High molecular weight hyaluronic acid down-regulates the gene
expression of osteoarthritis-associated cytokines and enzymes in
fibroblast-like synoviocytes from patients with early osteoarthritis
C.-T. Wang M.D.
y
, Y.-T. Lin M.D., Ph.D.
z
, B.-L. Chiang M.D., Ph.D.
z
,
Y.-H. Lin M.Sc.
z
and S.-M. Hou M.D., Ph.D., M.P.H.
y*
y Department of Orthopedics, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
z Department of Pediatrics, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
Summary
Objective: Activated synoviocytes play important roles in the progression of human osteoarthritis (OA). Intra-articular injection of high molecular weight hyaluronic acid (HMW-HA) has been used as viscosupplementation for knee OA but its effect on synoviocytes remains un-disclosed. This study aims to investigate the effects of HMW-HA on the gene expression of 16 OA-associated cytokines and enzymes, includ-ing interleukin (IL)-1b, IL-6, IL-8, leukemia inhibitory factor (LIF), tumor necrosis factor (TNF)-a, TNF-a convertinclud-ing enzyme (TACE), matrix metalloproteinase (MMP)-1, MMP-2, MMP-3, MMP-9, MMP-13, tissue inhibitor of metalloproteinase (TIMP)-1, TIMP-2, aggrecanase-1, aggre-canase-2, and inducible nitric oxide synthase (iNOS), in fibroblast-like synoviocytes (FLS) from patients with early stage OA.
Method: Synovial fluid-derived FLS were obtained from the knees of 15 patients with early stage OA. IL-1-stimulated or unstimulated FLS were cultured with or without the treatment of 600e800 kDa HMW-HA. Moreover, blocking experiments with anti-CD44 monoclonal antibodies (mAb) were used to examine the involvement of CD44 in HMW-HA effects. We designed and validated the real-time quantitative polymerase chain reaction (Q-PCR) assays with SYBR Green dyes for simultaneous quantification of the expression of the 16 genes.
Results: HMW-HA down-regulated IL-8 and iNOS gene expression in unstimulated FLS and down-regulated aggrecanase-2 and TNF-a gene expression in IL-1-stimulated FLS. CD44 blocking inhibited the down-regulatory effects of HMW-HA on gene expression.
Conclusion: HMW-HA may have a structure-modifying effect for OA by down-regulation of aggrecanase-2 in FLS. HMW-HA also has an anti-inflammatory effect by down-regulation of TNF-a, IL-8, and iNOS in FLS. These effects may be mediated through the interaction of CD44 and HMW-HA.
ª 2006 OsteoArthritis Research Society International. Published by Elsevier Ltd. All rights reserved. Key words: Hyaluronic acid, Osteoarthritis, Fibroblast-like synoviocytes, Aggrecanase, Cytokine, Enzyme.
Introduction
Recent studies suggest that osteoarthritis (OA) is a ‘‘non-classical’’ inflammatory disease1,2 and that activated synoviocytes play important roles in the progression of human OA3,4. Synoviocytes can produce various
OA-asso-ciated cytokines and enzymes, such as interleukin (IL)-1b, IL-6, IL-8, leukemia inhibitory factor (LIF), tumor necrosis factor (TNF)-a, TNF-a converting enzyme (TACE), matrix metalloproteinase (MMP)-1, MMP-2, MMP-3, MMP-9, MMP-13, tissue inhibitor of metalloproteinase (TIMP)-1, TIMP-2, aggrecanase-1, aggrecanase-2, and inducible nitric oxide synthase (iNOS)5e11. In addition, a recent study showed that fibroblast-like synoviocytes (FLS) can be obtained from synovial fluid, and the phenotype and func-tion of synovial fluid-derived FLS are the same as those
derived from the synovium of surgical specimens12. This cell model makes it possible to study the FLS from the pa-tients with early stage OA, whose tissue specimens are sel-dom available.
Hyaluronic acid (HA), a core component of extracellular matrix, comprises a repeat of N-acetylglucosamine and
D-glucuronic acid. HA has been identified as an important
modulator in many physiological and pathological pro-cesses13,14. Most of these responses are mediated through HAeCD44 interaction. Moreover, HA is a critical constituent component of normal synovial fluid and an important contributor to joint homeostasis15. In OA, both the concen-tration and molecular weight of intra-articular endogenous HA are decreased16. In clinical practice, intra-articular injec-tion of high molecular weight HA (HMW-HA) has been used as viscosupplementation for knee OA and its therapeutic ef-ficacy has been verified17. However, the exact mechanism
of HMW-HA in treating OA remains partially disclosed. Only a few studies examine the effects of HMW-HA on synovio-cytes18e20. Sasakiet al. found that HMW-HA inhibited the IL-1-induced expression of MMP-1 and MMP-3 in human synovial cells19. Takahashi et al. found that intra-articular injection of HMW-HA suppressed the mRNA expression
*Address correspondence and reprint requests to: Professor Sheng-Mou Hou, M.D., Ph.D., M.P.H., Department of Orthopedics, National Taiwan University Hospital and National Taiwan University College of Medicine, No. 7, Chung-Shan South Road, Taipei 100, Taiwan. Tel: 23123456x5274; Fax: 886-2-23971266; E-mail:[email protected]
Received 2 February 2006; revision accepted 10 May 2006.
ª 2006 OsteoArthritis Research Society International. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.joca.2006.05.009 International Cartilage Repair Society 1237
of MMP-3 and IL-1b in the synovium of rabbit OA model20. Qiu et al. found that intra-articular injection of HMW-HA decreased the mRNA expression of MMP-3 in the synovium of rabbit OA model18. Therefore, more studies are needed
to fully elucidate the effects of HMW-HA on synoviocytes. Reverse transcription (RT) followed by polymerase chain reaction (PCR) is the technique of choice for analyzing gene expression due to its high sensitivity. Among the various methods of quantification for RT-PCR, real-time quantitative PCR (Q-PCR) has advantages of extremely wide dynamic detection range and of higher reliability of results compared with end-point determinations using conventional PCR21. Real-time Q-PCR can quantify DNA fragment amplification using Taqman probes or SYBR Green fluorescence. SYBR Green is less expensive than Taqman probes and can pro-vide an equally accurate result in real-time Q-PCR if the PCR specificity is high and the artifacts such as primer dimers are minimal22, thus becoming an valuable and economical tool in researching gene expression.
The aims of this study were to investigate the effects of HMW-HA on the gene expression of 16 OA-associated cytokines and enzymes, including IL-1b, IL-6, IL-8, LIF, TNF-a, TACE, MMP-1, MMP-2, MMP-3, MMP-9, MMP-13, TIMP-1, TIMP-2, aggrecanase-1, aggrecanase-2 and iNOS, in synovial fluid-derived FLS from the patients with early stage OA. Furthermore, we also examine the involve-ment of CD44, a major HA-binding receptor, in these effects. These may clarify the mechanism of HMW-HA in treating OA. In this study, we designed and validated the real-time Q-PCR assays with SYBR Green dye for simulta-neous quantification of the expression of these 16 genes.
Methods
CULTURES OF SYNOVIAL FLUID-DERIVED FLS
The study protocol was approved by the Institutional Review Board of the National Taiwan University Hospital. Synovial fluid was aspirated from the knees of 15 patients with early stage OA (KellgreneLawrence grades I and II). The synovial fluid was centrifuged at 450g for 30 min, cell pellets were resuspended in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Grand Island, NY, USA) contain-ing 10% fetal calf serum (FCS), and incubated for 24 h at 37(C in a plastic culture flask. Non-adherent cells were washed out and medium was changed daily for the next 3 days. The remaining adherent cells were cultured for addi-tional 2 weeks in a flask before trypsinization, and then passed to new culture flasks. Cell passages 4 through 6 were FLS, which were used for phenotypic analysis and for experiments.
PHENOTYPIC ANALYSIS OF SYNOVIAL FLUID-DERIVED FLS
For surface marker staining, the anti-fibroblast surface molecule monoclonal antibodies (mAb) (clone D7-FIB; Abcam, Cambridge, UK) were used. FLS (passages 4 through 6) were released from culture by trypsinization, washed once, and resuspended in DMEM containing 1% FCS. Cells were incubated with the primary antibodies D7-FIB at 4(C for 30 min. After three washes, the cells were incubated with fluorescein isothiocyanate (FITC)-con-jugated goat anti-mouse IgG mAb (Serotec, Oxford, UK) as secondary antibodies at 4(C for 30 min and then washed
Fig. 1. Phenotypic features of synovial fluid-derived FLS (passages 4 through 6). (A) Light microscopy showed that FLS were homogeneous and spindle-shaped in morphology. (B) Flow cytometric analysis showed that more than 90% of FLS were positively stained with anti-fibroblast surface molecule mAb D7-FIB. (C) Immunocytochemistry revealed that the majority of FLS were positively stained with anti-prolyl-4-hydrox-ylase mAb 5B5. (D) Immunocytochemistry of negative control staining using isotype-matched control IgG did not display a positive reaction.
three times. Subsequently, the cells were fixed in a 1.25% paraformaldehyde solution and then analyzed using a FACS Calibur flow cytometer (BD Biosciences, San Jose, CA, USA). The positive and negative populations were determined according to the staining of unreactive isotype-matched control IgG.
For immunocytochemistry staining, the anti-fibroblast prolyl-4-hydroxylase mAb (clone 5B5; Abcam, Cambridge, UK) were used. FLS (passages 4 through 6) in 6-well culture plates were fixed with acetone for 15 min, washed twice with phosphate-buffered saline (PBS), then incubated with 1% bovine serum albumin (BSA) for 30 min and with the primary antibodies 5B5 for 1 h in a humid chamber at 4(C. After three washes with PBS, the cells were stained with VECTASTAIN Universal Quick Kit (Vector Laboratories Inc., Burlingame, CA, USA) according to the manufacturer’s protocol. Unreactive isotype-matched control IgG was used as primary antibodies for negative controls.
CELL STIMULATION AND TREATMENT
For all of the experiments, FLS were plated in 6-well cul-ture plates and serum starved for 24 h in DMEM containing 1% FCS to synchronize cells in a activating and non-proliferating phase. FLS were then cultured in DMEM containing 10% FCS and either (1) maintained as unstimu-lated and untreated controls, (2) treated with 100 mg/ml HMW-HA of MW 600e800 kDa (Artzdispo, Seikagaku Inc., Japan) for 24 h, (3) stimulated with 1 ng/ml IL-1 (Pepro-Tech, Rocky Hill, NJ, USA) for 24 h, or (4) stimulated with 1 ng/ml IL-1 and treated with 100 mg/ml HMW-HA for 24 h. In the CD44 blocking experiments, FLS were cultured in DMEM containing 10% FCS and either (1) stimulated with 1 ng/ml IL-1 for 24 h, (2) stimulated with 1 ng/ml IL-1 and treated with 100 mg/ml HMW-HA for 24 h, or (3) stimulated with 1 ng/ml IL-1 and treated with 100 mg/ml HMW-HA for
24 h after pretreatment with 20 mg/ml anti-CD44 blocking mAb (clone 5F12; Lab Vision Corp., Fremont, CA, USA) or 20 mg/ml isotype-matched control IgG (eBioscience, San Diego, CA, USA) for 1 h. The optimal concentration of HMW-HA (100 mg/ml) used in this study was determined according to the results of a preliminary doseeresponse study using 10 mg/ml, 100 mg/ml, or 1 mg/ml HMW-HA to treat the FLS from five patients (Fig. 5).
TOTAL RNA ISOLATION, QUANTIFICATION AND RT
FLS after treatments were lysed and total RNA was extracted with Trizol agent (Life Technologies, Rockville, MD, USA) according to the manufacturer’s protocol. Total RNA was quantified by spectrophotometry. DNase diges-tion was carried out using DNA-free (Ambion Inc., Austin, TX, USA) as per the manufacturer’s protocol. One micro-gram of total RNA was converted to cDNA using the Super-script II reverse tranSuper-scriptase (Invitrogen, Carlsbad, CA, USA). For each reaction, 4 mL 5 first-strand buffer (50 mM TriseHCl, pH 8.3, 375 mM KCl, 15 mM MgCl2),
2 mL of 0.1 M dTT, 5 U RNAsin, 500 mM dNTP mix, 200 pmol Oligo-dT, 25 U Superscript II reverse transcrip-tase and sterile water were added to the RNA to a volume of 20 mL. This reaction was then incubated at 42(C for 1 h. The finished cDNA products were stored in aliquots at80(C until needed.
PRIMER DESIGN AND REAL-TIME Q-PCR ASSAYS
The mRNA sequences were retrieved from the NCBI website. The NCBI BLAST server23was used to determine primer specificity, where reported similarities above 75% to non-specific sites within the human genome were undesir-able. Primers were designed to span the exoneexon bound-ary to eliminate the possible influence of the contamination
Table I
Primer sequences and product sizes in quantitative PCR Target gene Forward primer (50e30) Reverse primer (50e30) Genomic
DNA product size (bp) cDNA product size (bp) GenBank accession number 18S rRNA* AAGGAGACTCTGGCATGCTAAC CAGACATCTAAGGGCATCACAGAC 133 133 NG_002801
HPRT1* CCCTGGCGTCGTGATTAGT GTGATGGCCTCCCATCTCCTT 0 169 NM_000194 MMP-1 CATGCCATTGAGAAAGCCTTCC AGAGTTGTCCCGATGATCTCC 0 123 NM_002421 MMP-2 TGCTGGAGACAAATTCTGGA GATAC ACTTCACGCTCTTCAGACTTTGG 0 200 NM_004530 MMP-3 GACAAAGGATACAACAGGG ACCAAT TGAGTGAGTGATAGAGTGGGTACAT 0 122 NM_002422 MMP-9 TGCCCGGACCAAGGATACAG TCAGGGCGAGGACCATAGAG 0 182 NM_004994 MMP-13 GTCTCTCTATGGTCCAGGA GATGAA AGGCGCCAGAAGAATCTGT 0 144 NM_002427
TIMP-1 CCCAGAGAGACACCAGAGAAC CACGAACTTGGCCCTGATGAC 0 167 NM_003254
TIMP-2 GCACATCACCCTCTGTGACTT AGCGCGTGATCTTGCACT 0 110 NM_003255
Aggrecanase-1 ATGGCTATGGGCACTGTCTC GTGTTTGGTCTGGCACATGG 0 206 NM_005099
Aggrecanase-2 CCTGCCCACCCAATGGTAAATC CGGCCTACATTCAGTGCCATC 0 218 NM_007038
IL-1b GGCTTATTACAGTGGCAATGAGGA TCCATGGCCACAACAACTGA 0 180 NM_000576
IL-6 GACAGCCACTCACCTCTTCA TTCACCAGGCAAGTCTCCTC 0 211 NM_000600
IL-8 CTGCGCCAACACAGAAATT ATTGTA
TTCACTGGCATCTTCACTGATTCTT 0 170 NM_000584
LIF ATACGCCACCCATGTCACAA CCCTGGGCTGTGTAATAGAGAA 0 107 NM_002309
TNF-a CCCAGGCAGTCAGATCATCTT TCTCAGCTCCACGCCATT 0 140 NM_000594
TACE GTGACATGAATGGCAAAT GTGAGA
ACAATGGACAAGAATGCTGAAAGG 0 176 NM_003183
iNOS TCACCAGGAGATGCTGAACTAC GAGCACAGCTTTGACCAAGAC 0 142 NM_000625
bp¼ base pairs; rRNA ¼ ribosomal RNA; HPRT1 ¼ hypoxanthine phosphoribosyltransferase 1; MMP ¼ matrix metalloproteinase; TIMP¼ tissue inhibitor of metalloproteinase; IL ¼ interleukin; LIF ¼ leukemia inhibitory factor; TNF ¼ tumor necrosis factor; TACE ¼ TNF-a converting enzyme; and iNOS¼ inducible nitric oxide synthase.
of genomic DNA. Lengths of PCR products were designed to range between 100 and 220 bp. Primers were designed to have a length of 18e25 bp, 40e60% GC content, and a theoretical annealing temperature of approximately 59(C. Compatible 30end sequences that may cause primer dimer formation, of either the primer itself or its primer pair, were avoided by visual inspection of the candidate primer sequences. Negative water controls were also prepared for each primer set.
Platinum SYBR Green qPCR SuperMix (Invitrogen, Carlsbad, CA, USA) was used in all reactions. The ABI PRISM 7700 Sequence Detection System (Applied Bio-systems, Foster City, CA, USA) was used for all real-time Q-PCRs. The PCR thermal protocol applied consisted of a 2 min 95(C denaturation step, followed by 45 repeats of a 15 s 95(C denaturation step, a 30 s 59(C annealing step and a 30 s extension step at 72(C. A melting curve analysis was performed after final amplification period via a temperature gradient from 60(C to 95(C. All real-time Q-PCR amplified products were separated on agarose gel using the appropriate DNA ladder to confirm appropriate fragment sizes and lacking of primer dimers.
STANDARD CURVES
Real-time Q-PCRs for the primer sets of 18S ribosomal RNA (rRNA), hypoxanthine phosphoribosyltransferase 1
(HPRT1), IL-1b, IL-6, IL-8, LIF, TNF-a, TACE, MMP-1, MMP-2, MMP-3, MMP-9, MMP-13, TIMP-1, TIMP-2, aggre-canase-1, aggrecanase-2 and iNOS were tested for linearity of response by constructing standard curves. Based on the methods modified from the standard curve protocol of Whe-lanet al.24
, 1/10 dilution series of the purified PCR products was prepared in nuclease-free water and SYBR Green Q-PCR was performed for the primer sets on 102e109copies of their corresponding cDNA PCR products as templates. The target PCR products were purified by agarose gel elec-trophoresis. The absolute concentration of the purified PCR products was measured by using PicoGreen dsDNA quanti-tation reagent (Invitrogen, Carlsbad, CA, USA). The copy numbers of the purified PCR products were calculated from the DNA concentration by taking account of the size of each specific PCR product, the mean molecular weight of the nucleotide bases (340.5 g/mol), Avogadro’s number (6.022 1023), and the volume25.
STATISTICAL ANALYSIS AND RELATIVE QUANTIFICATION OF GENE EXPRESSION
Relative expression software tool (REST)26was used to
calculate comparative gene expression levels between samples after normalization to the control housekeeping gene (18S rRNA and HPRT1) expression. Using the ABI PRISM 7700 System software, the threshold cycle, at
Fig. 2. Melting curve analysis and agarose gel electrophoresis of real-time Q-PCR assays. (A) Melting curve analysis showed a single melting peak in each real-time Q-PCR assay. (B) Agarose gel electrophoresis also showed a single PCR product in each real-time Q-PCR assay.
which the threshold cycle value (CT) was measured, was
adjusted to the exponential phase of the amplification traces. The average of comparative gene expression ratio was determined from triplicate real-time Q-PCRs to take variation into account. A standard curve graph of CT vs
log value of starting copy number of the sample from the di-lution series was constructed. The slope of the standard curve was used to determine the amplification efficiency: Efficiency¼ 10(1/slope) 1. If calculated amplification effi-ciency was more than 1, then 100% amplification effieffi-ciency was assumed. Correlation coefficients were derived from the standard linear regression. Corrected comparative gene expression ratio adjusted by amplification efficiency
was calculated using REST. Pair-wise fixed reallocation randomization test provided in REST was used to distin-guish statistically significant results. Significance was set atP-value <0.05.
Results
PHENOTYPIC FEATURES OF SYNOVIAL FLUID-DERIVED FLS
FLS (passages 4 through 6) exhibited uniform spindle-shaped and fibroblast-like morphology [Fig. 1(A)]. The fibro-blast phenotype of FLS (passages 4 through 6) was confirmed by analysis of surface antigen and intracellular
Fig. 3. Standard curves and amplification efficiencies of real-time Q-PCR assays. The standard curves of all real-time Q-PCR assays showed high linearity over a wide range of starting copy numbers of cDNA templates. The equation represents the relationship of the threshold cycle value (CT) and the log value of starting copy number of the sample from the dilution series.R2¼ correlation coefficient; Eff ¼ calculated
protein. Flow cytometric analysis showed that more than 90% of FLS expressed the fibroblast surface marker [Fig. 1(B)]. Immunocytochemistry staining revealed the presence of fibroblast product prolyl-4-hydroxylase in the majority of FLS [Fig. 1(C)], whereas negative control stain-ing did not display a positive reaction [Fig. 1(D)].
PRIMER DESIGN AND CONFIRMATION OF PRIMER SPECIFICITY
Designed primer sequences and expected RT-PCR prod-uct sizes are listed in Table I. The specificity of RT-PCR products was confirmed with the high resolution agarose gel electrophoresis and a single band of the desired length was shown [Fig. 2(B)]. In addition, the melting curve analy-sis was performed. The Q-PCR assay for each gene showed a single product with a specific melting temperature as follows: 18S rRNA, 85.3(C; HPRT1, 79.7(C; MMP-1, 80.9(C; MMP-2, 86.3(C; MMP-3, 81.3(C; MMP-9, 88.2(C; MMP-13, 80.4(C; aggrecanase-1, 89.7(C; aggre-canse-2, 84.9(C; TIMP-1, 88.1(C; TIMP-2, 86.3(C; IL-1b, 86.8(C; IL-6, 82.3(C; IL-8, 80.5(C; LIF, 83.6(C; TNF-a, 87.3(C; TACE, 78.8(C; and iNOS, 82.2(C [Fig. 2(A)]. No primer dimers were generated during the applied 45 ampli-fication cycles of real-time Q-PCR.
Q-PCR AMPLIFICATION EFFICIENCIES AND LINEARITY
All real-time Q-PCR assays functioned optimally using PlatinumSYBRGreen qPCR SuperMix without the addi-tion of MgCl2to the reagent. The standard curves of Q-PCR
assays for all the 18 genes had correlation coefficients of 0.98 or higher and the amplification efficiencies were mostly beyond 90% in the investigated range from 102to 109 cop-ies of cDNA templates (Fig. 3).
DOSE-RESPONSE STUDY
In both the FLS with and without IL-1 stimulation, adding 100 mg/ml HMW-HA obviously decreased the gene expres-sion levels of most of the 16 OA-associated cytokines and enzymes. In comparison with 100 mg/ml HMW-HA, 1 mg/ml
HMW-HA had no additional down-regulatory effect on gene expression. On the other hand, 10 mg/ml HMW-HA had no obvious down-regulatory effect on gene expression (Fig. 5). Therefore, we used 100 mg/ml HMW-HA to conduct the following experiments.
HMW-HA down-regulated IL-8 and iNOS gene expression in non-stimulated FLS
In FLS without IL-1 stimulation, adding HMW-HA de-creased the expression levels of all the 16 genes. Among these, adding HMW-HA significantly decreased the IL-8 gene expression level to 3.1% (P < 0.05) and significantly decreased the iNOS gene expression level to 7.5% (P < 0.05) [Fig. 6(A)].
HMW-HA down-regulated aggrecanase-2 and TNF-a gene expression in IL-1-stimulated FLS
After IL-1 stimulation, the expression levels of 11 genes significantly increased to many folds (P < 0.05) (Fig. 4). In FLS with IL-1 stimulation, adding HMW-HA decreased the gene expression levels in most of the 16 genes. Among these, adding HMW-HA significantly decreased the aggre-canase-2 gene expression level to 33% (P < 0.05) [Fig. 6(B)]. Moreover, in the absence of HMW-HA, the TNF-a gene expression level significantly increased after IL-1 stimulation (P < 0.05) (Fig. 4), however, in the pres-ence of HMW-HA, it did not significantly increase after IL-1 stimulation. In other words, adding HMW-HA made the increase of TNF-a gene expression induced by IL-1 stimulation become insignificant.
CD44 blocking inhibited the down-regulatory effects of HMW-HA on gene expression
In FLS with IL-1 stimulation, pretreatment with anti-CD44 blocking mAb 5F12 inhibited the down-regulation of aggre-canase-2, TNF-a, IL-8, and iNOS gene expression caused by HMW-HA. However, pretreatment with isotype-matched control IgG had no effect on the down-regulation of gene ex-pression caused by HMW-HA (Fig. 7).
Fig. 4. Effect of IL-1 (1 ng/ml) stimulation on gene expression in FLS. Values are expressed in fold in comparison with control group after normalization to housekeeping gene. Data are expressed as meanS.E.M. *P < 0.05.
Discussion
This study showed that HMW-HA can down-regulate the gene expression of aggrecanase-2, TNF-a, IL-8, and iNOS in synovial fluid-derived FLS from the patients with early stage OA. HMW-HA down-regulated the gene expression of IL-8 and iNOS in unstimulated FLS and down-regulated the gene expression of aggrecanase-2 and TNF-a in
IL-1-stimulated FLS. These findings, to our knowledge, have never been reported by other researchers.
We found that HMW-HA down-regulated the gene ex-pression of aggrecanase-2 in IL-1-stimulated FLS. Aggreca-nase-2 has been identified to be the primary aggrecanase responsible for aggrecan degradation in a murine model of OA27. Thus, down-regulation of aggrecanase-2 suggests that HMW-HA may have a structure-modifying effect. We
Fig. 5. Dose-response study of the effect of HMW-HA (10 mg/ml, 100 mg/ml, and 1 mg/ml) on gene expression in FLS. (A) Effect of HMW-HA on gene expression in FLS without IL-1 stimulation. Values are expressed in fold in comparison with control group after normalization to housekeeping gene. (B) Effect of HMW-HA on gene expression in FLS with IL-1 (1 ng/ml) stimulation. Values are expressed in fold in
also found that HMW-HA down-regulated the gene expres-sion of iNOS, TNF-a, and IL-8 in FLS. In the literature, iNOS is detected in OA synovium, suggesting that increased local production of NO may contribute to the pathogenesis of in-flammatory arthritis by increasing synovial blood flow and by modulating cellular function within the synovium28. NO
also inhibits matrix synthesis in cartilage29. TNF-a is a pro-inflammatory cytokine that plays key roles in altering cartilage extracellular matrix turnover30. IL-8 is a cytokine that may play a regulatory function in the inflammatory process in OA31. Down-regulation of these inflammatory medi-ators suggests that HMW-HA may have an anti-inflammatory
Fig. 6. Effect of HMW-HA (100 mg/ml) on gene expression in FLS. (A) Effect of HMW-HA on gene expression in FLS without IL-1 stimulation. Values are expressed in fold in comparison with control group after normalization to housekeeping gene. (B) Effect of HMW-HA on gene expression in FLS with IL-1 (1 ng/ml) stimulation. Values are expressed in fold in comparison with IL-1 group after normalization to housekeep-ing gene. Data are expressed as meanS.E.M. *P < 0.05. **There was significant difference between IL-1 group and control group (Fig. 4), but
there was no significant difference between ‘‘IL-1þ HMW-HA’’ group and control group.
effect. In addition, we observed that HMW-HA also de-creased the gene expression of MMPs and other OA-asso-ciated cytokines in FLS, although not in a significant way. Therefore, based on these results, it may be reasonable to propose the use of HMW-HA in both inflammatory and non-inflammatory status for OA patients, because in the inflammatory status, HMW-HA could down-regulate aggre-canase-2 and TNF-a, while in the non-inflammatory status, HMW-HA could down-regulate IL-8 and iNOS in FLS and thus prevent inflammation and destruction of cartilage.
In the present study, we observed that the pretreatment of FLS with anti-pan CD44 blocking mAb 5F12 could reverse the down-regulatory effects of HMW-HA on the gene expres-sion of aggrecanase-2, TNF-a, IL-8, and iNOS. This finding suggests that the suppressive effects of HMW-HA on OA-associated cytokines and enzymes in FLS are mediated through the interaction of CD44 and HMW-HA, which is con-sistent with the finding of Shimizuet al. in rheumatoid syno-vial cells32. Similar suppressive effect of HMW-HA is also found in chondrocytes. Julovi et al. found that HMW-HA effectively inhibited IL-1b-stimulated production of MMP-1, MMP-3, and MMP-13 in human articular cartilage explant, which may involve direct interaction between HA and CD44 on chondrocytes33. Tanaka et al. also showed that HMW-HA suppressed MMP-1 and regulated on activation, normal T expressed and secreted (RANTES) production in chondrocytes via CD44eHA interaction34. However, how
HMW-HA interacts with CD44 remains undisclosed. Further studies are needed to elucidate the CD44eHA signaling pathway in OA synoviocytes.
Our results demonstrated that in FLS with IL-1 stimulation, CD44 blocking only partially inhibited the down-regulatory effects of HMW-HA on gene expression. This finding implies that in addition to interacting with CD44, there may be an-other mechanism for the observed inhibitory effects of HA in this system. It is possible that the HA ‘‘traps’’ the IL-1 in solution and so prevents IL-1 binding to its receptor.
We obtained synovial fluid-derived FLS as the source of RNA for RT-PCR in this study. Although there are many re-searches studying FLS in OA patients, most of the FLS were derived from the traditional synovium taken from the surgical specimens of total joint arthroplasty. This means that these FLS are derived from the synovium of the pa-tients with advanced stage OA because only such papa-tients require total joint arthroplasty. Therefore, these cell models only represent the FLS of advanced OA, so they may be inappropriate to study the pathophysiologic mechanisms of FLS in the progression of OA. It may be more logical to use the synovial tissue obtained from non-arthroplastic joint surgery of patients with early stage OA. This kind of sur-gery, however, is few. In a recent research, the FLS ob-tained from synovial fluid are shown to be phenotypically and functionally the same as those derived from the syno-vium of surgical specimens12. This makes it possible to study the FLS from the patients with early stage arthritis, whose tissue specimens are seldom available, and also al-lows longitudinal studies to determine whether or not the function of FLS is altered by the disease course and by ther-apy. We used this new cell model in studying the effects of HMW-HA on synoviocytes and thus can infer the situation in the patients with early stage OA, who are appropriate can-didates to receive HMW-HA treatment, from the results of this study17.
The current knowledge of the biology of FLS shows that the intriguing functional property of FLS may stem from their resemblance to bone marrow stromal cells: both cell types share common progenitors and display similar gene expres-sion profiles. It is therefore likely that, similar to the bone marrow stromal cells, FLS may support or modulate the effector character of resident or blood-derived cells in the arthritic joint35e37. Moreover, responses to cytokines in culture and in situ hybridization studies in diseased tissue suggest that FLS secrete MMPs more readily than other fibroblasts38,39. Since the characteristics of FLS are somewhat different from those of other fibroblasts, it is un-certain whether the data presented in our study could be obtained with primary cultures of fibroblasts from other sources. It will be of interest to investigate the effects of HMW-HA on other fibroblasts, such as skin fibroblasts, to explore the therapeutic potential of HMW-HA in other diseases.
In summary, we applied the real-time Q-PCR assays with SYBR Green dye, which were specially designed for simul-taneous quantification of the gene expression of the 16 OA-associated cytokines and enzymes with high sensitivity and specificity, in synovial fluid-derived FLS from the patients with early stage OA. Our results suggest that HMW-HA may have a structure-modifying effect for OA by down-reg-ulation of the gene expression of aggrecanase-2 in FLS. HMW-HA also has an anti-inflammatory effect by down-reg-ulation of the gene expression of TNF-a, IL-8, and iNOS in FLS. These effects may be mediated through the interaction of CD44 and HMW-HA. Further studies are needed to elu-cidate the molecular mechanism of the suppressive effects of HMW-HA on these cytokines and enzymes in OA synoviocytes.
Acknowledgments
This work was supported by Research Grants from the National Science Council and the National Taiwan University Hospital (NTUH 91-N009).
Fig. 7. Effect of CD44 blocking on gene expression in FLS. FLS were stimulated with IL-1 (1 ng/ml) and treated with HMW-HA (100 mg/ml) with or without pretreatment with anti-CD44 blocking mAb 5F12 (20 mg/ml) or isotype-matched control IgG (20 mg/ml). Values are expressed in fold in comparison with IL-1 group after
References
1. Attur MG, Dave M, Akamatsu M, Katoh M, Amin AR. Osteoarthritis or osteoarthrosis: the definition of inflammation becomes a semantic issue in the geno-mic era of molecular medicine. Osteoarthritis Cartilage 2002;10(1):1e4.
2. Pelletier JP, Martel-Pelletier J, Abramson SB. Osteoar-thritis, an inflammatory disease: potential implication for the selection of new therapeutic targets. Arthritis Rheum 2001;44(6):1237e47.
3. Malemud CJ, Islam N, Haqqi TM. Pathophysiological mechanisms in osteoarthritis lead to novel therapeutic strategies. Cells Tissues Organs 2003;174(1e2): 34e48.
4. Pelletier JP, McCollum R, Cloutier JM, Martel-Pelletier J. Synthesis of metalloproteases and interleu-kin 6 (IL-6) in human osteoarthritic synovial membrane is an IL-1 mediated process. J Rheumatol Suppl 1995; 43:109e14.
5. Borderie D, Hilliquin P, Hernvann A, Lemarechal H, Kahan A, Menkes CJ, et al. Inhibition of inducible NO synthase by TH2 cytokines and TGF beta in rheu-matoid arthritic synoviocytes: effects on nitrosothiol production. Nitric Oxide 2002;6(3):271e82.
6. Distler JH, Jungel A, Huber LC, Seemayer CA, Reich CF 3rd, Gay RE,et al. The induction of matrix metalloproteinase and cytokine expression in synovial fibroblasts stimulated with immune cell microparticles. Proc Natl Acad Sci U S A 2005;102(8):2892e7. 7. Garcia-Vicuna R, Gomez-Gaviro MV,
Dominguez-Luis MJ, Pec MK, Gonzalez-Alvaro I, Alvaro-Gracia JM,et al. CC and CXC chemokine receptors mediate migration, proliferation, and matrix metallo-proteinase production by fibroblast-like synoviocytes from rheumatoid arthritis patients. Arthritis Rheum 2004;50(12):3866e77.
8. Haupt JL, Frisbie DD, McIlwraith CW, Robbins PD, Ghivizzani S, Evans CH,et al. Dual transduction of insulin-like growth factor-I and interleukin-1 receptor antagonist protein controls cartilage degradation in an osteoarthritic culture model. J Orthop Res 2005; 23(1):118e26.
9. Ohta S, Harigai M, Tanaka M, Kawaguchi Y, Sugiura T, Takagi K, et al. Tumor necrosis factor-alpha (TNF-alpha) converting enzyme contributes to production of TNF-alpha in synovial tissues from patients with rheumatoid arthritis. J Rheumatol 2001;28(8):1756e63. 10. Yamanishi Y, Boyle DL, Clark M, Maki RA, Tortorella MD, Arner EC,et al. Expression and regula-tion of aggrecanase in arthritis: the role of TGF-beta. J Immunol 2002;168(3):1405e12.
11. Lotz M, Moats T, Villiger PM. Leukemia inhibitory factor is expressed in cartilage and synovium and can con-tribute to the pathogenesis of arthritis. J Clin Invest 1992;90(3):888e96.
12. Stebulis JA, Rossetti RG, Atez FJ, Zurier RB. Fibro-blast-like synovial cells derived from synovial fluid. J Rheumatol 2005;32(2):301e6.
13. Knudson CB, Knudson W. Hyaluronan and CD44: mod-ulators of chondrocyte metabolism. Clin Orthop 2004; (427 Suppl):S152e62.
14. Nagano O, Saya H. Mechanism and biological signifi-cance of CD44 cleavage. Cancer Sci 2004;95(12): 930e5.
15. Laurent TC, Fraser JR. Hyaluronan. FASEB J 1992; 6(7):2397e404.
16. Balazs EA, Watson D, Duff IF, Roseman S. Hyaluronic acid in synovial fluid. I. Molecular parameters of hyaluronic acid in normal and arthritis human fluids. Arthritis Rheum 1967;10(4):357e76.
17. Wang CT, Lin J, Chang CJ, Lin YT, Hou SM. Therapeu-tic effects of hyaluronic acid on osteoarthritis of the knee. A meta-analysis of randomized controlled trials. J Bone Joint Surg Am 2004;86-A(3):538e45. 18. Qiu B, Liu SQ, Peng H, Wang HB. The effects of sodium
hyaluronate on mRNA expressions of matrix metallo-proteinase-1, -3 and tissue inhibitor of metalloprotei-nase-1 in cartilage and synovium of traumatic osteoarthritis model. Chin J Traumatol 2005;8(1):8e12. 19. Sasaki A, Sasaki K, Konttinen YT, Santavirta S, Takahara M, Takei H, et al. Hyaluronate inhibits the interleukin-1beta-induced expression of matrix metal-loproteinase (MMP)-1 and MMP-3 in human synovial cells. Tohoku J Exp Med 2004;204(2):99e107. 20. Takahashi K, Goomer RS, Harwood F, Kubo T,
Hirasawa Y, Amiel D. The effects of hyaluronan on matrix metalloproteinase-3 (MMP-3), interleukin-1beta (IL-1beta), and tissue inhibitor of metalloproteinase-1 (TIMP-1) gene expression during the development of os-teoarthritis. Osteoarthritis Cartilage 1999;7(2):182e90. 21. Wilhelm J, Pingoud A. Real-time polymerase chain
re-action. Chembiochem 2003;4(11):1120e8.
22. Ponchel F, Toomes C, Bransfield K, Leong FT, Douglas SH, Field SL, et al. Real-time PCR based on SYBR-Green I fluorescence: an alternative to the TaqMan assay for a relative quantification of gene rearrangements, gene amplifications and micro gene deletions. BMC Biotechnol 2003;3(1):18.
23. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J Mol Biol 1990; 215(3):403e10.
24. Whelan JA, Russell NB, Whelan MA. A method for the absolute quantification of cDNA using real-time PCR. J Immunol Methods 2003;278(1e2):261e9.
25. Ovstebo R, Haug KB, Lande K, Kierulf P. PCR-based calibration curves for studies of quantitative gene expression in human monocytes: development and evaluation. Clin Chem 2003;49(3):425e32.
26. Pfaffl MW, Horgan GW, Dempfle L. Relative expression software tool (REST) for group-wise comparison and statistical analysis of relative expression results in real-time PCR. Nucleic Acids Res 2002;30(9):e36. 27. Glasson SS, Askew R, Sheppard B, Carito B,
Blanchet T, Ma HL,et al. Deletion of active ADAMTS5 prevents cartilage degradation in a murine model of osteoarthritis. Nature 2005;434(7033):644e8. 28. Grabowski PS, Wright PK, Van ’t Hof RJ, Helfrich MH,
Ohshima H, Ralston SH. Immunolocalization of induc-ible nitric oxide synthase in synovium and cartilage in rheumatoid arthritis and osteoarthritis. Br J Rheumatol 1997;36(6):651e5.
29. Studer R, Jaffurs D, Stefanovic-Racic M, Robbins PD, Evans CH. Nitric oxide in osteoarthritis. Osteoarthritis Cartilage 1999;7(4):377e9.
30. Westacott CI, Barakat AF, Wood L, Perry MJ, Neison P, Bisbinas I, et al. Tumor necrosis factor alpha can contribute to focal loss of cartilage in osteoarthritis. Osteoarthritis Cartilage 2000;8(3):213e21.
31. Goldring MB. The role of cytokines as inflammatory me-diators in osteoarthritis: lessons from animal models. Connect Tissue Res 1999;40(1):1e11.
32. Shimizu M, Yasuda T, Nakagawa T, Yamashita E, Julovi SM, Hiramitsu T, et al. Hyaluronan inhibits 1246 C.-T. Wang et al.: HA decreases cytokine and enzyme in FLS
matrix metalloproteinase-1 production by rheumatoid synovial fibroblasts stimulated by proinflammatory cytokines. J Rheumatol 2003;30(6):1164e72. 33. Julovi SM, Yasuda T, Shimizu M, Hiramitsu T,
Nakamura T. Inhibition of interleukin-1beta-stimulated production of matrix metalloproteinases by hyaluronan via CD44 in human articular cartilage. Arthritis Rheum 2004;50(2):516e25.
34. Tanaka M, Masuko-Hongo K, Kato T, Nishioka K, Nakamura H. Suppressive effects of hyaluronan on MMP-1 and RANTES production from chondrocytes. Rheumatol Int 2006;26(3):185e90.
35. Edwards JC. Fibroblast biology. Development and dif-ferentiation of synovial fibroblasts in arthritis. Arthritis Res 2000;2(5):344e7.
36. Kontoyiannis D, Kollias G. Fibroblast biology. Synovial fibroblasts in rheumatoid arthritis: leading role or cho-rus line? Arthritis Res 2000;2(5):342e3.
37. Ritchlin C. Fibroblast biology. Effector signals released by the synovial fibroblast in arthritis. Arthritis Res 2000;2(5):356e60.
38. Goldring MB, Krane SM. Modulation by recombinant interleukin 1 of synthesis of types I and III collagens and associated procollagen mRNA levels in cultured human cells. J Biol Chem 1987;262(34):16724e9. 39. Gravallese EM, Darling JM, Ladd AL, Katz JN,
Glimcher LH.In situ hybridization studies of stromely-sin and collagenase messenger RNA expression in rheumatoid synovium. Arthritis Rheum 1991;34(9): 1076e84.
