Hyaluronan fragments synergize with interferon-gamma to induce the C-X-C chemokines mig and interferon-inducible protein-10 in mouse macrophages.

Hallmarks of chronic inflammation and tissue fibrosis are increased influx of activated inflammatory cells, mediator release, and increased turnover and production of the extracellular matrix (ECM). Recent evidence has suggested that fragments of the ECM component hyaluronan play a role in chronic inflammation by inducing macrophage expression of chemokines. Interferon-gamma (IFN-gamma), an important regulator of macrophage functions, has been shown to induce the C-X-C chemokines Mig and IP-10. These chemokines affect T-cell recruitment and inhibit angiogenesis. The purpose of this investigation was to determine the effect of hyaluronan (HA) on IFN-gamma-induced Mig and IP-10 expression in mouse macrophages. We found a marked synergy between HA and IFN-gamma on Mig and IP-10 mRNA and protein expression in mouse macrophages. This was most significant with Mig, which was not induced by HA alone. The synergy was specific for HA, was not dependent on new protein synthesis, was not mediated by tumor necrosis factor-alpha, was selective for Mig and IP-10, and occurred at the level of gene transcription. These data suggest that the ECM component HA may influence chronic inflammatory states by working in concert with IFN-gamma to alter macrophage chemokine expression.

Activated macrophages play an essential role in inflammation through the release of a variety of mediators, including reactive oxygen species, reactive nitrogen species, proteases, chemokines, cytokines, and growth factors (1)(2)(3)(4)(5). Although the mechanisms controlling macrophage activation in inflammatory states are incompletely understood, recent studies suggest a role for extracellular matrix (ECM) 1 components in the acti-vation of inflammatory macrophages (6). In normal, healthy tissues, the ECM is composed of a complex array of proteins, proteoglycans, and glycosaminoglycans that play important roles in homeostasis (7) and maintenance of matrix structure. During an inflammatory response, there is increased production and degradation of the ECM, resulting in the accumulation of breakdown products of ECM components such as fibronectin, collagen, and glycosaminoglycans (8). Interestingly, these lower molecular mass ECM components have been shown to have different biological activities compared with their larger, native precursors (9 -11). In fact, fragments of the ECM components collagen and fibronectin have been recently shown to have proinflammatory properties (12). Likewise, recent work from our laboratory and others has shown that low molecular weight fragments of the ECM glycosaminoglycan hyaluronan (HA) may play a role in macrophage activation (13)(14)(15)(16) HA is a ubiquitously distributed component of the ECM. In its native form, it exists as a high molecular weight nonsulfated glycosaminoglycan polymer (17,18) made up of repeating disaccharide units of (␤,1-4)-D-glucuronic acid-(␤,1-3)-N-acetyl-D-glucosamine. High molecular mass HA is believed to have many functions in healthy tissue, such as water homeostasis, plasma protein distribution, and matrix structuring (17). However, at sites of inflammation and tissue injury, there is an accumulation of lower molecular mass HA species (19 -21), which have different biological functions than their high molecular mass precursors (13)(14)(15)(16). Recent studies have suggested that these lower molecular weight forms of HA may stimulate macrophages recruited to sites of inflammation to produce important mediators of tissue injury and repair (13)(14)(15)(16). This effect is mediated, in part, by interaction with the HA receptor CD44 (13)(14)(15)(16).
In the inflammatory milieu, there are numerous cytokines and chemokines that also influence macrophage expression of inflammatory genes. The cytokine IFN-␥ is an important modulator of macrophage effector functions and regulator of the inflammatory response (22). In addition to its antiviral activity, IFN-␥ also enhances certain macrophage functions, such as microbicidal and tumoricidal activity, through the production of reactive oxygen intermediates and reactive nitrogen intermediates (23,24). Yet despite its many proinflammatory roles, IFN-␥ also inhibits macrophage expression of certain LPSinduced chemokines, such as monokine chemoattractive protein-1 and KC (24).
IFN-␥ has recently been shown to induce macrophage expression of the novel chemokines monokine induced by interferon-␥ (Mig) and interferon-inducible protein-10 (IP-10) (25, 26). Both Mig and IP-10 are members of the C-X-C chemokine family and may play roles in chronic inflammation (27,28), as well as in viral and protozoan infections (25,29,30). In chronic infectious states, expression of Mig and IP-10 has been shown to correlate with IFN-␥ expression (29). Mig and IP-10 have each been implicated in T-cell trafficking, chemotaxis, and activation (31)(32)(33). Mig and IP-10 have also been shown to have angiostatic properties (34 -36) and thus may play an important role in regulating tissue granulation and remodeling by inhibiting angiogenesis. IP-10 is induced in macrophages by ␣, ␤, and IFN-␥, as well as by LPS (25,37,38). To date, Mig has only been shown to be induced in macrophages by IFN-␥ (26,39). Recent investigations have described synergistic enhancement of IFN-␥-induced Mig and IP-10 expression by TNF-␣ in fibroblasts (40,41).
In this report, we examined the effect of low molecular mass HA fragments on IFN-␥-induced expression of Mig and IP-10. We observed a striking synergy between HA and IFN-␥ on the expression of Mig and IP-10 mRNA and protein levels in murine macrophages that occurs at the level of gene transcription. These results identify a previously uncharacterized mechanism by which the ECM, acting in conjunction with IFN-␥, may regulate the immune response by influencing the expression of specific chemokines.
Chemicals and Reagents-Purified HA fragments from human umbilical cords were purchased from ICN Biomedicals, Inc. (Costa Mesa, CA). The HA-ICN preparation was free of protein (Ͻ2%) and other glycosaminoglycans with a peak molecular mass of 200,000 Da (44). Recombinant mouse IFN-␥ (specific activity, 3.0 ϫ 10 5 units/ml; endotoxin level less than 0.2 ng/mg) was from Genzyme Corporation (Cambridge, MA), the recombinant mouse TNF-␣ was purchased from R & D Systems, the anti-TNF-␣ was from Cappel (Aurora, OH), and the cycloheximide (10 g/ml) was from Sigma. Polymixin B was purchased from Calbiochem. Stock solutions of reagents were tested for LPS contamination using the Limulus amebocyte assay (Sigma).
Northern Analysis of mRNA Production-RNA was extracted from confluent cell monolayers using 4 M guanidine isothiocyanate and purified by centrifugation through 5.7 M cesium chloride for 12-18 h at 35,000 rpm as described (16). Ten g of total RNA was electrophoresed under denaturing conditions through a 1% formaldehyde-containing agarose gel and RNA was transferred to Nytran (Schleicher and Schuell) hybridization filters. Blots were briefly rinsed in 5ϫ SSC, RNA was cross-linked to the filter by UV cross-linking (Stratagene, La Jolla, CA), and blots were hybridized overnight with 10 6 cpm/ml of 32 P-labeled DNA labeled by the random prime method (Amersham Pharmacia Biotech). Following hybridization, blots were washed once in 2ϫ SSC/ 0.1% SDS at room temperature for 30 min with shaking, and then washed twice in 0.1ϫ SSC/0.1% SDS at 50°C with shaking for 20 min each wash. Blots were exposed at Ϫ70°C against Kodak XAR diagnostic film. Differences in RNA loading were documented by hybridizing selected blots with 32 P-labeled cDNA for aldolase (45). Densitometric scanning was performed using a Molecular Dynamics Personal Densitometer SI (Sunnyvale, CA).
Western Analysis of Protein Secretion-Western blot analysis was performed as described (46). Briefly, 200 g of macrophage-conditioned media was fractionated by SDS-polyacrylamide gel electrophoresis (10%), transferred to a nylon membrane, blocked and washed, incubated with the polyclonal anti-Mig antibody at a dilution of 1:2500 or polyclonal anti-crg-2 (murine IP-10 antibody from R & D Systems) at a dilution of 1:3000, and developed with a chemiluminescent system according to the manufacturer's instructions (Amersham Pharmacia Biotech). The recombinant Mig protein was prepared as described previously (26), and the recombinant crg-2 (murine IP-10) was purchased from R & D Systems.
Nuclear Run-on-Nuclei from confluent monolayers of MH-S cells were harvested by scraping in ice-cold phosphate-buffered saline and subsequently isolated by centrifugation through a sucrose cushion (47). Nuclei were then incubated for 30 min with 1 M dithiothreitol, 20 mM NTPs, and 100 Ci of [ 32 P]UTP in transcription buffer. The nuclei received a cold UTP chase for 10 min before the reaction was stopped by the addition of termination buffer, DNase (Promega, Madison, WI) and RNase-Inhibitor (Boehringer Mannheim). The nuclei were then incubated with tRNA (Sigma) for 15 min before the addition of 10% SDS, 0.2 M EDTA and proteinase K (Sigma). After 15 min of incubation, the RNA was extracted with phenol:chloroform:isoamyl alcohol, precipitated with 20% trichloroacetic acid, washed with 5% trichloroacetic acid/5% PPi, dissolved in 0.1% SET, and precipitated for a second time with 4 M NaAc and 100% ETOH at Ϫ80°C for 30 min. Purified radiolabeled RNA was washed once in 70% ETOH, dried with a speed vacuum concentrator (Savant), and resuspended in 100 l of water. 5 l of the radioactive RNA was counted, and all samples were normalized for counts using hybridization fluid. Normalized samples were hybridized with prehybridized Optitran-S membranes (Schleicher and Schuell) containing the cDNAs of interest. Blots were hybridized for 3-4 days and then washed once in 2ϫ SSC/0.1% SDS at room temperature for 5 min with shaking, and then washed twice in 0.1ϫ SSC/0.1% SDS at 50°C with shaking (20 min each wash). The blots were then exposed and quantitated with a PhosphorImager (Molecular Dynamics).
Transient Transfections-Transient transfections of the MH-S cells were performed using Lipofectin as described elsewhere (48). Briefly, 2.5 ϫ 10 6 cells were plated in 60-mm tissue culture dishes 1 day prior to transfection. The cells were washed twice with Opti-MEM (Life Technologies, Inc.) and incubated with the DNA/Lipofectin solution (5 g of DNA and 15 g of Lipofectin) for 7 h. The DNA/Lipofectin solution was then aspirated, 3 ml of Opti-MEM was added, and the cells were incubated overnight at 37°C. The contents of the dishes were scraped in 8 ml of fresh medium and divided between four 60-mm culture dishes. After incubation for 4 h to allow the cells to adhere, one dish from each transfection group received medium alone, one received HA (100 g/ ml), the third received IFN-␥ alone (300 units/ml), and the last received HA ϩ IFN-␥. The cells were harvested for analysis of chloramphenicol acetyltransferase (CAT) expression after 18 h of stimulation. The Ϫ1117/ϩ43 Mig/firefly luciferase, ␥RE-1 ϫ 4 Mig/firefly luciferase, and Ϫ299/ϩ7 IP-10/CAT construct were prepared as described previously (49,50).  (13)(14)(15)(16). Furthermore, recent work from our laboratory has shown that IFN-␥ selectively inhibits HA-induced MIP-1␣ and MIP-1␤ expression in primary mouse macrophages (51). We were therefore interested in the possible role for low molecular mass HA in regulating Mig and IP-10 expression in the presence of IFN-␥. In order to assess the combined effect of HA and IFN-␥ on Mig and IP-10 gene expression, macrophages were simultaneously stimulated with HA and IFN-␥ for 6 h, mRNA was isolated, and Northern analysis was performed. As shown in Fig. 1A, HA alone had virtually no effect on Mig or IP-10 mRNA expression in inflammatory thioglycollate-elicited peritoneal macrophages from C 3 H/HeJ-LPS hyporesponsive mice, whereas IFN-␥ induced moderate expression of Mig and IP-10 mRNA. However, we unexpectedly found that HA dramatically influences the effect of IFN-␥ on Mig and IP-10 mRNA expression (Fig. 1A). Similar results were observed with MH-S cells, a murine alveolar macrophage cell line (Fig. 1B). In these alveolar macrophages, the synergy is most apparent for Mig where, in addition to no effect on Mig gene expression by HA alone, IFN-␥ alone induced only a faint signal (Fig. 1B). However, the combination of HA and IFN-␥ markedly induced Mig mRNA expression (Fig. 1B). In this alveolar macrophage cell line, HA minimally induced IP-10 mRNA, and IFN-␥ had a moderate effect on IP-10 mRNA, but there was still marked enhancement of IP-10 mRNA when cells were stimulated both with HA and IFN-␥ (Fig. 1B).

HA Fragments Synergize with IFN␥ to Induce
HA and IFN␥ Exhibit a Time-and Dose-dependent Synergistic Induction of Mig and IP-10 mRNA Levels-To further delineate the effects of HA and IFN-␥ on Mig and IP-10 mRNA expression, we stimulated MH-S cells simultaneously with HA and IFN-␥ for varying time intervals and found that HA failed to independently induce Mig expression at any time point and only minimally induced IP-10 mRNA (data not shown). How-ever, the synergistic effect of HA and IFN-␥ on Mig and IP-10 mRNA expression was seen as early as 3 h after simultaneous stimulation of MH-S cells, peaked after 6 -9 h, and decreased toward baseline after 24 h of stimulation (data not shown).
We then determined the dose-response relationships for HA and IFN-␥ to induce Mig and IP-10 gene expression in MH-S cells. The effect of varying concentrations of HA with a constant concentration of IFN-␥ on Mig and IP-10 mRNA expression by the alveolar macrophage cell line MH-S is shown in Fig. 2A. The synergy between HA and IFN-␥ on Mig and IP-10 gene expression was observed with as little as 1 g/ml HA and was maximal at 10 g/ml HA in the presence of 300 units/ml IFN-␥. The converse is shown in Fig. 2B using a concentration of 100 g/ml HA. The synergy between HA and IFN-␥ occurs with as little as 1 unit/ml IFN-␥ and is maximal at 10 units/ml IFN-␥.
IFN␥ Synergizes Specifically with HA Fragments to Induce Mig and IP-10 mRNA Expression in Mouse Macrophages-In order to determine whether the synergy between HA and IFN-␥ was specific to low molecular mass HA fragments, we stimulated MH-S cells in the presence of IFN-␥ and numerous other glycosaminoglycans. Fig. 3 shows that IFN-␥ only synergizes with HA fragments and not high molecular mass HA, chondroitin sulfate A or B, or HA disaccharides to induce Mig or IP-10 mRNA expression. Thus, the synergy between HA and IFN-␥ on Mig and IP-10 gene expression in MH-S cells appears to be specific to the low molecular mass HA fragments.
HA and IFN␥ Synergize to Induce Mig Protein Production in Mouse Macrophages-Having identified the synergistic effect between HA and IFN-␥ on Mig and IP-10 mRNA levels, we investigated Mig and IP-10 production at the protein level. MH-S cells were simultaneously stimulated with HA in the presence of IFN-␥ for 20 h. Mig protein secretion was determined in the supernatants by Western blot analysis. Fig. 4A shows that there was little Mig protein present in the conditioned media from unstimulated cells or cells stimulated with HA or IFN-␥ alone. However, when the macrophages were stimulated with HA and IFN-␥ together, there was marked induction of Mig protein production in the conditioned media. Fig. 4B shows similar synergistic induction of IP-10 protein production by MH-S cells stimulated with both HA and IFN-␥. However, unlike Mig, there was some IP-10 protein induced by HA and IFN-␥ alone. Thus, the synergy between HA and IFN-␥ is also demonstrated at the protein level.

Cycloheximide Has a Minimal Effect on the Synergy between HA and IFN-␥ on Mig or IP-10 Gene Expression in Mouse
Macrophages-In order to further dissect the mechanism for the observed synergy, we examined the role of new protein synthesis on the synergy between HA and IFN-␥ on Mig and IP-10 mRNA expression in MH-S cells. We pretreated MH-S cells with cycloheximide (CHX) for 30 min before the addition of HA and IFN-␥ for 6 h. As shown in Fig. 5, CHX minimally inhibited the synergistic expression of Mig and IP-10 mRNA by HA and IFN-␥. Similarly, CHX had no effect on Mig or IP-10 gene expression from unstimulated cells or cells stimulated with HA alone, and it had little effect on cells stimulated with IFN-␥ alone (Fig. 5). Thus, the synergistic induction of Mig and IP-10 gene expression by HA and IFN-␥ does not appear to require new protein synthesis.

The Synergistic Induction of Mig and IP-10 Gene Expression by HA and IFN␥ Is Not Dependent on TNF-␣-Recently,
Ohmori et al. (40,41) have provided data that TNF-␣ and IFN-␥ synergize to induce Mig and IP-10 mRNA in fibroblasts. We have previously shown that HA induces TNF-␣ expression by mouse macrophages (16). Therefore, we investigated the possible role of TNF-␣ in the synergistic induction of Mig and IP-10 gene expression by HA and IFN-␥. First, using anti- TNF-␣ neutralizing antibodies, we found that TNF-␣ was not necessary for the induction of Mig or IP-10 mRNA by HA and IFN-␥ in MH-S cells (data not shown). In addition, we isolated thioglycollate-elicited peritoneal macrophages from mice that have had the TNF-␣ gene deleted (43). As shown in Fig. 6, HA and IFN-␥ synergize to induce Mig and IP-10 mRNA expression despite the complete absence of TNF-␣. Similar results were also found in thioglycollate-elicited peritoneal macrophages from wild type littermate controls (Fig. 6). Thus, the synergy between HA and IFN-␥ on Mig and IP-10 gene expression occurs independently of TNF-␣ expression. the baseline levels of IP-10 mRNA transcription were higher (Fig. 7). At earlier time points, after 1 h stimulation, there was modest enhancement of Mig and IP-10 mRNA transcription in cells treated with IFN-␥ alone, but the synergistic enhancement with HA was not present (data not shown). Thus, HA and IFN-␥ synergize to induce Mig and IP-10 gene expression at the level of transcription To further elucidate the mechanism responsible for the synergy between HA and IFN-␥ on Mig promoter activity, we transiently transfected MH-S cells with a 5Ј promoter construct spanning base pairs Ϫ1117 to ϩ43 upstream from a firefly luciferase reporter gene. As shown in Fig. 8A, IFN-␥ alone induced firefly luciferase activity by 12-fold, whereas stimulation of transfected cells with HA and IFN-␥ together showed over a 32-fold induction of luciferase activity over unstimulated cells. We also transiently transfected MH-S cells with a promoter construct containing four copies of the interferon-␥-responsive site, Ϫ200 to Ϫ167, of the Mig promoter upstream of a firefly luciferase reporter gene (␥RE-1 ϫ 4). IFN-␥ alone induced firefly luciferase activity by 6.2-fold over unstimulated cells, whereas stimulation with HA ϩ IFN-␥ only minimally induced (1.7-fold) luciferase activity. Thus, the ␥RE is not sufficient to account for the synergistic enhancement of Mig gene expression by HA fragments and IFN-␥.
Similar results were found with transient transfections of MH-S cells with a 5Ј IP-10 promoter construct spanning base pairs Ϫ299 to ϩ7 upstream from a CAT reporter gene. CAT activity was induced 10-fold by HA alone, 8-fold by IFN-␥ alone, and 28-fold after stimulation of transfected cells with HA and IFN-␥ (Fig. 8B). Together, these data suggest that the regulatory elements that convey synergy between HA and IFN-␥ on Mig gene transcription are contained within the approximately 1-kilobase proximal promoter, and for IP-10 gene transcription within the approximately 300-base pair proximal promoter. DISCUSSION The purpose of this study was to examine the effects of the ECM glycosaminoglycan HA on the expression of the inflammatory chemokines Mig and IP-10 in the presence of IFN-␥. Previous work in our laboratory has shown that low molecular mass HA fragments can stimulate mouse macrophages to express numerous chemokines (13)(14)(15). Mig is unique in that no stimulus other than IFN-␥ has been shown to induce its expression in macrophages. IP-10, on the other hand, has been shown to be induced by IFN-␣, -␤, and -␥, as well as LPS. The results presented herein identify a new role for the ECM in enhancing IFN-␥ induced Mig and IP-10 gene expression. Interestingly, the effect of HA on IFN-␥-inducible Mig and IP-10 expression appeared to be unique among the chemokines we examined. We have recently shown that IFN-␥ inhibits HAinduced expression of MIP-1␣, MIP-1b, and KC, while having no significant effect on RANTES and monokine chemoattractive protein-1 (51).
HA alone did not induce Mig and only minimally induced IP-10 mRNA and protein production in a variety of murine macrophages. However, HA enhanced the IFN-␥-induced steady state mRNA levels of Mig and IP-10 in both primary mouse macrophages (elicited peritoneal macrophages) and in MH-S cells, an alveolar macrophage cell line. The increased Mig and IP-10 mRNA levels were found to correlate with Mig and IP-10 protein production in cells stimulated with both HA and IFN-␥.
To further characterize the mechanisms by which HA and IFN-␥ synergize to induce Mig and IP-10 gene expression in macrophages, we performed time course and dose-response experiments. The peak synergy between HA and IFN-␥ occurs after 6 -9 h of stimulation and with as little as 1 g/ml of HA and 1 unit/ml of IFN-␥. Furthermore, the effect of HA on IFN-␥ induced Mig and IP-10 expression is specific to HA and not a general characteristic of glycosaminoglycans. Low molecular mass HA fragments alone, not high molecular mas native HA, chondroitin sulfate A or B, or HA disaccharides, influenced the effect of IFN-␥ on Mig or IP-10 expression. These results suggest that the synergistic effect of HA on IFN-␥-induced Mig and IP-10 mRNA expression is immediate, specific to HA fragments, and quite sensitive as it occurs with low concentrations of these mediators.
In an attempt to delineate whether new protein synthesis was necessary for synergy, we performed a series of experiments in the presence of the protein synthesis inhibitor cycloheximide. Evaluation of these experiments revealed that HAand IFN-␥-induced Mig and IP-10 expression was minimally blocked by CHX, suggesting that the synergy does not require synthesis of a secondary mediator.
Recent investigations have determined that TNF-␣ and IFN-␥ synergize to induce Mig and IP-10 mRNA expression in fibroblasts (40,41); therefore, we examined the role of TNF-␣ in mediating the effect of HA on IFN-␥-inducible Mig and IP-10 expression. Although HA independently induces TNF-␣ expression in macrophages, sufficient concentrations of TNF-␣ blocking antibodies to inhibit all HA-induced TNF-␣ activity failed to inhibit the induction of Mig or IP-10 by HA and IFN-␥ in MH-S cells (data not shown). Furthermore, thioglycollate-elicited peritoneal macrophages from TNF-␣-deficient mice exhibited the same synergy between HA and IFN-␥ on Mig and IP-10 expression as did littermate mice expressing TNF-␣. Thus, unlike with fibroblasts, the synergistic effect of HA and IFN-␥ on Mig expression is independent of TNF-␣.
The synergistic induction of Mig and IP-10 by HA and IFN-␥ is due to the up-regulation of Mig and IP-10 gene transcription. Nuclear run-on studies in the MH-S cells show that HA does not induce Mig mRNA transcription and only minimally induces IP-10. IFN-␥ alone has minimal effect on Mig and IP-10 transcription compared with unstimulated cells. However, cells stimulated with both HA and IFN-␥ have a marked enhance-ment of Mig and IP-10 gene transcription. Similarly, when MH-S cells were transfected with a Mig promoter construct containing the 5Ј regulatory region previously shown to convey the IFN-␥ responsiveness, there was over a 32-fold induction in Mig promoter activity compared with only a 12-fold induction in promoter activity by IFN-␥ alone. Similarly, when MH-S cells were transfected with a IP-10 promoter construct containing the 5Ј regulatory region previously shown to convey the IFN-␥ responsiveness, there was over a 28-fold induction in IP-10 promoter activity compared with only an 8-fold induction in promoter activity by IFN-␥ alone. Furthermore, in the case of Mig, the ␥RE, although enough to convey responsiveness to IFN-␥, failed to account for the HA-and IFN-␥-induced synergistic enhancement of Mig gene expression. Thus, the ␥RE is not sufficient to allow for the synergistic effects of HA and IFN-␥. Together, these results suggest that HA profoundly enhances IFN-␥ induced Mig and IP-10 steady-state mRNA expression and does so at the level of Mig and IP-10 gene transcription.
Recent studies have suggested that IFN-␥ induction of Mig is mediated by the transcription factor ␥RF-1 (50). ␥RE-1, the unique IFN-␥-responsive cis element in the Mig promoter, is an imperfect palindrome consisting of the following sequence: 5Ј-TTxxxATAAACxxxxxGTTTATXXXAA-3Ј (50). ␥RF-1 has recently been shown to consist of a complex containing STAT-1␣ (49). The electrophoretic and binding properties of ␥RF-1 are distinct, however, from the previously described dimeric form of STAT-1 also known as ␥-interferon-activated factor (GAF) (reviewed in Ref. 52). In contrast to the ability of GAF to bind to a monomeric STAT-1 binding site, purified ␥RF-1 does not bind to the STAT-1 site (49) but rather is specific for the imperfect palindrome resembling a STAT binding site. These studies are further supported by the report by Ohmori et al. (41), which has shown that in STAT 1-deficient mice, IFN-␥ fails to induce Mig gene expression in fibroblasts (41). One possible mechanism for synergy between HA and IFN-␥ could be enhanced phosphorylation of STAT-1 in the presence of HA. We investigated this possibility and found no effect of HA on IFN-␥-induced phosphorylation of STAT-1 (data not shown).
Similarly, investigators have defined the IFN-␥-responsive element in the IP-10 promoter in macrophages (40). The IP-10 promoter contains an interferon stimulus response element, as well as two B sites located at Ϫ228 to Ϫ102 in the 5Ј proximal promoter (40). Although the interferon stimulus response element is necessary and sufficient for IFN-␥ induced IP-10 gene expression, optimal response to IFN-␥ required both the interferon stimulus response element and one of two B sites.
Low molecular mass HA fragments have been shown to activate the NF-B/I-B␣ transcriptional regulatory system (53). Interestingly, analysis of the Mig promoter reveals that there are three NF-B binding sites on the 5Ј promoter downstream from the ␥RE-1 and, as noted, two NF-B sites downstream from the interferon stimulus response element on the 5Ј IP-10 promoter. Thus, HA may be synergizing with IFN-␥ to induce chemokine gene expression through interactions between a STAT-1␣ like protein and/or NF-B for Mig and IP-10 gene expression. Studies to address this possibility are ongoing.
Interactions between ECM and cytokines in regulating inflammatory gene expression may have an important role in determining the resolution of an inflammatory response. Recent studies have shown that HA and IFN-␥ promote the expression of macrophage-derived IL-12 and reactive nitrogen intermediates (14,15) while abrogating the ability of HA to induce MIP-1␣, MIP-1␤, and KC (51). These new data add Mig and IP-10 to the subgroup of chemokines produced in the presence of low molecular mass HA and IFN-␥. It is interesting to speculate that introducing IFN-␥ at the appropriate time in the inflammatory response will promote resolution through the direct effects on macrophage chemokine production. Thus, understanding the molecular mechanisms regulating the interaction between ECM-and IFN-␥-'inducible genes may lead to new approaches to ameliorate chronic inflammation.