SEARCH

SEARCH BY CITATION

Keywords:

  • extracellular matrix;
  • glycosaminoglycan;
  • lung injury

Abstract

  1. Top of page
  2. Abstract
  3. LITERATURE CITED

An important hallmark of tissue remodeling is the dynamic turnover of extracellular matrix (ECM). ECM performs a variety of functions in tissue repair including scaffold formation, modulation of fluid dynamics, and regulating cell behavior. During non-infectious tissue injury ECM degradation products are generated that acquire signaling functions not attributable to the native precursor molecules. Hyaluronan (HA) is a non-sulfated glycosaminoglycan which is produced in great abundance following tissue injury. It exists both in a soluble form and as side chains on proteoglycans. HA has critical roles in development as well as a variety of biological processes including wound healing, tumor growth and metastasis, and inflammation. HA fragments share structural similarities with pathogens and following tissue injury can be recognized by innate immune receptors. Elucidating the protean roles of HA in tissue injury, inflammation, and repair will generate new insights into mechanisms of diseases characterized by chronic inflammation and tissue remodeling. Anat Rec, 2010. © 2010 Wiley-Liss, Inc.

Tissue injury and remodeling occurs in a variety of settings that can involve infection, inflammation, and autoimmunity. The mechanisms that regulate tissue injury and repair following infection have become well understood, but the critical elements involved in determining the host responses to non-infectious or sterile inflammation remain unclear. Our laboratory has been interested in elucidating the mechanisms that contribute to lung injury, inflammation, and repair following non-infectious fibrotic lung injury. Tissue fibrosis is a leading cause of morbidity and mortality, and fibrotic lung diseases represent a major area of unmet medical need. Fibrosing lung diseases that occur in the absence of recognized inciting agents represent areas of investigation in need of new insights into both pathogenesis and determinants of progression. The most serious disorder of lung fibrosis is idiopathic pulmonary fibrosis (IPF) (Noble and Homer, 2004; Noble, 2006). IPF is a disorder characterized by unremitting deposition of components of the extracellular matrix (ECM) in the interstitium of the lung leading to destruction of gas-exchanging regions of the lung and ultimate suffocation. The pattern of deposition of ECM is unique in IPF and a number of components of the ECM accumulate including collagens, proteoglycans, fibronections, and the non-sulfated glycosaminoglycan hyaluronan (HA). Unremitting fibrosis involves the interactions between matrix components with mesenchymal cells leading to proliferation, tissue destruction, and further production of matrix.

One of the hallmarks of non-infectious lung injury is production of components of the ECM, and in recent years it has become evident that in addition to provide a substrate for cell migration, matrix can directly influence cell effector functions (Noble, 2002; Jiang et al., 2007). Our laboratory has focused on the role of HA in the pathobiology of non-infectious lung injury for a number of reasons. First and foremost, HA is produced in great abundance following tissue injury. The classic model for investigating non-infectious fibrotic lung injury for decades has been the instillation of bleomycin into the lung (Nettelbladt et al., 1989; Bray et al., 1991; Teder et al., 2002; Jiang et al., 2005). Bleomycin causes direct injury to the lung epithelium and orchestrates a sequence of events characterized by the influx of inflammatory cells, the resolution of inflammation and the development of lung fibrosis that is limited. While the model does not depict essential features of IPF, it is useful in testing paradigms for resolving and perpetuating fibrosis. Following bleomycin injury there is a massive accumulation of HA in both the alveolar spaces and interstitium of the lung (Teder et al., 2002) (Fig. 1). Our laboratory has been interested in understanding the significance of this accumulation of matrix. The vast majority of HA is cleared from the lung within 14 days after injury and that is when the preponderance of collagen deposition ensues (Teder et al., 2002). Interestingly, fibroblasts are the main source of both HA and collagen, and under physiologic conditions HA production ceases as collagen production increases. HA is non-sulfated polymer made up of repeating units of D-glucuronic acid and N-acetyl-glucosamine (Jiang et al., 2007). Under physiologic conditions in the unchallenged lung, HA exists as a very large polymer in excess of one million Daltons. Following tissue injury, there is an accumulation of HA degradation products that subsequently cleared from the normal lung within 14 days of injury. Interestingly, the accumulation of HA fragments coincides with the peak inflammatory response. We have been interested in the biological significance of HA fragment accumulation, particularly since HA has been shown to accumulate in fibrosing lung diseases. To begin to understand the potential significance of HA fragment accumulation, we began to explore the effects of HA fragments on macrophage functions in vitro (Jiang et al., 2005). HA has several cell surface receptors including CD44 and RHAMM, and we have been interested in the role of CD44 in regulating HA interactions with cells involved in lung disease pathogenesis. CD44 is highly expressed on lung macrophages and ligation of CD44 has been shown to result in the release of inflammatory mediators. Moreover, Savani and colleagues showed that the interaction of RHAMM with HA is important in recruitment of macrophages to the lung after bleomycin induced lung injury (Zaman et al., 2005). We found that HA fragments produced very different effects on macrophages than high molecular weight HA (Noble et al., 1996; Horton et al., 1998). HA fragments induced a variety of inflammatory mediators that have recognized functions in tissue injury, inflammation, and repair. In particular, we found that HA fragments induced the expression of a number of chemokines in macrophages (Noble et al., 1996; Horton et al., 1998; Jiang et al., 2005). Chemokines are critical mediators of inflammatory cell recruitment to sites of tissue injury. In addition, we found that HA fragments induced the activation of a critical regulator of innate immune responses, the transcriptional regulator NF-κB (Noble et al., 1996). This was the first demonstration that a matrix component, modified by the inflammatory milieu could active a regulatory system that senses host invasion by infectious agents. With these data we began to explore the concept that HA fragments serve the purpose of signaling that the host has been intruded upon and may function in analogous role in non-infectious injury that bacteria do under infectious conditions.

thumbnail image

Figure 1. Following bleomycin injury there is a massive accumulation of hyaluronan in both the alveolar spaces and interstitium of the lung. Lung tissue stained for hyaluronan with biotinylated hyaluronan binding protein at day 14 after bleomycin injury. Wild type (left panel) and CD44-deficient (right panel) mice. Adapted from Teder et al., 2002.

Download figure to PowerPoint

Having discovered that HA fragments could stimulate inflammatory macrophages to produce mediators of host repair, we sought to better define the role of CD44 in regulating HA functions. We had generated data that CD44 could mediate some aspects of HA fragment signaling such as TNFα and IGF-1 production, but not others such as MMP-12 (Noble et al., 1993). To explore the role of CD44 both in vitro and in vivo, we took advantage of CD44 null mice. These mice develop without obvious impairment and breed normally (Schmits et al., 1997). We found that macrophages from CD44 null mice still responded to HA fragments suggesting an alternative cell surface recognition system was involved. However, when we challenged the CD44 null mice with bleomycin we found that they had increased susceptibility to non-infectious lung injury (Teder et al., 2002). We explored this phenotype in great detail and determined that there was a marked impairment in the ability of the lung to resolve the inflammatory response. Furthermore, HA fragments were not cleared from the lung after injury. This impaired clearance resulted in unremitting inflammation that compromised the host. Since CD44 is expressed on all cells, we generated bone marrow chimeras to determine if hematopoietic CD44 was responsible for the inflammatory phenotype. Mice that expressed CD44 in bone marrow derived cells but not structural cells were able to clear HA fragments and resolve the inflammatory response (Teder et al., 2002). It was evident from these studies that HA and CD44 were of fundamental importance in mediating the host response to non-infectious injury. It was also evident that there must be another recognition system on macrophages that recognized HA fragments.

The clues to the recognition system were in the structure of the HA polymer. HA is a repeating pattern of disaccharides and innate immune cells recognize pathogens through pattern recognition receptors. The cell surface of gram positive organisms contain HA and the side chains of gram-negative organisms also have structural similarities, so we wondered if HA fragments could function like an infectious agent when generated in the context of the inflammatory response. The best studied group of pattern recognition receptors are the Toll-like receptors (TLRs) (Medzhitov and Janeway, 2000; Akira and Takeda, 2004). We had been using macrophages from C3H/HeJ mice, for our in vitro studies, to avoid the effects of contaminating endotoxins. C3H/HeJ mice are defective in TLR4 signaling (Poltorak et al., 1998), and so we were skeptical that TLR4 would be the key TLR involved in HA recognition. The first step in determining if TLRs might be involved was to utilize MyD88 null macrophages. MyD88 is an important adaptor for most but not all TLR agonist signaling (Medzhitov and Janeway, 2000; Akira and Takeda, 2004). We were excited to find out that MyD88 null macrophages did not respond effectively to HA fragments (Jiang et al., 2005). This directly indicted the TLR system in HA recognition. We then tested macrophages from TLR1–5 and TLR9 null mice and found that TLR4 and TLR2 null macrophages had reduced response to HA fragments, and all other TLRs tested still had similar response to HA fragments as wild type. This was discouraging but we returned to our concept that perhaps both gram-positive (TLR2) and gram-negative (TLR4) recognition systems were necessary. We generated TLR2/TLR4 double knockout mice and stimulated macrophages with HA fragments. Macrophages from the TLR2/TLR4 double knockout mice failed to respond to HA fragments (Jiang et al., 2005) (Fig. 2). These data strongly implicated the innate immune system in the recognition of HA fragments. HA-TLR interactions have been demonstrated in dendritic cells (Termeer et al., 2002), macrophages (Taylor et al., 2007), and microvascular endothelial cells (Taylor et al., 2004). Recently, a study by Gallo and associates showed that HA fragments interact with a receptor complex including TLR4, CD44, and MD-2 in non-infectious injury (Taylor et al., 2007). This is analogous to LPS-TLR-CD14-MD2 interactions in infectious injury (Medzhitov and Janeway, 2000).

thumbnail image

Figure 2. Hyaluronan fragments stimulate chemokine expression through both TLR4 and TLR2. A: Cxcl2 mRNA expression by elicited peritoneal macrophages from wild type (wt), TLR2−/−, TLR4−/−, or TLR2−/−4−/− mice treated with HA fragments or LPS in the presence or absence (underlined) of polymyxin B, detected by Northern analysis. B: TNFα protein expression by elicited peritoneal macrophages from wt or TLR2−/−4−/− treated with the indicated concentration of HA in the presence of polymyxin B for 24 hr. Adapted from Jiang et al., 2005.

Download figure to PowerPoint

Having determined that HA fragments were a component of the innate immune response under conditions of sterile inflammation, we examined the role of TLR signaling in the injury, inflammation, and repair response in vivo. We challenged TLR2/TLR4 null mice with bleomycin and found that they had increased susceptibility to non-infectious lung injury (Fig. 3). This was surprising since macrophages from TLR2/TLR4 null mice did not produce inflammatory mediators in response to HA fragment in vitro (Jiang et al., 2005). We examined the phenotype in more detail and observed that although mortality was increased, there was actually a decrease in neutrophil recruitment in the absence of TLR2 and TLR4. This presented a conundrum and clues to explain this disconnect between an inhibition of inflammation and an increase in tissue injury were revealed by examining the lung parenchyma in the TLR2/TLR4 null mice after injury. Much to our surprise there was clear evidence of increased tissue damage in the TLR2/TLR4 null mice after injury (Jiang et al., 2005). We then examined the lung epithelial cells and found that there was evidence of increased apoptosis in the TLR2/TLR4 null mice after bleomycin treatment (Jiang et al., 2005). These data suggested that somehow TLR2/TLR4 were protective from acute lung injury due to a non-infectious insult. We explored this further and discovered that HA was on the cell surface of primary lung epithelial cells from wild type mice but this cell surface expression was significantly diminished in lung epithelial cells from TLR2/TLR4 null mice. The implication was that cell surface HA was protective against acute lung injury. To test this hypothesis in another way, we generated transgenic mice that express hyaluronan synthase 2 (HAS2) specifically on lung epithelial cells using the CCSP (CC10) promoter. We found that increasing lung epithelial cell surface HA afforded increased protection against acute lung injury (Jiang et al., 2005) (Fig. 4). Furthermore, when we isolated primary lung epithelial cells from TLR2/TLR4 null mice, we found an increased susceptibility to bleomycin-induced apoptosis and high molecular weight HA was protective against acute lung injury in a TLR dependent manner (Jiang et al., 2005). Moreover, Salathe and colleagues demonstrated that HA binding to RHAMM stimulates ciliary beating to play a role in airway mucosal host defense (Forteza et al., 2001).

thumbnail image

Figure 3. TLR2/TLR4 null mice had increased susceptibility to non-infectious lung injury. Wild type and TLR2−/−4−/− mice were subjected to intratracheal bleomycin treatment. Percentages of surviving animals were plotted over a 21-day period. Adapted from Jiang et al., 2005.

Download figure to PowerPoint

thumbnail image

Figure 4. Overexpression of high molecular mass HA ameliorates lung injury in CC10-HAS2 transgenic mice. CC10-HAS2 transgenic mice have improved survival following high dose of bleomycin treatment compared with littermate controls. Adapted from Jiang et al., 2005.

Download figure to PowerPoint

HA and HA receptors appear to have protean functions in lung injury inflammation and repair. These are summarized in a schematic published as an editorial (O'Neill, 2005) on our studies (Jiang et al., 2005). Soluble HA fragments generated during inflammation stimulate macrophages to produce mediators that are focused on repairing the lung. However, HA on the epithelial cell surface serves a protective function by engaging TLR2 and TLR4 to produce low levels of NF-κB activation that are protective against exogenous insults (O'Neill, 2005). This is analogous to the role commensal gut flora function in protecting gut epithelium form injury (Rakoff-Nahoum et al., 2004). HA continues to be a remarkably interesting molecule that has important roles in a variety of biological properties. HA is produced in great abundance in the interstitium of the lung by mesenchymal cells, and it will be exciting to explore the functional significance of this in non-infectious lung injury, inflammation, and repair.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. LITERATURE CITED