Adenovirus-mediated Gene Transfer of Mutated IκB Kinase and IκBα Reveal NF-κB-dependent as Well as NF-κB-independent Pathways of HAS1 Activation*

It has become increasingly clear that hyaluronan is more than the simple matrix molecule it was once thought to be but instead takes part in a multitude of biological functions. Three genes encode for hyaluronan synthases (HAS). We demonstrated earlier that HAS2 and HAS3 are constitutively activated in type-B synoviocytes (fibroblast-like synoviocytes) and, furthermore, that the only gene that readily responds to stimulation with a series of proinflammatory cytokines is HAS1. Here we probe the involvement of the transcription factor NF-κB in induced and noninduced HAS activation. Transforming growth factor (TGF) β1 as well as interleukin (IL)-1β are both strong inducers of HAS1 transcription. Stimulation of fibroblast-like synoviocytes with IL-1β resulted in rapid degradation of IκBα, an event that was preceded by IκBα phosphorylation. Interestingly, TGFβ1 neither affected IκBα levels, nor did it cause phosphorylation of IκBα. In addition, TGFβ1 had no effect on IκBβ and IκBϵ levels. Electrophorectic mobility shift assays demonstrate that IL-1β is a potent inducer of NF-κB translocation; however, TGFβ1 treatment did not result in shifting bands. Two adenovirus constructs were used to further clarify differences in TGFβ1- and IL-1β-induced HAS1 activation. Overexpressing IκBα completely abolished the IL-1β effect on HAS1 but did not interfere with TGFβ1-induced HAS1 mRNA accumulation. Identical results were obtained when a dominant negative IKK was overexpressed. Interestingly, neither overexpression of IκBα nor of IKK had any effect on HAS2 and HAS3 mRNA levels. Taken together, HAS1 can be activated by distinct pathways; IL-1β utilizes NF-κB, and TGFβ1 does not. Furthermore, HAS2 and HAS3 are activated without the involvement of NF-κB.

It has become increasingly clear that hyaluronan is more than the simple matrix molecule it was once thought to be but instead takes part in a multitude of biological functions. Three genes encode for hyaluronan synthases (HAS). We demonstrated earlier that HAS2 and HAS3 are constitutively activated in type-B synoviocytes (fibroblast-like synoviocytes) and, furthermore, that the only gene that readily responds to stimulation with a series of proinflammatory cytokines is HAS1. Here we probe the involvement of the transcription factor NF-B in induced and noninduced HAS activation. Transforming growth factor (TGF) ␤1 as well as interleukin (IL)-1␤ are both strong inducers of HAS1 transcription. Stimulation of fibroblast-like synoviocytes with IL-1␤ resulted in rapid degradation of IB␣, an event that was preceded by IB␣ phosphorylation. Interestingly, TGF␤1 neither affected IB␣ levels, nor did it cause phosphorylation of IB␣. In addition, TGF␤1 had no effect on IB␤ and IB⑀ levels. Electrophorectic mobility shift assays demonstrate that IL-1␤ is a potent inducer of NF-B translocation; however, TGF␤1 treatment did not result in shifting bands. Two adenovirus constructs were used to further clarify differences in TGF␤1-and IL-1␤-induced HAS1 activation. Overexpressing IB␣ completely abolished the IL-1␤ effect on HAS1 but did not interfere with TGF␤1-induced HAS1 mRNA accumulation. Identical results were obtained when a dominant negative IKK was overexpressed. Interestingly, neither overexpression of IB␣ nor of IKK had any effect on HAS2 and HAS3 mRNA levels. Taken together, HAS1 can be activated by distinct pathways; IL-1␤ utilizes NF-B, and TGF␤1 does not. Furthermore, HAS2 and HAS3 are activated without the involvement of NF-B.
Over the past few years, the list describing essential functions of hyaluronan (HA) 2 in physiological as well as pathophysiological events has become increasingly longer (1)(2)(3). Hyaluronan, a major component of the extracellular matrix, plays a role in cell proliferation, cell migration, inflammation, tumorogenesis, angiogenesis, and in embryonic development (4,5). Among the ailments strongly associated with unfettered HA production is rheumatoid arthritis (RA). In healthy joints, HA levels are low but play a vital role for proper joint function. Nevertheless, HA levels in RA affected joints can be enormous, even resulting in measurably elevated concentrations of HA in plasma (6,7).
Unfettered HA release, in association with inflammatory events, eventually results in degradation of the long chain HA molecules. Although little is known about the differences in physiological function of nondegraded HA molecules, it has been evident for some time that HA fragments exert a series of undesirable effects that might be linked to the progression of RA. Among these are the chemoattractant properties, induction of angiogenesis and the activation of proinflammatory genes (8,9). For these reasons it seems vital to widen our understanding of the mechanisms involved in the regulation of the genes encoding HA synthases (HAS). Three genes, encoding for a plasma membrane protein, are responsible for HA synthesis (10). We demonstrated earlier that in type B synoviocytes (FLS) HAS2 and HAS3 are constitutively activated and that in this cell type HAS1 can be induced by a series of cytokines. We also showed that the p38 MAPK plays an essential role in cytokine induced HAS1 transcription (11).
Despite the considerable progress made in HA biology, still very little is known about intracellular signaling pathways participating in the regulation of these genes. Here we describe parts of our efforts to test the involvement of certain transcription factors in the activation of HAS genes. A broader insight into signaling pathways might provide hints at the function of the HAS genes as well as offer new ways to test the involvement of this important group of genes in RA and other diseases.
Many transcription factors have been said to take part in the regulation of genes. Among these factors, nuclear factor B (NF-B) seems to stand out for the central role this transcription factor plays in the activation of most proinflammatory genes (12,13). With regard to RA, it has been recognized that NF-B also plays an essential role in both inflammatory and destructive mechanisms associated with joint destruction (14,15). Taken together, a large body of research suggests that genes that depend on NF-B for their activation are linked to the onset and/or progression of inflammation. Based on this assumption, we investigated the requirement of NF-B activation for the genes encoding HAS in FLS. The experiments reported here are a continuation of our efforts to further understanding of the effects of TGF␤, IL-1␤, and TNF␣ on HAS activation.
Cell Culture-Human type B (fibroblast-like) synoviocytes were purchased from Dominion Pharmakine (Derio, Bizkaia, Spain). FLS were cultured as previously described (11). In brief, FLS were propagated in T75 tissue culture flasks or culture dishes (Iwaki, Funabashi, Chiba, Japan) (15 cm in diameter) in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% heat-inactivated fetal bovine serum (Sigma), L-glutamine, and 50 units/ml penicillin/streptomycin. The medium was changed every 3 days. For experiments, FLS were detached using trypsin and transferred to 6-or 24-well plates (Iwaki). Viability of cells was confirmed by phase contrast microscopy and occasionally by staining cells with trypan blue.
Adenovirus-mediated Gene Transfer of Mutant mIKK and IB␣-Purified adenoviruses, containing the expression construct for dominant negative mutant IB kinase 2 (mIKK), IB␣, as well as green fluorescent protein (GFP), were a generous gift from Prof. R. de Martin (Medical University of Vienna, Vienna, Austria) (16 -18). FLS were incubated with serial dilutions of each adenovirus to determine optimal concentrations for gene expression in this cell type. The transfection efficiency was monitored on a fluorescence microscope and/or by Western blot after 3 days in culture. In subsequent experiments a dilution of the stock material of 1:10 4 was used, and in some experiments a dilution of 1:10 5 was chosen.
Western Blot Experiments-For SDS-PAGE and Western blotting, the cells were washed twice in ice-cold PBS and subsequently dissolved in SDS sample buffer (62.5 mM Tris/HCl, pH 6.8, 2% w/v SDS, 10% glycerol, 50 mM dithiothreitol, 0.01% w/v bromphenol blue). For 3-cm culture dishes and 10-cm dishes, 100 and 500 l of SB, respectively, was used. Aliquots of whole cell protein extract (10 -25 l/well) were separated on a mini gel (10%). The proteins were blotted onto polyvinylidene difluoride membranes (Amersham Biosciences) using a semi-dry apparatus (Bio-Rad). The blots were flashed with double distilled H 2 O, dipped into MetOH, and dried for 20 min before proceeding with the next steps. Subsequently, the blots were transferred to a blocking buffer solution (1ϫ PBS, 0.1% Tween 20, 5% w/v nonfat dry milk) and incubated for 1 h. The membranes were then incubated with the appropriate diluted primary antibody in 5% bovine serum albumin, 1ϫ PBS, and 0.1% Tween 20 at 4°C overnight in a roller bottle. Following three washing steps in wash buffer (1ϫ PBS, 0.1% Tween 20), the blots were incubated with appropriate secondary antibodies diluted in PBS. After 45 min of gentle agitation, the blots were washed five times for 5 min in wash buffer, and the proteins were made visible using either LumiGLO (New England Biolabs, Beverly, MA) or Renaissance Plus (PerkinElmer Life Sciences) and Kodak BioMax MR films.
RNA Isolation and Reverse Transcription-RNA isolation and reverse transcription were performed as described (11). Small aliquots of RNA were used to check the quality of RNA using agarose gel and ethidium bromide or Vistagreen (Molecular Probes, Eugene, OR) for visualization. First strand cDNA synthesis was performed exactly as described by the supplier of the RT-PCR Kit (Amersham Biosciences) using 1 g of RNA/reaction.
RT-PCR and Data Analysis-Experiments on the block cycler were performed, analyzed, and quantitated as described (11). Real time RT-PCR and data analysis were done as follows: RNA was resuspended in water, quantitated on a spectrophotometer, and set to equal concentrations. Reverse transcription was performed using the Strata Script First Strand Synthesis System (Stratagene, Amsterdam, The Netherlands). Total RNA (500 ng) was transcribed, using random hexamer primers, and diluted with water to a final concentration of 250 ng to improve amplification efficiency. For real time RT-PCR, one-step SYBR Green RT-PCR amplification was conducted on a MX3000P real time PCR system (Stratagene), using Brilliant SYBR Green QPCR Master mix (Stratagene). After optimization of the primer pairs, RT-PCR was performed using the following standard conditions: initial denaturation for 10 min at 95°C, denaturation for 10 s, annealing for 15 s at 57°C, and extension for 15 s at 72°C. Each RT-PCR experiment included a dissociation curve to verify the specificity of the amplicon, as well as no-template controls.
mRNA for hypoxanthine-guanine phosphoribosyl transferase 1 (HPRT) was co-amplified and used as a control for quantification. Primers used in real time PCR experiments were selected based on published data (19,20), ordered from MWG Biotech AG (Ebersberg, Germany), and dissolved at a concentration of 100 pmol/l in Tris-EDTA. Electrophoretic Mobility Shift Assay-Nuclear extracts from FLS were prepared as described (21). The double-stranded oligonucleotides used in all of the experiments were end-labeled using T4 polynucleotide kinase and [␥-32 P]ATP. After labeling and purification by chromatography, 5 g of nuclear extract was incubated with 100,000 cpm of labeled probe in the presence of 1.5 g of poly(dI-dC) at room temperature for 20 -30 min followed by separation of this mixture on a 6% polyacrylamide gel in Tris/glycine/EDTA buffer at pH 8.5. As also published previously (21), a series of control experiments were performed that confirmed the specificity of EMSA.

HAS mRNA Expression Profile in FLS-Stimulation of FLS with a series of proinflammatory stimuli leads to the activation of HAS1.
Although the genes HAS2 and HAS3 are constitutively expressed in FLS, in unstimulated FLS mRNA levels of HAS1 are very low or undetectable. FLS respond to a series of stimuli with up-regulation of HAS1 (11). Shown in Fig. 1 is a representative experiment demonstrating the effects of TNF (5 ng/ml), IL-1 (1 ng/ml), and TGF (1 ng/ml) on mRNA levels of HAS1, HAS2, and HAS3. FLS were exposed to these stimuli for 10 h. Aliquots of the resulting RT-PCR were separated on an agarose gel and scanned on a fluorescence imaging device. Although HAS2 and HAS3 mRNA levels are already high in unstimulated cells, HAS1 mRNA levels increase manifold in response to all of the above stimuli. In parentheses, indicated to the right of Fig. 1, are the number of PCR cycles at which the PCRs were terminated.
IB Degradation Following IL-1, TNF, and TGF Treatment-In resting cells, IB is bound to the NF-B protein complex and plays an important role in preventing its translocation to the nucleus. Activation of the NF-B pathway is essential for the activation of most proinflam-matory genes (22). We used Western blot experiments to investigate whether exposure of FLS to IL-1 (1 ng/ml), TNF (5 ng/ml), and TGF (1 ng/ml) will lead to the degradation of IB␣. Presented in the upper panel of Fig Treatment of FLS with IL-1 but Not with TGF Leads to IB␣ Phosphorylation-IB␣ degradation is preceded by phosphorylation of IB␣, tagging it for subsequent degradation by specific enzymes. Measuring phosphorylation of IB␣ provides another means to assess the involvement of the NF-B signaling pathway by a given stimulus. FLS were treated with IL-1 and TGF for 10 min. Whole cell protein extracts were separated by PAGE. The results of such Western blots, using phospho-specific anti-IB␣ antibodies, reveal that stimulation with TGF does not result in phosphorylation of IB␣. As demonstrated in the lower panel of Fig. 2A, treating FLS for 10 min with IL-1 leads to a strong phosphorylation specific signal. Contrary to the result obtained with IL-1, IB␣ phosphorylation was detected neither in TGF-treated cells nor in resting cells.
It is conceivable that there are differences in the time course of IB␣ degradation induced by differences in the upstream signaling pathways utilized by these two stimuli. Shown in Fig. 2B is a representative experiment where IB␣ degradation was monitored for an extended period of time. Similar to IL-1, TNF treatment of FLS for 10 min also results in complete degradation of IB␣ within a very short time frame. More importantly, TGF treatment did not result in reduced IB␣ levels when those levels were monitored for up to 1 h. Shown in this particular experiment are the IB␣ levels at 0, 10, 20, and 50 min after TGF (1 ng/ml) stimulation. Equal loading is demonstrated by staining a second blot with a control antibody (labeled Contr).
Other members of the IB protein family might be involved in TGFinduced signaling. A Western blot experiment investigating the effect of TGF treatment on IB␣, IB␤, and IB⑀ is shown in Fig. 2C. FLS were exposed to TGF and TNF, respectively, for 10 min. Similar to the results of the Western blot experiment shown in Fig. 2B, TNF treatment for 10 min resulted in complete degradation of IB␣ but left IB␤ and IB⑀ unaffected. TGF treatment, on the other hand, had no effect on any of the IB proteins monitored.  As shown in this RT-PCR experiment, in resting cells mRNA levels for HAS1 are barely detectable, but high mRNA levels for HAS2 and HAS3 are evident. Also, TNF␣, IL-1, and TGF are potent inducers of HAS1 transcription. Although mRNA for HAS2 and HAS3 are not significantly affected, stimulation with the above reagent leads to a dramatic increase in measurable HAS1 mRNA levels. FLS were left untreated (MED) or stimulated for 10 h with 5 ng/ml TNF␣, 1 ng/ml IL-1, or 1 ng/ml TGF. mRNA for actin was co-amplified as quality control. The values given in parentheses are the number of cycles at which PCR was terminated.

FIGURE 2. Effects of TGF, IL-1, and TNF␣ on IB phosphorylation and degradation.
A comparison of the IB␣ specific band in the upper panel of A demonstrates that IL-1 (1 ng/ml) treatment leads to complete degradation of IB␣. As shown here, levels of IB␣ in TGF (1 ng/ml)-treated FLS are similar to the levels found in untreated cells (MED). Furthermore, as shown in the lower section of A, IL-1 induces IB␣ phosphorylation, but TGF treatment is without effect. FLS were left untreated or stimulated for 10 min. In addition, the blots were also probed for tubulin to confirm equal loading. The use of specific antibodies for the nonphosphorylated (IB␣) and the phosphorylated (p-IB␣) forms of IB is indicated on the right. IB degradation is a very early event in intracellular signaling cascades. Nonetheless, there is no evidence for TGF-induced IB degradation, even if levels of IB␣ were monitored for an extended period of time. Shown in B is an experiment where IB␣ levels were examined at 0, 10, 20, and 50 min following stimulation with TGF. Shown in C are data demonstrating that TGF neither affected IB␣, IB␤, nor IB⑀. FLS were stimulated for 10 min with 1 or 10 ng/ml TGF (lanes TGF (1) and TGF (10), respectively) or TNF␣ (5 ng/ml). cells as controls, FLS were treated with IL-1 (5 ng/ml) for 45 min, after which nuclear protein extract was isolated und used for a series of experiments. First, supershift experiments were performed utilizing antibodies specific for NF-B family members. Aliquots of nuclear extracts of IL-1-treated FLS were used without the addition of antibodies (Fig. 3B, none) or were preincubated with antibodies directed against NF-B p50, p65, p52, c-Rel, and RelB for 30 min at room temperature prior to the addition of labeled oligonucleotides. Such experiments reveal the presence of the prototypical NF-B p65/p50 heterodimer complex in IL-1-stimulated FLS. Also shown in Fig. 3B are experiments demonstrating that the addition of 500-fold excess of unlabeled homologous NF-B oligonucleotides are able to compete for NF-〉-DNA binding (lane cold NF-B). On the other hand, the addition of 500-fold excess of unrelated AP-1-oligonucleotides (lane cold AP-1) was without any effect on NF-B-DNA interactions. Similarly, adding similar amounts of unlabeled SP1 or CRE oligonucleotides was without effect on NF-B-DNA interactions (data not shown).
TGF Effects on the Gene HAS1 Are Not Mediated through the Activation of PKC nor through the Activation of the Transcription Factors SMAD3/4 and STAT3-We demonstrated earlier that 0.5-1 ng/ml TGF is sufficient to maximally activate HAS1 transcription (11). Interestingly, increasing the concentration of TGF to 10 ng/ml nearly abolished the effect on HAS1 transcription seen at lower concentrations of TGF. Like TGF, phorbol myristate acetate (PMA) activates HAS1 transcription in FLS (11). PMA is known to be a potent activator of PKC (23). Similarly, PKC-dependent activities have been reported in TGF-induced gene expression (24,25). For this reason we tested whether TGF might exert its effects on HAS1 through the activation of PKC. Staurosporine, a broad spectrum inhibitor of protein kinases, inhibits PKC (IC 50 ϭ 50 pM), cAMP-dependent protein kinase (IC 50 ϭ 7.5 nM), as well as protein kinase G (IC 50 ϭ 8.5 nM). In a series of experiments, FLS were preincubated with staurosporine at concentrations ranging from 0.5 nM (10 ϫ IC 50 for PKC) to 100 nM (more than 10 ϫ IC 50 for protein kinase G) for 45 min, TGF (1 ng/ml) was added for an additional 10 h at which RNA was isolated, and RT-PCR was performed. We also utilized the PKC inhibitor RO-31-8220 (at the final concentration of 50 nM) as well as the cAMP-dependent protein kinase inhibitor KT 5720 (at the final concentration of 250 nM). No inhibition of TGF (1 ng/ml)-mediated HAS1 activation could be seen in such experiments (data not shown). Remarkably, blocking protein kinases with staurosporine actually resulted in consistently elevated levels of HAS1 mRNA detected in otherwise untreated FLS (data not shown).
The activation of members of the SMAD family is among the best studied pathways involved in TGF-mediated intracellular signaling events. Innumerable studies describe the utilization of SMAD proteins as the main intracellular signaling pathways utilized by this compound (26 -28). For that reason, SMAD3 is sometimes also referred to as TGF␤ response effector SMAD3. In an attempt to investigate whether activation of SMAD proteins might account for the effect of TGF on HAS1 in FLS, the activation and nuclear translocation of SMAD1, SMAD3, and also that of STAT1 and STAT3 proteins were evaluated by EMSA. Shown in Fig. 3C is one of three EMSA experiments where FLS were left untreated or were exposed to TGF (1 and 10 ng/ml, respectively) for 45 min. Nuclear protein extract and EMSA were performed as described under "Experimental Procedures." A weak protein-DNA complex can be observed in unstimulated cells (Fig. 3, B and C, MED). Exposing FLS to 1 ng/ml TGF for 45 min did not result in a significant increase in SMAD3/4-DNA complexes. Interestingly, increasing the concentration of TGF to 10 ng/ml resulted in the activation of SMAD3/4. Similarly, STAT3 activation was not detected in FLS stimulated with 1 ng/ml TGF, yet again, using 10-fold higher concentrations of TGF led to a readily detectable increase in STAT3-DNA complexes. Although SMAD3/4 as well as STAT3 were activated in FLS stimulated with 10 ng/ml of TGF, no activation of SMAD1 and STAT1 could be observed in similar EMSA experiments using oligonucleotides resembling consensus SMAD1 and STAT1 sequences; neither 1 nor 10 ng/ml TGF treatment resulted in significant activation of these transcription factors (data not shown). Also included in Fig. 3C (far right panel) is yet another control experiment, demonstrating that PMA results in the activation of NF-B when used at a concentration (2.5 ng/ml) that significantly increases HAS1 mRNA accumulation in FLS (11).
Overexpressing IB␣ and Dominant Negative mIKK Prevents IL-1induced HAS1 Activation but Does Not Effect TGF-induced HAS mRNA Accumulation-As demonstrated many times, nuclear translocation of NF-B can be prevented by overexpression of a mutant version of IKK (16,29). Such a measure seems well suited to clarifying the involvement of NF-B in IL-1-and TGF-induced HAS activation. FLS were transfected with adenovirus constructs containing either a mutant IKK or a GFP expression sequence as described under "Experimental Procedures." Shown in Fig. 4A is such a representative experiment performed in duplicate. As indicated for Fig. 4A, FLS (nontransfected or IB␣ transfected) were left untreated (MED) or stimulated with IL-1 and TGF, respectively, for 10 h. The effect of preventing NF-B translocation was studied by comparing induced and noninduced HAS1, HAS2, and HAS3 mRNA levels.
Such experiments demonstrated that overexpression of IB␣ affected only HAS1 mRNA levels in significant ways. Stimulating FLS with IL-1 and TGF treatment results in Ͼ50and Ͼ30-fold higher levels of HAS1 mRNA, respectively. More importantly, as shown in the panel labeled HAS1, whereas the IL-1 effect on HAS1 mRNA levels in cells expressing IB␣ is completely abolished, no inhibitory effect could be observed on TGF-induced HAS mRNA accumulation. As shown in Fig.  1, HAS2 and HAS3 mRNA levels in resting FLS are relatively high. Interestingly, overexpression of IB␣ had neither an effect on the high basal levels of these genes, nor were mRNA levels of HAS2 and HAS3 significantly altered in FLS that were stimulated with IL or TGF (Fig. 4A,  panels HAS2 and HAS3, respectively).
Activation of IKKs upstream of IB␣ plays an important role in the signaling cascade leading to the degradation of IB␣ (30,31). We used a mutant IKK2 adenovirus expression construct to validate data obtained using the IB␣ construct and to further dissect possible differences in signaling pathways in IL-1-induced versus TGF-induced HAS activation. Parts of 6-well tissue culture dishes containing FLS grown to high density were treated with the mIKK adenovirus for 3 days. Subsequently, FLS were left untreated (lane MED) or were stimulated with IL-1 and TGF, respectively, for 10 h, at which time the experiment was terminated and RNA was extracted. A comparison of the data presented in Fig. 4B reveals similar effects as observed using the IB␣ construct to block NF-B-mediated signaling. As demonstrated here again, treatment of FLS with IL-1 and TGF results in Ͼ50and Ͼ30-fold higher levels of HAS1 mRNA, respectively, whereas levels of HAS2 and HAS3 mRNA are not significantly affected. Similar to the results presented in Fig. 4A using an IB␣ construct, overexpression of mIKK also leads to a complete inhibition of IL-1-induced HAS1 activation. Furthermore, like high levels of IB␣, overexpression of mIKK has no effect on TGFinduced activation of this gene. Again, HAS2 as well as HAS3 mRNA levels are not significantly affected by the presence of high amounts of mIKK in the cytoplasm.
Real Time PCR, Standard Curves, and PCR Analysis-Pictures depicting representative real time PCR experiments for HAS1, HAS2, and HAS3 as well as HPRT are given in Fig. 5. The name of the gene illustrated is given on the top of each panel. The lower panels show amplification curves, and the upper panels show the standard curves used for quantification (calculation of x-fold induction).
Overexpression of GFP Does Not Influence HAS1 mRNA Levels-Shown in Fig. 6A is one of three experiments demonstrating the speci-

. Effects of overexpression of IB␣ and dominant negative IKK (mIKK) on TGF-and IL-1-induced HAS mRNA levels.
Three days after transfection with the indicated adenovirus constructs, FLS were stimulated with TGF (1 ng/ml) and IL-1 (1 ng/ml) for 10 h. Real time RT-PCR was used to calculate the effect of this procedure on mRNA levels. Fig. 4 demonstrates ϳ40-fold induction of HAS1 in IL-1-stimulated cells and 20 -30-fold higher levels of HAS1 mRNA in TGF-treated cells. As shown in A, overexpressing IB␣ resulted in a complete inhibition of IL-1-induced HAS1 mRNA, but the same procedure did not affect the TGF-mediated increase of HAS1 mRNA levels. Similarly, as shown in B, overexpressing mIKK also completely blocked IL-1-induced HAS1 transcription but left TGF-induced HAS1 mRNA levels essentially unchanged. Furthermore, as shown in A as well as in B, HAS2 and HAS3 mRNA levels were largely unaffected by stimulation with IL-1 or TGF. In addition, mRNA levels of these genes were not significantly altered in cells that were transfected with the IB␣ or mIKK virus construct prior to stimulation with the above stimuli. In A, the label V-IB indicates experiments in which cells were manipulated to over express IB␣. In B, the label mIKK points at data derived from cells transfected with the IKK adenovirus construct. ficity of the mIKK-and IB␣-mediated inhibition of IL-1-induced HAS1 activation. For 3 days, FLS were incubated with adenovirus constructs (GFP, mIKK, and IB␣, respectively) after which cells were exposed to IL-1, to TGF, or were left untreated (MED). mRNA levels were quantitated by real time PCR. As demonstrated in the section labeled IL-1, overexpression of mIKK as well as IB␣ completely abolished the 60-fold induction. More importantly, overexpression of GFP in these cells had no effect on HAS1 mRNA levels induced by IL-1 treatment. Fig. 6A demonstrates again that TGF-induced HAS1 mRNA levels are not affected by overexpressing mIKK or IB␣, respectively. This figure clearly demonstrates that HAS1 mRNA levels are affected neither by the presence of the virus construct itself, nor by the high concentrations of GFP present in transfected cells.
Shown in Fig. 6B is a control experiment demonstrating the effectiveness of the virus construct as a means of inducing high levels of IB␣ protein in the cytoplasm of cells transfected with IB␣ adenovirus expression construct. In this particular experiment, FLS (two wells each) were left untreated (lanes MED), transfected with the IB␣ adenovirus construct (lanes V-IB), or with the GFP virus construct (lanes V-GFP) and incubated for 3 days. Whole cell extract was prepared and separated by SDS-PAGE. As demonstrated in this figure, levels of IB␣ in cells exposed to the IB␣ adenovirus construct are significantly higher than those in untreated cells. Furthermore, treating cells with the GFP adenovirus construct does not lead to changes in the IB␣ levels.

DISCUSSION
The genes encoding HA seem to be implicated in a series of ailments ranging from cancer to rheumatic disorders, but these genes also play a crucial role in a variety of physiological processes ranging from embryogenesis to essential functions of the immune system (3)(4)(5)(32)(33)(34). Very little is known about the intracellular and extracellular mechanisms involved in the regulation of these genes. In the present study we inves-  DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 tigated the involvement of the transcription factor NF-B in induced and noninduced HAS activation in in vitro experiments using FLS.

NF-B and HAS1 Activation
We reported earlier that in FLS, of the three HAS genes, HAS1 is the gene that readily responds to stimulation with a series of cytokines including IL-1␣, IL-1␤, IL-8, TGF, and TNF␣ (11). HAS1 mRNA levels are very low in resting FLS but mRNA for HAS2 and HAS3 is readily detectable in resting cells. Interestingly, HAS1 was the only gene whose mRNA levels were significantly altered in response to the above stimuli.
Here we report that the gene HAS1 can be activated in a NF-B-dependent as well as in a NF-B-independent manner. Although the utilization of the NF-B pathway is a prerequisite for IL-1-induced HAS1 activation, TGF-induced activation of HAS1 occurs independently of NF-B. These conclusions are based on a series of experiments analyzing the activation of the NF-B pathway in response to TGF and IL-1.
As shown here, stimulating FLS with IL-1 resulted in rapid phosphorylation and subsequent degradation of IB␣. In contrast, TGF treatment had no effect on IB␣. Furthermore, TGF treatment did not lead to the phosphorylation of IB␣, which in most cases is a prerequisite for subsequent IB␣ degradation. Because it is feasible that in FLS other IB members might play a dominant role in the retention/release of NF-B, the effect of TGF on IB␤ and IB⑀ was investigated as well. Similar to IB␣ levels, TGF treatment of FLS did not lead to any significant changes in cytoplasm levels of those IB proteins.
Although in most reports IB␣ degradation has been shown to be essential for the release and the subsequent translocation of the NF-B complex into the nucleus, there are a few reports demonstrating the activation of NF-B without prior degradation of IB␣ (35). To test the possibility of TGF-induced NF-B activation without prior degradation of IB, EMSA experiments were performed. The results of these experiments clearly demonstrate that although IL-1 induces the translocation of NF-B into the nucleus as well as subsequent binding to oligonucleotides resembling consensus NF-B binding sites, TGF does not.
EMSA are to a certain degree artificial systems that might not entirely mimic in vivo conditions. We therefore resorted to using two adenovi-rus constructs that would allow us to validate the interpretation of Western blot and EMSA experiments. First, FLS were transfected with an adenovirus construct containing the expression sequence for IB␣. As demonstrated in Fig. 6B, at the time of the experiment, high levels of IB␣ could be demonstrated by Western blot. More importantly, overexpressing IB␣ led to complete inhibition of IL-1-induced HAS1 transcription. On the other hand, the ϳ30-fold higher HAS1 mRNA levels induced by TGF treatment were not reduced. Identical results were obtained when instead of the IB␣ construct, an adenovirus gene transfer system was used that results in the expression of a mutant kinasenegative version of IKK. That this kinase, like IB, plays an essential role in the NF-B-mediated signaling pathway has been shown before (30,31).
The outcome of these experiments yet again demonstrates the involvement of NF-B in IL-1-induced HAS1 transcription and also bolsters the conclusion that TGF-induced signaling leading to HAS1 activation indeed occurs independently of NF-B. The effects of overexpression of IB␣ and mIKK are attributable only to the elevated levels of these proteins and are not mediated either by nonspecific effects of the virus construct itself or by the transfection conditions used. This conclusion is supported by the demonstration that the adenovirus-GFB construct affects neither IL-1-nor TGF-induced HAS1 activation.
We described earlier that the maximal level of HAS1 mRNA can be observed in FLS stimulated with TGF at a concentration of 0.5-1 ng/ml (11). We also demonstrated that increasing the amount of TGF to 10 ng/ml not only reduced the stimulatory effect on HAS1 significantly but also reduced levels of the constitutively elevated HAS3 mRNA in FLS. That the TGF effect greatly depends on the concentration of TGF as well as on the effector cells used has been shown in numerous studies (36 -38). It is also of interest that Westergren-Thorsson et al. (39) reported that TGF enhances HA production in lung but not in skin fibroblasts. Although Sugiyama et al. (40) reported up-regulation of the gene HAS1 in skin fibroblasts stimulated by TGF, several reports confirm our findings and reported TGF-induced activation of HA release in FLS (41,42). Others also studying the effect of TGF on HA reported no effect of TGF on HA or even demonstrated TGF-induced inhibition of HA release (43). Although likely, it is currently not clear whether the above discrepancies can be explained solely by differences in the concentration, the length of exposure, or the biological activity of TGF used.
Because TGF does not activate the NF-B pathway, we performed a series of experiments aimed at a better understanding of the mechanisms that account for TGF induced HAS1 activation. Although PKC has been implicated in TGF-induced signaling pathways, our experiments blocking PKC with staurosporine indicate that in FLS, PKC is not involved in TGF-mediated HAS1 activation. Similarly, a cAMPdependent protein kinase as well as a protein kinase G inhibitor also failed to reduce TGF-mediated HAS1 mRNA accumulation. In addition, as demonstrated by the representative EMSA experiment shown in Fig. 3C, SMAD3/4 as well as STAT3 can be activated in FLS. However, at the concentrations of TGF that lead to induction of HAS1, TGF neither activated SMAD3/4 nor STAT3. This is surprising, because especially proteins of the SMAD family have been described to be closely associated with TGF-induced intracellular signaling (26,28,44).
The fact that in FLS, HAS2 and HAS3 are constitutively activated speaks against a role of NF-B in the activation of these genes. With few exceptions e.g. in B-cells, NF-B is not activated in resting cells (45). The data presented here also support the conclusion that in FLS neither HAS2 nor HAS3 depend on the activation of NF-B. As demonstrated in Fig. 1, mRNA levels of HAS2 and HAS3 in unstimulated FLS are indeed high, but EMSA experiments do not demonstrate significant Shown in A is a representative experiment demonstrating that overexpression of GFP with an adenovirus construct affected neither IL-1-nor TGF-induced HAS1 mRNA levels. As in Fig. 4, data in A again demonstrate that IB␣ and mIKK do not affect TGF effects but efficiently blocked IL-1-induced HAS1 transcription. Shown in B are data demonstrating the efficiency of the adenovirus construct to overexpress IB␣. FLS were transfected with the IB␣ and the GFP virus construct. After 3 days, transfection efficiency for GFP was confirmed by fluorescence microscopy (data not shown) and IB␣ by Western blot experiments. Protein extract of cells exposed to the IB␣ and adenovirus GFP construct are indicated by the labels V-IB and V-GFP, respectively. Extracts from untreated cells are marked with the label MED.
NF-B-DNA interactions in nuclear proteins isolated from resting FLS. Furthermore, neither HAS2 nor HAS3 mRNA levels decrease in response to obstruction of the NF-B signaling pathway through overexpression of mIKK and IB␣, respectively.
The very nature of the HA molecules has complicated and considerably hampered investigations in functional differences of the HAS gene products. Therefore, very few studies directly address biological effects that might be attributable to dissimilarities of HA molecules. Nevertheless, we have previously speculated about the fact that HAS1 is the sole HA encoding gene that readily responds to a series of proinflammatory cytokines (11). If a generalization from published data is legitimate, and the assumption of the NF-B dependence of proinflammatory genes also holds true for HA encoding genes, HAS1 would fulfill yet another criterion to justify its classification as a proinflammatory gene. Currently it is unclear which of the intracellular signaling pathways are utilized in NF-B-independent HAS1 activation. Nevertheless, similar phenomena have been observed by others studying mediators of inflammation. It has been shown, for example, that TNF␣, IL-1, IL-6, and IL-8 can also be regulated in both NF-B-dependent and NF-B-independent ways (46).
The critical role of IL-1 in the propagation of RA and inflammation in general is well accepted (47,48). On the other hand, participation of TGF in inflammatory processes is less recognized, although certain observations clearly link TGF to inflammation (49,50). With regard to RA, not only are there reports demonstrating that TGF is indeed abundant in the rheumatoid synovium (51), it has also been shown that TGF is able to exert a series of detrimental effects on FLS isolated from RA patients. Among these is the activation of IL-1, TNF, IL-8, MIP1␣, and MMP-1. In addition, most of these effects of TGF seem specific to the arthritic synovial fibroblasts (52).
Concentrations of HA are considerably higher in the affected joints and plasma of patients suffering form RA than in both healthy subjects and in osteoarthritis patients (7). Still, to date, no causal link between up-regulation of HAS genes and the subsequent infiltration of leukocytes in affected joints in RA has been established. It nevertheless seems plausible that unfettered HA release contributes to initiation and/or propagation of RA. Such a view is also strongly supported by the recent demonstration that HA can cause all classical signs of RA (53). Taken together, these data seem to support our working hypothesis that unregulated HA metabolism does play a significant role in initiation and/or progression of RA.
In summary, we demonstrate the existence of two independent pathways of HAS1 activation. In addition, our data also seem to rule out a role of the transcription factor NF-B in the activation of the genes HAS2 and HAS3. The presented data might offer new insight in TGFinduced signaling pathways and contribute to a better understanding of regulatory mechanisms in HA biology.