Despite evidence that points to unfettered hyaluronic acid (HA) production as a culprit in the progression of rheumatic disorders, little is known about differences in regulation and biological functions of the three hyaluronan synthase (HAS) genes. Testing the effects of drugs with proven anti-inflammatory effects could help to clarify biological functions of these genes. In this study, we demonstrate that leflunomide suppresses HA release in fibroblast-like synoviocytes (FLS) in a dose-dependent manner. We further demonstrate that leflunomide suppresses HA synthase activity, as determined by 14C-glucuronic acid incorporation assays. Additional experiments revealed that in FLS, leflunomide specifically blocked the induction of HAS1. HAS2 and HAS3, genes that are, in contrast to HAS1, constitutively expressed in FLS, are not significantly affected. Leflunomide can function as a NF-κB inhibitor. However, EMSA experiments demonstrate that at the concentrations used, leflunomide neither interferes with IL-1β- nor with PMA-induced NF-κB translocation. Furthermore, reconstituting the pyrimidine synthase pathway did not lead to the restoration of IL-1β-induced HAS1 activation. More importantly, two tyrosine kinase inhibitors mimicked the effect of leflunomide in that both blocked IL-1β-induced HAS1 activation without affecting HAS2 or HAS3. These data point at HAS1 activation as the possible cause for unfettered HA production in rheumatoid arthritis and might explain, at least in part, the beneficial effects of leflunomide treatment. These findings also support the concept that IL-1β-induced HAS1 activation depends on the activation of tyrosine kinases, and indicate that leflunomide blocks HA release by suppressing tyrosine kinases rather than through inhibition of NF-κB translocation.
Rheumatoid arthritis (RA), 3 a disorder of unknown etiology, affects large parts of the population (1, 2, 3, 4). This debilitating disease is characterized by chronic inflammation of affected joints caused by infiltrating cells. Structural damage is caused by a series of well-described mediators of inflammation. Among the known participants in inflammatory processes are adhesion molecules, cytokines, and chemokines. Although the consequences of chronic inflammation are well described, very little is known about early events that cause, for example, unwanted migration of leukocytes into joints.
Controlled hyaluronic acid (HA) production/release is undoubtedly essential for many physiological mechanisms, for example, the proper functioning of joints. Nevertheless, unfettered HA production results in many detrimental effects and might directly as well as indirectly contribute to the progression of RA. This hypothesis is supported by a series of findings, demonstrating for example that increased HA synthesis is nearly always associated with inflammatory reactions, irrespective of its cause (5, 6, 7, 8, 9). In addition, as has been demonstrated many times, degradation products of HA possess a series of unwanted properties such as acting as chemoattractants, inducing blood vessel growth, as well as activating and inducing the release of proinflammatory molecules in leukocytes and endothelial cells (10, 11, 12, 13, 14, 15). That detection of increased levels of HA in human synovium effusion serves as a sensitive indicator of altered connective tissue cell function has been recognized for some time (16). Furthermore, while in healthy joints only a fine layer of HA covers and protects joint surfaces, RA is characterized by uncontrolled HA production, often leading to enormous amounts of HA in affected joints. Therefore, significantly elevated levels of HA can be detected even in the serum of RA patients, and elevated plasma HA levels correlate positively with destruction of involved joints (11). As a result, it has been suggested to use HA serum levels as indicators of disease progression (17). In strong support of the hypothesis that HA is a key player in RA, Wang and Roehrl (18) showed in an animal model that within 50 days of weekly HA injections, animals became chronically sick, exhibiting on-and-off RA symptoms for months. The animals in this study also developed all classical pathological symptoms of RA, such as swollen joints, edema, large numbers of CD4+ cells, and infiltrating macrophages (18).
HA can be synthesized by three distinct enzymes that are the products of transcription and translation of the genes hyaluronan synthase (HAS) 1, HAS2, and HAS3 (19). Earlier, we tested the effects of a series of proinflammatory cytokines on the expression of the three HAS genes in fibroblast-like synoviocytes (FLS) (20). These studies revealed that in cultured FLS, the genes HAS2 and HAS3 are constitutively expressed, while mRNA for HAS1 is very low or undetectable. More importantly, stimulation with proinflammatory cytokines revealed that in FLS HAS1 mRNA is readily inducible, resulting in a manifold increase in HAS1 mRNA as well as in significantly elevated levels of HA as determined by ELISA. Interestingly, very little or no changes were noticed monitoring mRNA levels of HAS2 and HAS3 in such experiments. We concluded from these experiments that HAS1 may be the gene in FLS whose unfettered activation by proinflammatory cytokines might be involved in elevated HA levels associated with RA.
With the objective of expanding our understanding with regard to the involvement of the three HAS genes and their regulation in rheumatic diseases, we tested a series of drugs that have been used successfully to treat various forms of rheumatic disorders. In this work, we report the results of experiments treating FLS with leflunomide, a drug that has been shown to be very effective not only in the treatment of rheumatic disorders (21, 22, 23, 24), but also as a remedy in preventing the rejection of allografts as well as xenografts (25, 26, 27).
Materials and Methods
If not stated otherwise, reagents were from Sigma-Aldrich. The erbstatin analog, 2,5-dihydroxymethylcinnamate and genistein, were from Calbiochem. UDP-[14C]glucuronic acid (418.3 mC/mmol) was from PerkinElmer. Oligonucleotides for EMSA experiments were from Promega; HA-ELISA were from Corgenix. Abs for supershift experiments were from Santa Cruz Biotechnology.
Human fibroblast-like synoviocytes (FLS) were a gift from G. Partsch (Ludwig Boltzmann Institute, Vienna, Austria) (28) or were purchased from Dominion Pharmakine. FLS were cultured, as previously described (20). In brief, FLS were propagated in T75 tissue culture flasks or culture dishes (Iwaki, Funabashi) (15 cm diameter) in DMEM (Sigma-Aldrich) supplemented with 10% heat-inactivated FBS (Sigma-Aldrich), l-glutamine, and 50 U/ml penicillin/streptomycin. Medium was changed every 3 days. For experiments, FLS were detached using trypsin and transferred to 6- or 24-well plates (Iwaki, Funabashi). For ELISA experiments, FLS were cultured in 24-well plates, and for RT-PCR experiments in 6-well plates, respectively.
RNA isolation and RT-PCR
RNA isolation, reverse transcription, and PCR were performed, as described (20). Small aliquots of RNA were used to check the quality of RNA using agarose gel and ethidium bromide or Vistagreen (Molecular Probes) for visualization. First-strand cDNA synthesis was performed exactly as described by the supplier of the RT-PCR kit (Amersham Biosciences) using 1 μg of RNA per reaction. Aliquots were used for PCR. A Techne cycler (Techgene) and an Eppendorf cycler (Eppendorf) were used for PCR under the following standard conditions: initial denaturation, 4 min at 94°C; annealing, 55°C or 62°C (HAS3); amplification, 20 s at 72°C; denaturation, 20 s at 94°C; 21–35 cycles, followed by final extension for 10 min at 72°C. Care was taken to work out exact PCR conditions to ensure that the amplification reaction was stopped in the log phase of amplification. As control for equal usage of mRNA, either actin or GAPDH or both were used.
Primers were from MWG Biotec and were dissolved at a concentration of 100 pmol/μl in Tris-EDTA. The sequence of the primers for HAS1, HAS2, HAS3, GAPDH, and actin mRNA as well as PCR conditions were published before (20). Aliquots of PCR were separated on agarose gels. The specificity of PCR was confirmed by comparing the size of the amplified fragment with the calculated length as well as by sequencing the PCR products.
Nuclear extracts from FLS were prepared, as described (29). The double-stranded oligonucleotides used in all experiments were end labeled using T4 polynucleotide kinase and [γ-32P]ATP. After labeling and purification by chromatography, 5 μg of nuclear extract was incubated with 100,000 cpm of labeled probe in the presence of 1.5 μg of poly(dI-dC) at room temperature for 20 min, followed by separation of this mixture on a 6% polyacrylamide gel in Tris/glycine/EDTA buffer at pH 8.5. For specific competition, 7 pmol unlabeled NF-κB oligonucleotides was included, and for nonspecific competition, 7 pmol double-stranded AP-1 oligonucleotides and 7 pmol cAMP-responsive element-like (CRE) nucleotide were used. For supershift assays, 1 μl of the specific supershift Ab anti-NF-κB p56 subunit mAb (Santa Cruz Biotechnology) was added to the nuclear extract simultaneously with the labeled probe.
Aliquots of culture medium were removed at indicated time points, centrifuged (5 min at 2000 × g), and tested for the presence of HA via a procedure provided by Corgenix. OD values were used to calculate HA levels using a third-order polynominal regression analysis performed with a universal assay calculation program (AssayZap; Biosoft).
In vitro HAS assay
HAS activity was monitored using a modification of previously described methods (30). Briefly, FS, cultured in 15-cm tissue culture dishes, were washed, incubated in hypotonic lysis buffer (LB) (10 min), and harvested into 1 ml of LB supplemented with aprotinin, leupeptin, and PMSF. A Dounce homogenizer (pestle B) was used to disrupt cells. Nuclei were pelleted by spinning tubes at 1,000 × g for 4 min. Samples were centrifuged at 16,000 × g for 25 min to pellet membrane fragments. LB (50 μl) with protease inhibitors was used to resuspend membrane pellets. An aliquot, diluted in LB plus 1% SDS, was used for protein measurement using the bicinchoninic acid assay (Pierce). All of the above steps were performed in the cold room using prechilled solutions. The in vitro HA synthase assays using 25 μg of cell membrane extract were assembled exactly as described (30). After incubation for 1 h at 37°C, the reaction was stopped by boiling, and the mixtures were further incubated with or without Streptomyces hyaluronate lyase (40 turbidity-reducing units per 50-μl aliquot) at 37°C overnight and then treated with 200 mg/ml Pronase at 37°C for 5 h for deproteinization.
After boiling the samples in the presence of 1% SDS (w/v), mixtures were transferred to Microcon centrifugal filter devices that retained molecules larger than 100,000 Da (Microcon YM-100; Millipore). Unincorporated [14C]glucuronic acid was removed by filtration (5 min at 5,000 × g). Subsequently, LB (200 μl) was added to the sample reservoir. Centrifugation and resuspension of the retentate in LB were repeated three times to ensure complete removal of unincorporated [14C]glucuronic acid. After the final spin, sample reservoirs were placed upside down in a new vial and centrifuged at 1,000 × g to recover polysaccharides. Scintillation mixture was added to determine radioactivity. [14C]Glucuronic acid incorporated into hyaluronan polymer was calculated from the Streptomyces hyaluronidase-sensitive radioactivity.
Quality control and data analysis
Special care was taken to terminate PCR in the log phase of amplification. As demonstrated earlier (20, 31), a series of cycles are routinely tested to define optimal PCR conditions for a given gene. Viability of cells was confirmed by phase contrast microscopy and occasionally by staining cells with trypan blue.
Agarose gels were stained with ethidium bromide and scanned on a Fluorimager 595 (Amersham Biosciences). Data were analyzed and quantitated using ImageQuant Software (Amersham Biosciences). mRNA for GAPDH or actin or both were used as controls for RT-PCR, and scanner readings were used to recalculate PCR data.
IL-1β-induced HA release is blocked by leflunomide
FLS are cells that are able to produce and release considerable amounts of HA (12, 14, 32, 33). In this study, we demonstrate that leflunomide is able to suppress induced HA release in FLS in a dose-dependent manner. We tested the effects of leflunomide on basal and IL-1β- as well as IL-1α-induced HA release. IL-1 was added, and cells were incubated for up to 24 h. Where indicated, FLS were treated with 5 and 50 μM leflunomide only, or were pretreated with leflunomide 30–45 min before the addition of IL-1. Aliquots of cell culture supernatant were collected at times ranging from 4 to 24 h.
Shown in Fig. 1 is one representative experiment in which HA levels were quantitated by an HA-specific ELISA. Measurements were done in duplicates and demonstrate readily detectable HA levels in culture medium of unstimulated FLS. In control experiments, HA concentrations were also measured in the complete culture medium (DMEM plus 10% FCS) used to culture FLS. HA levels in complete medium were negligibly low (≤30 ng/ml). In the experiment shown in Fig. 1, HA levels were quantitated after 16 h. HA in wells containing unstimulated cells was found at concentrations of 618 ± 54 ng/ml. As shown in this figure, treating FLS with 5 μM leflunomide had no effect on noninduced HA release. Increasing the concentration of leflunomide to 50 μM (column LEF 50) led to an insignificant reduction of HA levels (618 ± 54 ng/ml HA in unstimulated vs 499 ± 47 ng/ml in 50 μM leflunomide-treated cells). Stimulating cells with IL-1β resulted in a >100% increase in measurable HA levels (1325 ± 112 ng/ml). More importantly, pretreating FLS with 50 μM leflunomide completely abolished the effect of IL-1β (1325 ± 112 ng/ml HA in IL-1β-treated vs 521 ± 54 ng/ml HA in leflunomide- and IL-1β-treated cells). Lowering the concentration of leflunomide to 5 μM still results in a reproducible and significant reduction of IL-1β-induced HA release (1325 ± 112 ng/ml vs 729 ± 68 ng/ml).
Leflunomide prevents IL-1β-activated HA synthase activity
A series of enzymes are known that can be involved in controlling levels of HA (34, 35). Activation of such hyaluronidases by leflunomide might account for the observed effect. Therefore, we subsequently tested whether the effects on HA levels noted by ELISA are a consequence of leflunomide-induced changes in HA degradation patterns, or due to leflunomide-induced changes of HA synthase activity. An in vitro HA synthase assay was established, as published previously (30). FLS were grown to high density in 15-cm tissue culture dishes and treated with IL-1β (5 ng/ml), with or without leflunomide (50 μM) for 8 h. Leflunomide was added 30–40 min before the addition of IL-1β. HA synthase assays were performed, as described (20, 30). Hyaluronidase from Streptomyces was used in control experiments that confirmed HA specificity of the assay.
As shown in Fig. 2, leflunomide inhibits IL-1β-induced HA synthase activity. Although in this particular experiment dpm in untreated cells (column Medium) were 388 ± SD 140 (SD), there was a significant increase to 1275 ± SD 205 in cells treated with IL-1β. Exposure to leflunomide before the addition of IL-1β resulted in a significant reduction to 450 ± SD 72 dpm. Two independent experiments were performed; shown are the results of one experiment done in duplicates. Decays per minute are given on the y-axis, and culture conditions are indicated on the x-axis. Differences between IL-1β-treated and leflunomide plus IL-1β-treated cells are considered significant (p = 0.033).
Leflunomide selectively blocks HAS1 and has no significant effect on HAS2 and HAS3 mRNA levels
HA synthase assays measure net effects on HA synthase activity and cannot distinguish among the three HAS genes. To gain a better understanding of the mechanisms behind leflunomide effects on HA synthase, RT-PCR experiments were performed. FLS were left untreated, preincubated with 5 or 50 μM leflunomide for 30 min before IL-1β (5 ng/ml) or PMA (2.5 ng/ml) treatment, or were treated only with IL-1β and PMA, respectively. Experiments were terminated after 6 h; total RNA was isolated; and levels of HAS1, HAS2, and HAS3 mRNA were evaluated by RT-PCR.
Shown in Fig. 3, A and B, are representative experiments demonstrating that leflunomide suppresses IL-1β-induced HAS1 mRNA in a dose-dependent manner. Furthermore, as demonstrated in Fig. 3,A, leflunomide does not exert any significant effects on mRNA levels of HAS2 or HAS3. HAS1 mRNA levels are low in quiescent, unstimulated FLS (column Medium), stimulation with PMA (column PMA), as well as with IL-1β (column IL-1β) led to a significant increase of detectable HAS1 mRNA levels. In the three independent experiments performed, leflunomide alone (column Lef) had no significant effect on any of the HAS genes. More importantly, leflunomide dramatically affected induced HAS1 mRNA accumulation. As a comparison of the columns Lef + PMA and Lef + IL-1, with controls shown in columns PMA and IL-1, demonstrates, treatment with leflunomide (50 μM) repeatedly led to a reduction of HAS1 mRNA levels by 75–100%. Shown in Fig. 3 are gel scans and graphs resulting from quantitating HAS1 mRNA levels by densitometry.
Shown in Fig. 3 B is a representative experiment that demonstrates the dose effects of leflunomide. Although in such experiments 5 μM leflunomide was sufficient to suppress IL-1β-induced HAS1 mRNA levels by 10–50%, increasing the amount of leflunomide to 50 μM resulted in 75–100% inhibition. mRNA for the gene actin was coamplified and used to adjust HAS mRNA levels.
Because IL-1β might also contribute to elevated HA levels by suppressing the synthesis of HA-degrading enzymes, we tested whether IL-1β affects the activation of HA-degrading enzymes. As reported earlier (20), mRNA for HYAL1, HYAL2, as well as PH-20 are readily detected in FLS. Of the remaining hyaluronidase genes, HYAL3 is not measurable in synoviocytes, but readily detectable in carcinoma cells used in our control experiments. HYAL4 is a chondroitinase that has no activity against HA, and HYALP1 is a pseudogene. Shown in Fig. 3 C are data demonstrating that IL-1β treatment did not affect mRNA levels of the genes encoding hyaluronidases in synoviocytes. Taken together, these experiments indicate that the elevated HA levels resulting from exposure to IL-1β are due to activation of the gene HAS1 rather than the down-regulation of HA-degrading enzymes
Leflunomide does not block IL-1β-induced NF-κB translocation in FLS
NF-κB is a transcription factor that plays an essential role in the activation of many proinflammatory genes (36, 37). IL-1β is known to induce activation and translocation of NF-κB in several cell types. Furthermore, convincing evidence has been presented by others (38) indicating the potential of leflunomide as an inhibitor of NF-κB activation. We were interested in investigating whether inhibition of NF-κB might account for leflunomide-mediated inhibition on HAS1 mRNA.
First, the effect of IL-1β on NF-κB activation in FLS was tested. Cells were cultured in 10-cm tissue culture plates to high density and subsequently stimulated with IL-1β. Nuclear extract preparation and EMSA were performed, as described before (29). Shown in Fig. 4,A is one of two experiments performed to determine the optimal point in time for subsequent experiments. As demonstrated in Fig. 4,A, IL-1β (5 ng/ml)-treated FLS respond within a very short time period with the translocation of NF-κB into the nucleus. Although 15 min of IL-1β treatment is sufficient to result in readily detectable NF-κB binding to its consensus element in EMSA experiments, maximal levels of NF-κB can be observed between 30 and 45 min. IL-1β concentrations as low as 0.1 ng/ml resulted in readily detectable NF-κB translocation; maximal activation was obtained at 3–5 ng/ml; and increasing IL-1β concentrations further did not result in any further significant increase (data not shown). On the left side of this EMSA figure, the label κB indicates the position of the NF-κB/DNA complexes, and NS indicates the position of non-κB/DNA-binding protein complexes. Given on top of Fig. 4 B are the culture conditions (minutes of IL-1β exposure). The label Free Pr. indicates the column in which nuclear protein extract was left out of the EMSA reaction mixture.
Next, we tested whether leflunomide has the potential to affect IL-1β-induced NF-κB activation. FLS were treated with 5 and 50 μM leflunomide for 45 min; after that, IL-1β (5 ng/ml) was added. A representative EMSA experiment is shown in Fig. 4 B. As a comparison of the columns labeled Medium and IL-1, respectively, demonstrates, IL-1β treatment results in significantly higher levels of NF-κB/DNA complexes. More importantly, neither pretreating FLS with 5 μM nor with 50 μM leflunomide led to a significant reduction in subsequent IL-1β-induced NF-κB translocation.
Competition experiments as well as supershift experiments were performed, as described earlier (29), ensuring specificity of EMSA experiments. Shown in Fig. 4 C are such control experiments, demonstrating that while unlabeled NF-κB/DNA-binding elements can compete for κB binding, unlabeled AP-1 or CRE consensus elements cannot.
Reconstituting the pyrimidine synthesis pathway with uridine does not restore IL-1β-induced HAS1 activation
Leflunomide is a known inhibitor of pyrimidine synthesis (39), and adding exogenous uridine reportedly demonstrated the involvement of this mechanism in many leflunomide-mediated effects (40). We tested whether such mechanisms might account for the observed phenomena with regard to the activation of the HAS1 gene. In several independent experiments, the addition of exogenous uridine did not reverse the leflunomide-mediated effect on IL-1β-induced HAS1 up-regulation. In some experiments, uridine (100 μM) was added together with IL-1β; in other experiments, at the same time as leflunomide. Such experiments (data not shown) had no marked effect and demonstrated that the suppressive effect of leflunomide could not be overcome by exogenous uridine, making it unlikely that inhibition of pyrimidine synthesis by leflunomide accounts for the effect on HAS1.
IL-1-induced HAS1 activation relies on tyrosine kinase activation
Besides being an NF-κB inhibitor, leflunomide has been shown to interfere with tyrosine kinase activities. To explore the involvement of protein tyrosine kinases in IL-1β-induced HAS1 activation, genistein and the stable erbstatin analog 2,5-dihydroxymethylcinnamate, both of which have been shown to function as tyrosine kinase inhibitors, were used (41, 42). Shown in Fig. 5 are the results of an experiment in which FLS were treated with genistein and erbstatin, followed by stimulation with IL-1β. Depicted in Fig. 5 are mRNA levels for the genes HAS1 and actin, as indicated on the right side of this figure. As demonstrated in this figure, both tyrosine kinase inhibitors suppress IL-1β-induced HAS1 mRNA accumulation. Although no significant effect was obtained when genistein was added at concentrations of 5 μM, a higher concentration of 50 μM genistein repeatedly blocked the effects of IL-1β on HAS1 by >60%. Similarly, while concentrations of 5 μM erbstatin were not sufficient to significantly affect IL-1β-induced HAS1 mRNA levels, increasing the concentration to 50 μM resulted in >90% inhibition.
Arbitrary fluorescence units are given on the y-axis, and culture conditions are indicated on the x-axis. In the graphs, HAS1 mRNA levels were normalized using mRNA levels of actin.
We tested the effect of leflunomide on noninduced as well as IL-1α-, IL-1β-, and PMA-induced HA activation and release. The presented data demonstrate that in FLS, HAS1 is the gene that is readily activated by these stimuli, while the constitutively activated genes encoding HAS2 and HAS3 are not. EMSA experiments exclude leflunomide-induced inhibition of NF-κB translocation as the mode of action of this drug. Furthermore, inhibition of de novo pyrimidine synthesis by leflunomide as the mode of action could also be ruled out by reconstitution experiments. We conclude that the effect of leflunomide on HAS1 transcription is due to its properties as a tyrosine kinase inhibitor. Such a conclusion is supported by experiments demonstrating that two tyrosine kinase inhibitors mimic the effects of leflunomide on HAS1 regulation.
These and data published earlier (29) demonstrate that HAS1 is a gene that, in contrast to other HAS genes, is readily activated by a series of proinflammatory stimuli. Whether it is also the HAS1 gene product that acts as the ligand for CD44, therefore facilitating and contributing to undesired cell migration in affected joints, is currently under investigation.
This study has been motivated by our interest in the role that the different HAS genes might play in the pathogenesis of RA. Undoubtedly, HA is essential for many physiological processes; nevertheless, the presence of abnormally large amounts of HA in joints as well as in serum of RA patients is a hallmark of this disorder (15). That HA is much more than an inert matrix molecule has become increasingly clear. A series of studies demonstrate, for example, an exceptional role of HA in the progression of certain forms of cancer (43, 44, 45, 46, 47). More important in the context of this manuscript are reports that clearly demonstrate the importance of HA in cell adhesion and migration, events that are associated with inflammatory processes also seen in RA. One of the prerequisites for molecules involved in inflammation and migration is the ability to be regulated in a tightly controlled manner. In FLS, only HAS1 seems to fulfill such requirements. Interestingly, plenty of effort was put into understanding the activation and regulation of CD44, the principal binding partner of HA that has been shown to be essential for the migration of many cell types (48, 49, 50). It is, however, surprising how little is known about the regulation/activation of HA, the counterpart to CD44.
Leflunomide has recently been approved for the treatment of RA and has been used for several years with considerable success (51). It is thought that the molecular mechanisms of leflunomide in RA include interference with IL-2 release and IL-2R binding, modulation of Th2-dependent B cell function. Inhibition of various tyrosine kinases, demonstrated also in animal models, has been reported to account for effects of leflunomide (52). Other beneficial effects of leflunomide noted by investigators seem to be due to its inhibition of cell adhesion and cell migration (53). However, the modus operandi favored by most investigators is the well-documented inhibition of de novo pyrimidine synthesis by leflunomide. Data presented in this manuscript seem to exclude such a mechanism as the mode of action, because replenishing cell culture medium with uridine did not restore IL-1-induced HAS1 activation.
Of great importance for the understanding of leflunomide as an anti-inflammatory remedy are also reports demonstrating its inhibitory effects on transcription factors that have been shown to be essential for the activation of most proinflammatory genes (36). NF-κB is a transcription factor that has been studied in great detail. The demonstration that leflunomide can act as an inhibitor of NF-κB activation seemed to explain and account for the beneficial effects of leflunomide treatment (22, 38). Interestingly, at least in T cells, inhibition of NF-κB activation by leflunomide is also a process that can be overcome by addition of exogenous uridine (38). Our EMSA data demonstrate clearly that inhibition of IL-1β-induced NF-κB translocation is not the mode of action of leflunomide in FLS.
IL-1, with its myriad effects on cell signaling, has been shown to activate protein tyrosine kinases (54, 55). Protein tyrosine kinases are thought to play essential roles in signal transduction in many cell types and have been demonstrated to play important roles in cell activation associated with inflammation. Leflunomide has also been described to possess tyrosine kinase inhibitor properties (56, 57).
The concentration of leflunomide necessary for complete HAS1 inhibition in in vitro experiments is readily attainable and maintained in humans without significant toxicity (58). Furthermore, the serum concentration of this drug that prevents cardiac allograft rejection has been shown to range from 10 to 100 μM (59). As reported for activation of T cells, leflunomide was able to inhibit Src family (p56lck and p59fyn)-mediated protein kinase at IC50 of 125–175 and 22–40 μM, respectively, while the IC50 values for autophosphorylation and phosphorylation of histone 2B were 160 and 65 μM, respectively (56). The same group also demonstrated the ability of leflunomide to inhibit protein tyrosine phosphorylation induced by anti-CD3 Abs. In such cases, the IC50 values of total intracellular tyrosine phosphorylation ranged from 5 to 45 μM. These concentrations are well within the range used in our experiments and considerably lower than concentrations necessary to inhibit NF-κB activation (38).
The conclusion that the observed effects of leflunomide are indeed due to inhibition of protein tyrosine kinase activities is supported by: 1) our EMSA experiments that exclude the effects of NF-κB as the modus operandi; 2) reports that demonstrate the need for considerably higher concentrations of leflunomide in experiments in which this drug has been shown to inhibit the activation of NF-κB; and 3) the failure to restore IL-1-induced HAS1 activation by exogenous uridine as a measure to reconstitute a possibly blocked pyrimidine synthase pathway. The assumption that IL-1β-induced HAS1 up-regulation is indeed due to the well-described effect of leflunomide on the activation of tyrosine kinase is supported by our data demonstrating that two well-defined protein tyrosine kinase inhibitors were similarly effective in suppressing IL-1β-induced HAS1 mRNA accumulation.
Collectively, these data suggest that tyrosine kinase activity is essential for IL-1β-induced HAS1 activation. These data also point at the mechanism by which leflunomide lowers HA production, a mechanism that might account for the beneficial effects of leflunomide in treating RA. Furthermore, the presented data on the mechanism of leflunomide-mediated inhibition of HA up-regulation may be helpful as a tool to further dissect the role of tyrosine kinases in pathways leading to HAS1 transcription, and might also contribute to a better understanding of the mode of action of this class of immunomodulatory drug.
I thank C. Pollaschek for performing RT-PCR experiments.
The author has no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported in part by grants from the City of Vienna; the Austrian Ministry of Social Security and Generations; the Austrian Ministry of Education, Science, and Culture; and the Austrian National Bank.
Abbreviations used in this paper: RA, rheumatoid arthritis; CRE, cAMP-responsive element-like; FLS, fibroblast-like synoviocyte; HA, hyaluronic acid/hyaluronan; HAS, hyaluronan synthase; LB, lysis buffer.