-like receptors (7, 18, 43). Indeed, in macrophages, the induction of MIP-1α, MIP-1β, RANTES, MCP-1, IP-10, and TNF-α are all NF-κB dependent. Although our macrophage data are consistent with a role for NF-κB in HA fragment-induced IP-10 in airway epithelial cells, they are in contrast with HA-induced IL-8 production in epithelial cells that utilizes the MAP kinase pathway (Figs. 4–6). In fact, inhibition of NF-κB activation by incubation of epithelial cells with the proteasome inhibitor PS-1 led to the superinduction of IL-8 (Fig. 4). Such data are consistent with other reports showing that proteasomal inhibition leads to the increased production and stability of IL-8 expression (8, 19). In contrast to our findings, it is clear that under certain conditions IL-8 transcription is in fact NF-κB dependent. For example, Pseudomonas-induced IL-8 in airway epithelium is dependent predominantly on NF-κB activity, whereas respiratory syncytial virus-infected airway epithelial cells rely on a cooperative binding of AP-1 and NF-κB sites (28, 31). Additionally, the transcription factors NF-κB, NF-IL-6, and AP-1 are all necessary for IL-8 expression in airway epithelial cells by ozone (17). On the other hand, consistent with our data (Figs. 4–6), IP-10 transcription has only been associated with the NF-κB pathway (1, 20).

Whereas both IL-8 and IP-10 have been associated with chronic airway inflammation, their respective pathophysiological roles are unclear. IL-8 is a potent chemoattractant for neutrophils, and IP-10 attracts monocytes and lymphocytes. In many inflammatory disorders, such as chronic bronchitis, there is an accumulation of neutrophils in the lumen and monocytes in the airway wall (12). In addition to being chemoattractants, both IL-8 and IP-10 are thought to play a role in angiogenesis. IL-8 has been shown to be proangiogenic, whereas IP-10 has angiostatic properties (21–23). In this regard, it has been proposed that for pulmonary fibrosis IL-8 promotes angiogenesis and subsequent fibrosis whereas IP-10 counteracts this effect (21–23). Inasmuch as the induction of IL-8 and IP-10 by HA fragments is by distinct biochemical pathways, our data suggest that MAP kinase inhibitors might be able to mitigate inflammation by preferentially inhibiting HA fragment-induced IL-8 production while allowing for the production of IP-10.

In the setting of lung injury/inflammation, extracellular matrix destruction and turnover is part of the normal repair process. With normal healing, the lower-molecular-weight active HA fragments generated at sites of injury are further broken down and removed, allowing resolution of the inflammation. In the case of chronic insults, we propose that the extracellular matrix is not only the target of inflammation but the accumulation of lower-molecular-weight HA fragments serves to perpetuate dysregulated, unremitting inflammation. The persistent HA fragment generation further propagates and amplifies the inflammatory process by continually stimulating inflammatory gene expression by epithelial cells. Furthermore, exciting new data by Jiang et al. (18) have strongly implicated HA fragments as being of paramount importance in a model of bleomycin-induced lung injury. Additionally, this raises the possibility of high-molecular-weight HA as a therapy via competitive inhibition of the active low-molecular-weight fragments (Fig. 7). In fact, Cantor et al. (4, 5) have already demonstrated in a hamster model of elastase-induced emphysema that coadministration of HA ameliorates emphysema formation.

GRANTS

This work was supported by grants from the National Heart, Lung, and Blood Institute (K08-HL-039932 and RO1-HL-0961402) and the Flight Attendant Medical Research Institute (M. R. Horton).

FOOTNOTES

• The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.