| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Division of Dermatology, University of California, San Diego and VA Medical Center, San Diego, California, USA
1Correspondence: Mail Code 9111B, 3350 La Jolla Village Dr., San Diego, CA 92161, USA. E-mail: rgallo{at}vapop.ucsd.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGLYCOSAMINOGLYCAN STRUCTUREHEPARAN SULFATE/HEPARIN AND...CHONDROITIN/DERMATAN SULFATE AND...HYALURONAN AND INFLAMMATIONCONCLUSIONSREFERENCES |
|---|
Key Words: heparan sulfate • chondroitin sulfate • dermatan sulfate • hyaluronan
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGLYCOSAMINOGLYCAN STRUCTUREHEPARAN SULFATE/HEPARIN AND...CHONDROITIN/DERMATAN SULFATE AND...HYALURONAN AND INFLAMMATIONCONCLUSIONSREFERENCES |
|---|
. However, advances in glycobiology have made apparent that the active participation of GAGs, and their proteoglycan backbone, is important in a variety of cell communication events. Among the myriad of cell behaviors influenced by these molecules, recent data have shown GAGs initiate and control events associated with inflammation (2
, 3)
. Examples of processes influenced by GAGs include cytokine/chemokine production, leukocyte recruitment, and inflammatory cell maturation (4
5
6)
. In such settings GAGs released after injury function similarly to pathogen associated molecular patterns (PAMPs), and serve as a molecular signal of host injury in the absence of microbial involvement. Work with direct clinical relevance has suggested that specific GAGs, and their low molecular weight fragments, are elevated in patients with osteoarthritis, rheumatoid arthritis, psoriasis, scleroderma, and inflammatory bowel disease, and may adversely influence the course of these disorders. Conversely, other observations have led to models proposing that increased levels of GAG can lead to a decrease in the inflammatory response. Reports of the participation of GAGs as both pro-inflammatory and anti-inflammatory mediators have led to significant progress in understanding the mechanism by which GAGs participate in inflammation. Such progress includes how GAG synthesis enzymes participate in encoding information within the linear carbohydrate chain that defines the GAG and how this translates into function. Overall, it is now apparent that the presence of GAGs must be recognized when considering complex and coordinated steps of inflammation and tissue repair. This review will discuss these aspects of GAG biology with specific emphasis on important observations relevant to understanding inflammatory processes in human disease. |
GLYCOSAMINOGLYCAN STRUCTURE |
|---|
|
TOPABSTRACTINTRODUCTIONGLYCOSAMINOGLYCAN STRUCTUREHEPARAN SULFATE/HEPARIN AND...CHONDROITIN/DERMATAN SULFATE AND...HYALURONAN AND INFLAMMATIONCONCLUSIONSREFERENCES |
|---|
Four structurally distinct GAG families exist: heparan sulfate (HS)/heparin, chondroitin (CS)/dermatan sulfate (DS), keratan sulfate (KS), and hyaluronan (HA). GAGs are linear polysaccharides, comprised of repeating disaccharide units of an amino sugar, either N-acetyl-D-glucosamine (D-GlcNAc) or N-acetyl-D-galactosamine (D-GalNAc), and an uronic acid, either D-glucuronic (D-GlcA) or L-iduronic acid (L-IdoA) (Fig. 1
), except for KS, which consists of a galactose in place of the uronic acid. For simplification purposes, the absolute configurations of the amino sugars and uronic acids as stated above will be abbreviated as GlcNAc, GalNAc, GlcA, and IdoA. HS and CS/DS, but not HA, are assembled via a serine residue to protein cores, delineating them as proteoglycans (PG). While slight differences exist in the basic sugar backbone of the GAG chain, subsequent sulfation, deacetylation and epimerization modifications distinguish individual GAG chains and are critical for their specific activity. Sulfation plays a particularly important role in HS and CS/DS activity. In contrast, HA is not modified by sulfation or epimerization. Due to the way GAG chains are constructed and modified, incredible diversity in GAG chains can be achieved where no two chains are exactly alike.
|
HS and CS/DS share many similarities, including their initial tetrasaccharide linkage sequence to PG protein cores (Fig. 1)
. GAG chain synthesis on a PG occurs at a serine residue and is initialized by xylosyltransferase activity and the addition of a xylose (Xyl) in the ER (7)
. Two galactose residues are subsequently added in the cis/medial Golgi to the Xyl by galactosyltransferase I and galactosyltransferase II (7)
. The fourth residue is a GlcA added by glucuronyltransferase I and occurs in the trans Golgi (7)
. At this point in chain synthesis, there is no distinguishing whether this chain will become a HS or CS/DS GAG, as these enzymes have been shown to be critical for initiating chain formation for both of these GAG families (7
, 8)
. What determines the identity of the GAG is which enzyme adds the next sugar residue; the addition of an 
-GlcNAc by -N-acetylglucosaminyltransferase I delineates the chain to be of the HS family, but addition of a ß-GalNAc by ß-N-acetylgalactosaminyltransferase I distinguishes the chain as a member of the CS/DS family (7
, 8)
. Although no definitive code has been determined to dictate the type of GAG produced, it has been proposed that modifications, such as sulfation or phosphorylation, within the linkage region may influence GAG choice (9)
. The addition of the GlcNAc or GalNAc residue and subsequent modifications of GAG chains on PG occurs in the trans network of the Golgi (7)
.
HS consists of repeating units of GlcNAc and GlcA that are assembled to the nonreducing end of the GAG by HS copolymerases, exostosin 1 (EXT1) and exostosin 2 (EXT2) (8
, 10
11
12
13)
. The role of EXT1 and EXT2 was first described as possible tumor suppressor genes and the loss of EXT1 or EXT2 was found to result in bony exostoses, or cartilage-capped tumors, in long bones (14)
. Subsequent studies have shown that EXT1 is an ER resident type II transmembrane glycoprotein involved in the synthesis of HS and mutations in these genes result in the disorder Hereditary Multiple Exostosis (15
, 16)
. Modifications on HS include N-deacetylation and N-sulfation of the glucosamine, epimerization of the GlcA to IdoA, 2-O-sulfation of the uronic acid (usually IdoA), and 6-O-sulfation and 3-O-sulfation of the glucosamine (8)
. Heparin is distinguished from HS by being more heavily modified, although HS can contain heparin-like regions. Heparin is much less abundant in vivo than HS and exists primarily in mast cells (3)
. In addition, the importance of sufficient sulfate pools within the cell for GAG chain assembly has been shown by Hastbacka et al. (17)
. A defect in the sulfate transporter DTDST (also referred to as SLC26A2 or CSF1R) leads to undersulfated GAGs, especially affecting bone and joints, and is responsible for the osteochondrodysplasia disease Diastrophic dysplasia (DTD) (17)
. Recently a mouse defective in DTDST has also been described with a similar phenotype to that of humans with DTD (18)
.
Modification reactions on the HS chain occur while the chain is being polymerized and begins with the N-deacetylase/N-sulfotransferase (NDST) enzymes that remove the N-acetyl groups and replace them with sulfate groups. This modification is a prerequisite for all subsequent modifications (8)
. Studies in Drosophila have revealed an early NDST gene (sulfateless) that is necessary for proper assembly of the HSPG dally. Dally participates in the signaling of a Wnt family member (wingless) and its receptor Frizzled 2, providing an example of how a NDST modifies biological activity (19)
. Four NDST’s have been described (NDST1, 2, 3, and 4) in mice and humans, each has slightly different functional abilities and arise from four different genes (8
, 20)
. Mice deficient in NDST1 contain low-sulfated sugar residues and die shortly after birth due to lung complications (21)
. NDST2 knockout mice are viable, produce HS but not heparin (22
, 23)
(Table 1
). An example of the specificity and activity of GAG chains was shown by Humphries et al. and their study of the PG serglycin in mast cells (22)
. Serglycin can possess either heparin or CS chains that are critical to their ability to store granule proteases. The absence of heparin chains on serglycin, due to the absence of NDST-2, results in the inability to properly store two chymases and an exopeptidase. This provides evidence that it is the specificity of the GAG chain, and not just a negatively charged carbohydrate chain, that is required for proper mast cell function (22
, 24)
. In addition, the serglycin knockout mouse is viable and fertile but exhibits severe abnormalities in mast cell protease storage and overall function (24)
.
|
Epimerization of the C5 carbon of GlcA results in the formation of IdoA and one C5 epimerase has been described for HS/heparin (8)
. Deficiency of the C5 epimerase is embryonic lethal and deficient mice display abnormalities in kidney and lung development (25)
. The activity of the HS 2-O-sulfotransferase (HS2ST) is also specific for HS/heparin (8
, 20)
. Knockout mice deficient in HS2ST die neonatally due to kidney development failure (26
27
28
29)
. In addition, it has been proposed that the C5 epimerase and HS2ST interact with one another or form a complex, possibly for sulfating the IdoA so that it is unable to epimerize back to GlcA (20)
. HS 6-O-sulfotransferases add 6-O-sulfates to N-sulfated glucosamines and three isoforms of this enzyme have been described (HS6ST1, 2, and 3) (8
, 20)
. Each enzyme has slightly different substrate specificities and expression patterns. Finally, the most unique sulfation modification of HS takes place with addition of a 3-O-sulfate to the glucosamine residue by HS 3-O-sulfotransferases. Six different isoforms of this enzyme have been described (HS3ST1, 2, 3A, 3B, 4, and 5) (8
, 20)
. As with the HS6STs, each HS3ST prefers specific substrate sequences and have been suggested to be involved in producing HS sequences important in specific biological processes. For example, HS3ST1 and HS3ST5 have been suggested to add on 3-O-sulfates important in modifying the HS chain to make anticoagulant HS (8)
. However, HS3ST1-deficient mice are viable and exhibit no obvious procoagulant phenotype (30
, 31)
. Shulka et al. have described a role for HS3ST3 (mouse and human isoforms 3-OST-3B and human isoforms of 3-OST-3A) in herpes simplex virus-1 (HSV-1) entry into host cells. Using a resistant cell line, the authors determined that the cells became susceptible to HSV-1 when the HS3ST3 enzyme modified the cell surface HS chains, resulting in the ability of the HSV-1 gD glycoprotein to mediate viral entry (32)
.
CS is distinct from HS and is composed of repeating disaccharide units of GalNAc and GlcA, added on to the nonreducing end of the growing CS chain (7)
. Studies suggest that GlcA transferase and GalNAc transferase functions are on separate proteins although one protein, chondroitin synthase, appears to possess both activities (20)
. Modifications of the CS chain include epimerization of the C5 carbon of GlcA to form IdoA; this delineates the CS chain as DS, providing DS with a sugar modification similar to HS. In addition to epimerization, sulfation modifications may occur at the 2-position of the IdoA (and, to a lesser extent, GlcA) and at the 4- position or the 6-position of the GalNAc (7)
. Unlike HS, CS/DS do not undergo N-deacetylation on the GalNAc.
Modifications of the CS/DS chains occur while the chain is undergoing polymerization although the details of CS/DS biosynthesis have not been worked out as well as they have for HS. Early in chain modification, 6-O-sulfation occurs by the action of chondroitin 6-sulfotransferase (C6ST) onto nonsulfated GalNAc residues in the medial/trans Golgi region (7
, 20
, 33)
. Next, 4-O-sulfation of the GalNAc takes place in the later trans Golgi. However, this apparently can occur concurrently with epimerization to produce DS (7
, 20)
. There exists 4-sulfotransferases that differentiate between CS and DS (C4ST and D4ST), but the epimerase prefers 4-O-sulfated substrates, suggesting that the enzymes are working at the same time (7
, 20)
. Three C4STs have been described (C4ST1, 2 and 3) that prefer GlcA-rich regions (20)
. Alternatively, one D4ST has been identified that is specific for the IdoA regions present in DS (7)
. It has been suggested that D4ST works immediately after epimerization to prevent back-epimerization (20)
. Another enzyme, GalNAc4-6ST adds sulfates onto GalNAc already modified by C4ST (7
, 20
, 33)
.
Studies on the C6ST-deficient mouse describe a decrease in T cells in the spleen, indicating a role for C6ST in maintaining naïve T lymphocytes in the spleen (34)
. However, 6-O-sulfation on to 4-sulfated-GalNAc was unaffected due to the presence of the GalNAc4-6ST enzyme. A recent familial study on patients with no chondroitin 6-O-sulfotransferase-1 function describes a form of spondyloepiphyseal dysplasia with major spinal involvement (35)
. Unlike the C6ST knockout mice, these patients exhibited severe skeletal abnormalities. It remains unclear why such discrepancies exist between the deficient mice and human patients, but a better understanding of these enzymes, possible isoforms, and substrate specificities may help to elucidate these observations. In a related but distinct observation of the morphological impact of a 6-O-sulfotransferase deficiency, the disease Macular Corneal Dystrophy (MCD) types I and II was found to be a defect in CHST6, a 6-O-sulfotransferase responsible for the sulfation of the N-acetylglucosamine of keratan sulfate (36)
. A defect in CHST6 results in the improper sulfation of keratan sulfate in corneal cells, which is needed to maintain transparency of the cornea (36)
.
The final step in CS/DS synthesis is the addition of 2-sulfates on IdoA by the 2-O-sulfotransferase (CS/DS2ST). Only one CS/DS2ST enzyme has been identified and it prefers to add 2-sulfates to IdoA next to 4-sulfated GalNAc. Although a few reports suggest the possibility of 3-O-sulfation of the GlcA of CS in nonmammalian species (37
, 38)
, more work is required to determine if this modification exists in vivo and has functional relevance.
HA consists of repeating disaccharide units of GlcNAc and GlcA, but is unique in that it contains no chain modifications and is directly synthesized into the extracellular space, not assembled via a covalent linkage to a PG. In contrast to the three other GAG families, the HA polymerase enzymes, HA synthases (HAS) are 7-transmembrane proteins inserted directly in the plasma membrane (39
40
41)
. Three mammalian hyaluronan synthases have been discovered (HAS1, 2, 3) that are conserved evolutionally in mammalian species but are present on different chromosomes (39
, 40)
. In addition, bacteria and viruses contain HAS genes and are capable of synthesizing HA (42)
. Hyaluronan synthases are unique in that they possess two enzymatic components, one to add on the GlcA and another to add on the GlcNAc. Unlike the other GAGs, addition of the monosaccharide occurs at the reducing end of the growing HA chain on the inside of the membrane, and the growing polysaccharide is extruded out through the HAS complex into the extracellular space. This unique synthesis mechanism allows for unrestrained polysaccharide length, accounting for how HA exists in vivo in excess of 106 Da (40)
. Although HA is not covalently attached to a PG intracellularly, HA possesses a variety of protein binding partners extracellularly, such as SHAP and aggrecan, and is most likely not found in vivo as a free GAG (43
, 44)
.
Similar to HS and CS/DS, HA does exist in biologically inactive and active forms and this appears to be dependent on HA size (45
46
47
48)
. Contributing to differences in size, HAS’s have been found to produce variations in HA chain length; HAS1 and HAS3 produce 2 x 105 to 2 x 106 Da whereas HAS2 synthesizes larger HA in excess of 2 x 106 Da (39)
. In addition to their different chromosome locations, the HAS genes differ in their regulatory mechanisms and function independently of each other (39)
. Knockout studies have shown the importance of HA in development, as the HAS2 knockout is embryonic lethal due to cardiac and vascular abnormalities (49)
. However, HAS1 and HAS3 knockouts are viable and fertile (39)
. Ongoing investigations with these animals will provide more details to the functional roles of hyaluronan synthases and HA production and activity in vivo.
|
HEPARAN SULFATE/HEPARIN AND INFLAMMATION |
|---|
|
TOPABSTRACTINTRODUCTIONGLYCOSAMINOGLYCAN STRUCTUREHEPARAN SULFATE/HEPARIN AND...CHONDROITIN/DERMATAN SULFATE AND...HYALURONAN AND INFLAMMATIONCONCLUSIONSREFERENCES |
|---|
, 4)
. Glypican, another cell surface HSPG, consists of six family members (glypican-1 to -6), and are tethered by a GPI anchor (50
, 51)
. Other HSPGs include perlecan, agrin, and betaglycan (50
, 51)
. In addition, there are proteins that can be modified by HS chain addition under certain conditions, such as CD44 and collagen XVIII, thus delineating them as HSPG (50
, 51)
. Several PG deficient mice have been studied, including three of the syndecans, glypican-3 and perlecan (Table 1)
(52
53
54
55
56
57
58
59)
.
After injury, GAGs are released from their PG backbone or from the cell membrane, becoming soluble. These soluble GAGs can then be further modified to alter chain length or reveal specific domains to convey a signal that was previously masked. For example, soluble syndecan ectodomains are found after injury, but these did not promote FGF-2 activity without involvement of a HS-degrading enzyme, heparanase (60)
. Heparanase is an endoglycosidase that degrades HS to lower molecular weight fragments. It is released by several cells involved in inflammation, such as platelets, neutrophils, monocytes, and T lymphocytes (61
, 62)
, and is proposed to play a role in cancer metastasis (61
, 63)
. Increased levels of heparanase correlate to increases incidence of metastasis and blocking of heparanase has resulted in a decrease in tumor growth (63)
. Heparanase works on the soluble ectodomains of syndecan-1, releasing oligosaccharides that are "heparin-like" in their more heavily sulfated and epimerized residues (60)
. These HS oligosaccharides contain the correct sequence motifs to promote FGF-2 activity (a minimal pentasaccharide containing N-sulfated glucosamines and 2-sulfated IdoA residues (64
, 65)
), describing a role for the HSPG syndecan-1 in the regulation of growth factor activity that is not apparent prior to inflammation (60)
. Several other growth factors have similar dependence on the presence of heparin-like domains for their activity (Table 2
). Thus, HS modifications as a consequence of processing events that occur during inflammation have been hypothesized to be an important step in the tissue repair process.
|
Other HSPGs have been associated with inflammation and later events in wound healing. CD44 is a multi-functional cell surface receptor that is well known for its interactions with HA. However, certain CD44 isoforms can be modified by HS chain addition delineating CD44 as a HSPG. CD44/HSPG is expressed on monocytes, but not lymphocytes, after the monocytic differentiation to macrophages or after stimulation by IL-1ß or LPS (66)
. CD44/HSPG is capable of binding FGF-2, VEGF and heparin binding-EGF (HB-EGF) but not chemokines, such as MCP-1 or IL-8 (66)
. Staining of inflamed synovial membrane macrophages revealed that they express CD44/HSPG and that staining was greatest where high levels of FGF-2 were also present. This led to the proposal that CD44/HSPG is involved in the regulation of growth factors during the inflammation stage of wound healing. Due to the importance of macrophages in the regulation of tissue repair and remodeling, the expression of CD44/HSPG, which is induced in the first 24 h after stimulation with LPS, may facilitate and regulate the activity of growth factors to adjacent cells (66)
. The authors suggest a disease correlate with their observations in regard to inflammatory arthritis. High levels of proinflammatory cytokines, such as IL-1 and TNF, are continuously elevated leading to the production of growth factors such as FGF, VEGF, and PDGF by monocytes and macrophages. This in turn promotes fibroblast proliferation and angiogenesis, which is regulated by the presence of CD44/HSPG and syndecan-2 on the activated macrophages (66)
. In addition to CD44/HSPG, syndecan-2 has also been described to be up-regulated by activated macrophages and is capable of binding to macrophage-derived growth factors such as FGF-2, VEGF, and HB-EGF (67)
. These activated macrophages expressing syndecan-2 were capable of presenting FGF-2, which is also produced and released by macrophages, to their appropriate target cells for activation (67)
.
In addition to growth factors, HS influences the behavior of a variety of enzymes and enzyme inhibitors. The HS structure required for anticoagulation activity is defined based on a minimum pentasaccharide of 2-sulfated IdoA, and 6- and N-sulfated glucosamines containing a rare 3-sulfate modification (68)
. Such modifications of HS are tightly controlled and dependent on cell type. Endothelial cells are able to produce highly sulfated HS that is capable of binding antithrombin III on the vascular wall (68)
. HS and heparin facilitate a conformational change in both the inhibitor (i.e., ATIII) and the target-inhibited protein (i.e., thrombin). This results in the formation of an irreversible complex that is subsequently cleared from circulation. Another serpin, heparin cofactor II (HCII), is activated by HS/heparin (68)
. HCII is an inhibitor of thrombin only and cannot act as a substitute for ATIII deficiency (68)
. However, clot-bound thrombin is resistant to inactivation by HCII. Liaw et al. have shown that HS/heparin is capable of binding to both thrombin and fibrin and therefore, prevents HCII from inhibiting its target, thrombin (69)
. Additionally, ectodomains of syndecan-1 and syndecan-4 found in wound fluid are capable of binding to neutrophil elastase and cathepsin G, two proteases released by infiltrating leukocytes into injured tissue (70)
. These HSPG protect neutrophil elastase and cathepsin G from inactivation by serpins, suggesting that HSPG are involved in maintaining the balance of protease and anti-protease activity in wound fluid (70)
.
HS influences the recruitment of leukocytes to sites of inflammation and injury. This occurs by HSPG regulation of gradients of chemokines and cytokines produced by EC that have been stimulated by proinflammatory factors, such as IL-1ß and TNF. Numerous cytokines and chemokines have been implicated in these processes including several members of the interleukin family (IL-2, -3, -4, -5, -7, -8, -10, and -12), GM-CSF, RANTES, IP-10, MCP-1, and MCP-4 (2)
. One of the most well studied interactions is that of IL-8 and HSPG. The CXC-chemokine IL-8 is generated by endothelial cells (EC) but held at the cell surface by their interaction with HSPG. Studies differ as to which HSPGs may be involved, syndecan-1 (71)
and syndecan-2 (72)
have both been implicated as the HSPG that retains IL-8 on the EC surface and creates a gradient for inflammatory cell recruitment. Marshall et al. found IL-8 complexed with HS and syndecan-1 in the supernatant of EC and levels of this complex increased when a neutralizing antibody to plasminogen activator inhibitor-1 (PAI-1) was added to the cell culture (71)
. They have hypothesized that plasmin, which is activated by plasminogen activator, induces shedding of IL-8. This subsequently results in soluble IL-8 that binds neutrophils and prevents them from binding EC and subsequent transendothelial migration (71)
. However, PAI-1, which is produced by platelets, is capable of stabilizing the chemoattractant form of IL-8 at the cell surface of EC (71)
.
In addition to their involvement in cytokine and chemokine regulation, HSPG have been implicated in selectin binding. HSPG are capable of interacting with P-selectin and L-selectin (2)
. Koenig et al. have proposed that the minimum size requirement of HS is a tetradecasaccharide; L-selectin prefers more sulfated and epimerized HS oligosaccharides whereas P-selectin bound HS that was less modified (73)
. Studies in the rat kidney determined that L-selectin was important in the infiltration of leukocytes into the kidney (74)
. Subsequent studies determined that the HSPG involved in leukocyte recruitment is collagen XVIII, a basement membrane HSPG (75)
. While verifying that collagen XVIII is the HSPG responsible for L-selectin-mediated cell adhesion, Kawashima et al. also determined that collagen XVIII is capable of binding to monocyte chemoattractant protein-1 (MCP-1) and presenting it to THP-1 monocytic cells (75)
. This resulted in the enhanced 4ß1 integrin binding to VCAM-1, the next step in the more firm adhesion of infiltrating inflammatory cells to the endothelium (2
, 75)
. This suggests that the HSPG collagen XVIII may provide a link between initial rolling of inflammatory cells via L-selectin followed by the induction of chemokine-induced, integrin-dependent adhesion (75)
.
A final interesting example of HS interactions in immune defense has come in the study of innate immune defense by antimicrobial peptides. These highly cationic peptides have been shown to bind and be inactivated by HSPG, a mechanism exploited by pseudomonas aeroginosa that releases a factor capable of inducing the release of HSPG from pulmonary epithelium. Upon release, inactivation of antimicrobial peptides by HSPG provides a survival advantage for the bacteria and leads to increased infection (76)
.
|
CHONDROITIN/DERMATAN SULFATE AND INFLAMMATION |
|---|
|
TOPABSTRACTINTRODUCTIONGLYCOSAMINOGLYCAN STRUCTUREHEPARAN SULFATE/HEPARIN AND...CHONDROITIN/DERMATAN SULFATE AND...HYALURONAN AND INFLAMMATIONCONCLUSIONSREFERENCES |
|---|
78
79
80)
. In addition to CS chains, a smaller list of PG that contains DS chains include versican, decorin, biglycan, and endocan (5)
. The two most well studied DSPG are decorin and biglycan; they are small, leucine-rich PG that contain 1 DS chain (decorin) or 1-2 DS chains (biglycan), and secreted DSPGs prominent in connective tissue cells (5
, 81
82
83
84
85)
. Table 1
lists PG knockout mice, as well as GAG synthesizing enzyme deficiencies, and possible human disease correlations that have been described.
As discussed for HS, CS function is dependent on their structure, size, and sequence, all of which may be altered in inflammation. After injury, CS/DS and CSPG/DSPGs become soluble and are a major component in wound fluid (86)
. DS comprises over half of the total soluble GAG content of wound fluid. This soluble DS has been shown to be capable of activating growth factors such as FGF-2 and FGF-7, also known as keratinocyte growth factor (KGF) (86
, 87)
. As described for HS, studies have shown that specific size and sequence requirements exist for DS to promote growth factor activity. Thus, like HS, the synthesis and release of DS with injury may modify several cell behaviors including those associated with activation of growth factor (GF) activities (Fig. 2
). Data indicates that the minimum size of DS required to promote FGF-2 proliferation is an octasaccharide. 4-O-sulfation on the GalNAc is sufficient to promote activity, increasing sulfation to a 2/4-O-disulfated disaccharide does not appear to increase activity (88)
. The sequence required for FGF-7-dependent cell proliferation consists of a minimum decasaccharide oligo that possess 4-O-sulfation of the GalNAc (88)
. Another growth factor, hepatocyte growth factor/scatter factor (HGF/SF), requires a minimal size of a DS octasaccharide that consists of unsulfated IdoA residues and 4-O-sulfated GalNAc (89)
. In addition, CS-E has been shown to bind to several heparin binding factors including midkine, pleiotropin, HB-EGF, FGF-16, and FGF-18 (90)
, although studies to determine which sequences, and therefore which enzymes, may partake in their synthesis, still need to be determined.
|
DS chains can also regulate coagulation events early in inflammation. Similar to HS, DS binds to HCII, a serine protease inhibitor that inhibits the procoagulation effects of thrombin (91)
. However, unlike HS and heparin, DS has no activity in promoting ATIII activity, another serine protease inhibitor that has more widespread effects on several proteases (91)
. Despite similarities between HCII and ATIII, HCII cannot compensate for total loss of ATIII (68)
. A minimum sequence of a hexasaccharide containing 2-O-sulfated IdoA and 4-O-sulfated GalNAc residues are required for HCII binding, which accounts for only 
5% of the total disaccharide content of DS (92)
. Structural studies have revealed that DS is only capable of binding to thrombin and not fibrin, which allows thrombin susceptibility to inactivation by HCII (69)
. This is in contrast to heparin, which binds both thrombin and fibrin, and protects thrombin from inactivation by inhibitors (69)
. This suggests that DS has possible therapeutic applications, as it differs from heparin in target molecules and also it’s activity in clot-bound fibrin. Another molecule associated with DS and involved early in clot formation is platelet factor-4 (PF-4). DS binds PF-4 (93)
, and this interaction is dependent on the presence of 4/6-O-disulfated disaccharides (94)
. PF-4 is released at high concentrations from platelets participating in the early events of clot formation.
An area of research that has been minimally explored is what factors influence changes in GAG structures and sequences and what this may mean for later stages in inflammation. For example, studies investigating the effects of pro- and anti-inflammatory cytokines on endothelial cells (ECs) have shown that that IL-6 and IL-10 induce decorin synthesis in ECs that can lead to effects in angiogenesis (95)
. Despite increases in decorin mRNA in ECs after incubation with cytokines, decorin protein only increased in ECs grown on a collagen lattice. This highlights an area of research that is largely undeveloped: how inflammatory factors in turn regulate the activity of GAG synthesizing and modifying enzymes.
DS, like HS, also binds several molecules involved in the recruitment, rolling, and subsequent extravasation of leukocytes. DS signals human microvascular dermal ECs to up-regulate ICAM-1 expression on their cell surface (96)
. Mice injected with DS increased their soluble levels of circulating ICAM-1, which was not observed with injection of other GAGs such as HS (96)
. The DS chains of PGs bind several factors involved in leukocyte adhesion, specifically L-selectin, P-selectin, and CD44. Versican is a CSPG capable of binding these factors; the CS chain binds to the same site of L-selectin and P-selectin as HS and other GAGs, but to a different site than HS on CD44 (97)
. The specific modifications required to bind were CS-B (or DS) and CS-E disaccharides for L-selectin and P-selectin. For CD44, HA along with all variants of CS chains, and also unsulfated chondroitin were capable of binding (97)
. Binding of CS/DS to L-selectin and P-selectin is sulfation dependent while the binding to CD44 is sulfation independent (98)
. These more highly sulfated CS/DS chains that are capable of binding L-selectin and P-selectin also bind chemokines SLC, IP-10 and SDF-1ß (98)
. These oversulfated CS/DS chains inhibit chemokine activity due to their high affinity and possibly may provide a reservoir of these cytokines (98)
.
|
HYALURONAN AND INFLAMMATION |
|---|
|
TOPABSTRACTINTRODUCTIONGLYCOSAMINOGLYCAN STRUCTUREHEPARAN SULFATE/HEPARIN AND...CHONDROITIN/DERMATAN SULFATE AND...HYALURONAN AND INFLAMMATIONCONCLUSIONSREFERENCES |
|---|
, 45)
. Several cell types including fibroblasts, endothelial cells, and keratinocytes synthesize HA into the extracellular space (99
100
101)
. A large proportion of HA is found in the skin, 56% of total HA in this tissue (6)
.
Turnover of HA occurs daily with the half-life of HA ranging from a couple minutes in the blood to 12 h in the skin to several weeks in the vitreous body of the eye (99)
. This observation raises important questions regarding HA function, considering the great energy investment required to make an HA chain. After injury or during inflammation, HA is degraded or broken down into smaller molecular weight fractions (45
, 46
, 102)
. Degradation can occur enzymatically by HA-degrading enzymes, or nonenzymatically, such as mechanical injury or by free radicals (45)
, and has direct implications in several inflammatory and tissue-repair events (Fig. 3
). HA-degrading enzymes include both mammalian hyaluronidases (103)
, and bacterial hyaluronidases (104)
. These bacterial enzymes appear to assist bacterial spreading through the extracellular matrix (104)
. However, whether these enzymes produce HA products that are capable of eliciting an immune response has not been well studied. Low molecular weight HA (LMW HA) participates in a variety of signaling events including cellular proliferation, migration, differentiation, angiogenesis, and induction of proinflammatory cytokines and chemokines. It has been suggested that this process of HA breakdown is a highly regulated event, and that the inability to remove or properly degrade HA leads to improper repair (45)
. Due to the absence of any GAG chain modifications, HA function appears to be dependent on its size.
|
As described for HS and DS, although via a different mechanism, HA is involved in the recruitment of leukocytes to sites of inflammation by its interaction with CD44. HA production by EC is induced by proinflammatory cytokines such as TNF, IL-1ß (105)
. HA expressed by the ECs and CD44 on leukocytes initiates the interaction between leukocytes and the endothelium. This interaction is dependent on the activation of ECs to produce HA and on the expression of a specific HA binding CD44 isoform by leukocytes (105)
. However, CD44 on T cells and B cells is not always capable of binding HA, suggesting that this interaction is tightly controlled (106)
.
In addition to binding leukocytes, HA has also been implicated in regulating the activation of leukocytes. Mummert et al. have described a role for HA in the activation of CD4+ T cells by dendritic cells (DC) (107)
. DC express both HA synthases and hyaluronidases, produce HA and display HA on their cell surface (107)
. Their work showed that HA binding proteins interfere with DC cluster formation, IL-2 and IFN
production, and cell proliferation. This led them to conclude that HA on DC facilitates interactions and activation between DC and T cells. CD44 on T cells interacts with HA on endothelium to mediate adhesion and this appears to be dependent entirely on CD44 (107)
. However, CD44 could not account fully for HA-dependent T cell activation, suggesting the involvement of other receptors in cellular activation.
Animal models of pulmonary injury have supported the role of HA in inflammation by evaluation of the consequences of CD44 expression after bleomycin injury. Wild-type mice develop lung fibrosis but survive, whereas 75% of the CD44-deficient mice die by day 14 after injury (108)
. Unlike their CD44-deficient counterparts, CD44 wild-type mice were able to clear infiltrating cells between days 5 and 10. However, in CD44-deficient mice, infiltrating cells continually increased until the mouse died. Similarly, cytokine levels in the wild-type mice decreased by day 7 after injury but persisted in the CD44-deficient mice. Wild-type mice possessed LMW HA after injury that eventually returned to HMW HA, whereas CD44-deficient mice had a variety of HA oligosaccharide sizes that did not exist in the wild-type mice (108)
. These experiments support a necessary role for CD44 and proper HA clearance after injury. The absence of CD44 leads to uncontrolled HA accumulation and inflammatory response and results in the death of CD44-deficient mice.
Another signal transducing receptor for HA is RHAMM (Receptor for HA Mediated Motility). Similar to CD44, RHAMM can be alternatively spliced, and has also been found both extracellularly and intracellularly (109)
. In addition to signaling properties, RHAMM appears to be involved in cellular motility and migration due to its interactions with the cytoskeleton. These observations have implicated RHAMM in aspects of HA and tumor progression (110)
. Indeed, RHAMM knockout mouse studies have shown that RHAMM is involved in cell proliferation and cell migration, important events in wound healing and tumor pathology (111)
. RHAMM knockout mice have a decrease in size and number of fibromatosis as compared with wild-type controls (111)
. In addition, Nedvetzki et al. have described a role for RHAMM in compensating for CD44 in CD44-deficient mice (112)
. The authors have proposed that when CD44 is absent during development, RHAMM is capable of compensating and performing overlapping functions (112)
. RHAMM is not up-regulated in the absence of CD44; instead, it appears that RHAMM is used when HA is not internalized by CD44.
HA has also been implicated in interacting with molecules associated with inflammation. Toll-like receptor 4 (TLR4), best know for its ability to recognize lipopolysaccharide (LPS), is required for the ability of LMW HA fragments to induce the maturation of DC and stimulate release of proinflammatory cytokines such as IL-1ß, TNF, and IL-12 (113
, 114)
. The induction of inflammatory cytokines through TLR4 occurred in the presence of HA fragments, but not high molecular weight or intermediate molecular weight HA. Microvascular dermal ECs also respond to the presence of LMW HA fragments and release IL-8 in a TLR4-dependent manner (115)
. These data support other recent evidence suggesting that tissue components can serve to activate innate pattern recognition receptors such as TLR4. Endogenous factors such as intracellular components of necrotic cells, extracellular matrix components, and heat shock proteins, normally not seen in healthy tissue, are able to induce an immune response through TLRs (116
117
118)
.
Other molecules also influence the effects of HA and inflammation. TSG-6 is a binding partner of HA, and its expression increases after injury or during inflammatory conditions, similar to HA. TSG-6 is not found in normal, healthy adult tissue but is synthesized by several cell types after stimulation with proinflammatory mediators and growth factors. Therefore, TSG-6 has been implicated as an important protein during resolution of inflammation (119)
. TSG-6 is found in the sera of patients with bacterial sepsis, systemic lupus erythematosus, inflammatory bowel disease, and multiple forms of arthritis (119)
. TSG-6 stabilizes and increases the interaction of CD44 and HA and may play a role in leukocyte rolling and attachment (120)
. TSG-6 has also been found complexed to various forms of Inter--Inhibitor (II) (119)
. II is comprised of 2 heavy chains (HC1 and HC2) complexed to one light chain (bikunin) via a chondroitin sulfate chain (121)
. It has been suggested that II is an anti-inflammatory agent activated by TSG-6 (121)
. Roles for TSG-6 and II include the transfer of HC1 and HC2 of II by TSG-6 to HA, protecting HA from degradation (121)
. Heavy chains bound to HA have been found in patients with arthritis. In addition, bikunin is responsible for the protease inhibitory activity of II, and can be found free in plasma (121)
. TSG-6 appears to potentiate the anti-plasmin activity of II, an important step in the resolution of inflammation (119)
.
Further support of the importance of HA in inflammation comes from several studies describing associations between HA and human inflammatory diseases. HA is elevated in disease states including osteoarthritis, rheumatoid arthritis, psoriasis, and scleroderma (99)
. Detailed studies have shown a direct link between HA and colon inflammation (122)
. Viral infection or exposure to a viral mimic induced HA cell surface expression in human colonic SMCs. Staining of involved tissue from patients with inflammatory bowel disease showed enhanced HA staining, an observation suggested by the authors to contribute to chronic inflammatory conditions (123)
.
The biologic effects of HMW HA have now expanded to include recognition of activity as an anti-angiogenic, anti-proliferative and anti-inflammatory molecule. HMW HA is commonly used therapeutically in the treatment of osteoarthritis by intra-articular injection. Pharmaceutical-grade high molecular weight HA may in part aide knee pain by restoring viscoelasticity of synovial fluid and perhaps increasing endogenous HA production (124)
. However, while some studies indicate that pain in patients injected with HMW HA does decrease, the mechanism responsible for how HA functions in vivo after injection is unclear.
HMW HA has been proposed to be one factor involved in scarless fetal dermal wound healing. In contrast to adult wounds or even the later gestational age fetus, early gestational age fetuses heal without scar formation (125)
. One factor suggested for this difference is the interaction of HA and platelets. Platelets bind to HA via CD44 and this appears to regulate their binding after injury (126)
. However, in the fetal wound environment, HA inhibits platelet aggregation and subsequent proinflammatory cytokine release (127)
. In addition, fetal fibroblasts make HA and express 2- to 4-fold more HA receptors than adult fibroblasts (127)
and HA persists in the fetal wound environment considerably longer than in the adult, 3 wk compared with 3 to 5 days (125)
. The observation that fetal wounds have considerably less leukocyte infiltration than adult wounds can be accounted for by a variety of factors including a decrease in proinflammatory cytokines, increased fibroblast migration, an increase in matrix metalloproteinases (MMPs) and a decrease in tissue inhibitor of metalloproteinases (TIMPs) (125)
. While HA has been proposed as a factor in fetal wound healing and data has suggested that it is a participant in both fetal and adult wound healing, how it contributes to both scarless and scar-forming wound healing has yet to be determined.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGLYCOSAMINOGLYCAN STRUCTUREHEPARAN SULFATE/HEPARIN AND...CHONDROITIN/DERMATAN SULFATE AND...HYALURONAN AND INFLAMMATIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGLYCOSAMINOGLYCAN STRUCTUREHEPARAN SULFATE/HEPARIN AND...CHONDROITIN/DERMATAN SULFATE AND...HYALURONAN AND INFLAMMATIONCONCLUSIONSREFERENCES |
|---|
