Synovitis is a common feature of rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE), but the pattern of joint involvement differs in each disease. This study was undertaken to investigate the global gene expression profiles in synovial biopsy tissue from the swollen knees of untreated SLE patients (n = 6), RA patients (n = 7), and osteoarthritis (OA) patients (n = 6).
Synovial biopsy samples were obtained from the affected knees of patients in the 3 groups by needle arthroscopy. Half of the material was used for extraction of total RNA, amplification of complementary RNA, and high-density oligonucleotide spotted hybridization arrays. On the remaining tissue samples, real-time reverse transcription–polymerase chain reaction (RT-PCR) and immunohistochemical experiments were performed to confirm the microarray data.
SLE synovial biopsy tissue displayed a significant down-regulation of genes involved in extracellular matrix (ECM) homeostasis and a significant up-regulation of interferon-inducible (IFI) genes. Real-time RT-PCR experiments confirmed the up-regulation of selected IFI genes (IFI27, IFI44, and IFI44L) in the SLE synovial tissue. Immunohistochemical analyses showed that 3 molecules involved in ECM regulation, chondroitin sulfate proteoglycan 2, latent transforming growth factor β binding protein 2, and fibroblast activation protein α, were significantly down-regulated in SLE synovium. In contrast, immunostaining for IFI27, Toll-like receptor 4, and STAT-1 resulted in higher quantitative scores in SLE synovial tissue, which could be attributed to the fact that the RA samples had a large population of inflammatory cell infiltrates that were negative for these markers.
Arthritis in SLE has a very distinct molecular signature as compared with that in OA and RA, characterized by up-regulation of IFI genes and down-regulation of genes involved in ECM homeostasis.
Systemic lupus erythematosus (SLE) is a systemic autoimmune disorder affecting mainly young women. Arthritis may occur in 40–60% of patients and usually affects the small joints of the hands, wrists, and knees. Unlike that in rheumatoid arthritis (RA), arthritis in SLE does not induce severe bone damage and erosions, although joint deformities may occur in the course of the disease (similar to Jaccoud's arthropathy), and some patients may develop erosive disease that resembles RA (1, 2). Magnetic resonance imaging studies have confirmed the uncoupling of severe chronic inflammation from bone involvement in the course of arthritis in SLE, by identifying a subset of SLE patients whose disease has prominent soft tissue manifestations (capsular swelling, proliferative tenosynovitis, and synovial hypertrophy) but few or no bone erosions (3).
The pathophysiologic mechanisms of arthritis in SLE have not been investigated thoroughly. Histologic studies performed on SLE synovial tissue indicate the presence of moderate proliferation of the synovial lining layer and inflammatory cell infiltrates in the sublining, together with mild vascular hyperplasia (4). Yet, the underlying molecular mechanisms of joint inflammation in SLE remain unknown.
In the present study, we assessed global gene expression profiles in the synovial tissue from a small group of SLE patients with knee arthritis as compared with patients with osteoarthritis (OA) and patients with RA. This study is the first to profile gene expression levels with the use of high-density oligonucleotide spotted microarrays for comparison of SLE synovitis with the synovitis of OA and RA. Our results indicate that SLE arthritis is characterized by a very specific molecular signature that is distinct from that of OA and RA, with up-regulation of interferon (IFN)–inducible (IFI) genes and down-regulation of genes involved in extracellular matrix (ECM) homeostasis.
PATIENTS AND METHODS
Patients and synovial samples.
Synovial biopsy tissue (15–20 synovial samples per patient) was obtained by needle arthroscopy of the affected knee of patients with SLE (n = 6), patients with RA (n = 7), and patients with OA (n = 6). For each patient, 4–6 synovial samples were snap-frozen in liquid nitrogen and stored at −80° for later RNA extraction. The same amount of tissue was also kept at −80° for future immunostaining experiments on frozen sections. The remaining material was stored in formaldehyde and paraffin-embedded for conventional optical evaluation and immunostaining of selected cell markers.
All patients with SLE met the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) revised classification criteria for SLE (5), all were female, and the mean age was 32 years (range 19–40 years). All SLE patients had active articular disease at the time of synovial tissue sampling, and none had received immunosuppressive therapy; some of the SLE patients were receiving nonsteroidal antiinflammatory drugs. All patients with RA met the ACR classification criteria for RA (6) and all had early (<1 year's duration) active disease at the time of tissue sampling. Among the patients with RA, 2 were female and 5 were male, and the mean age was 51 years (range 37–69 years). In these patients, the mean C-reactive protein (CRP) level was 25 mg/liter (range 9–96 mg/liter), and the mean Disease Activity Score in 28 joints (including the CRP) (7) was 5.08 (range 3.76–5.82). None of the RA patients had received any treatment, except with nonsteroidal antiinflammatory drugs. Among the patients with OA, 5 were female and 1 was male, and the mean age was 63.2 years (range 51–73 years).
All patients had a swollen knee at the time of the needle arthroscopy procedure. The biopsy samples were harvested before initiation of disease-modifying antirheumatic drugs or any other immunosuppressive therapy. All of the RA patients were subsequently treated with methotrexate. Tumor necrosis factor (TNF) blocking agents were added to the regimen of 5 of these patients at a later stage. All of the SLE patients subsequently received antimalarial drugs. Combination therapy with methotrexate was later started because of persistent joint involvement in 3 of these patients. Azathioprine was started for severe hematologic manifestations in 2 other patients. The study was approved by the ethics committee of the Université catholique de Louvain, and informed consent was obtained from all patients.
Total RNA was extracted from synovial biopsy tissue using the Nucleospin RNA II extraction kit (Macherey-Nagel, Düren, Germany), which included DNase treatment of the samples. At least 1 μg total RNA could be extracted for further processing from all but 2 samples from SLE patients and all but 1 sample from OA patients. Labeling of RNA (complementary RNA [cRNA] synthesis) was performed according to a standard Affymetrix procedure (One-Cycle Target Labeling kit; Affymetrix UK, High Wycombe, UK). Briefly, total RNA was first reverse transcribed into single-stranded complementary DNA (cDNA) using a T7-Oligo(dT) Promoter Primer kit (Affymetrix UK) and Superscript II reverse transcriptase. RNase H was then added together with Escherichia coli DNA polymerase I and E coli DNA ligase, followed by a short incubation with T4-DNA polymerase in order to achieve synthesis of the second-strand cDNA. The purified double-stranded cDNA served as the template for the in vitro transcription reaction, which was carried out overnight in the presence of T7-RNA polymerase and a biotinylated nucleotide analog/ribonucleotide mix. At the end of this procedure, the biotinylated cRNA was cleaned and fragmented by a 35-minute incubation at 95°C.
GeneChip Human Genome U133 Plus 2.0 arrays (spotted with 1,300,000 oligonucleotides informative for 47,000 transcripts and originating from 39,000 genes; Affymetrix UK) were hybridized overnight at 45°C in monoplicate cultures with 10 μg cRNA. The slides were then washed and stained using the EukGE-WS2v5 Fluidics protocol on the GeneChip Fluidics Station (Affymetrix UK) before being scanned on a GeneChip Scanner 3000. For the initial normalization and data analysis steps, data were retrieved using Affymetrix GCOS software. The frequency of positive genes on each slide was 48–55%. After scaling of all probe sets to a value of 100, the amplification scale was reported to be between 1.1 and 2.5 for all slides. The signals yielded by the poly(A)-RNA, hybridization, and housekeeping controls (GAPDH 3′:5′ ratio <2) were indicative of the good quality of the amplification and hybridization procedures.
For RT-PCR analyses, cDNA was synthesized from RNA that originated from a subset of samples (n = 4 SLE, n = 5 RA, and n = 4 OA), using RevertAid Moloney murine leukemia virus reverse transcriptase (Fermentas, St. Leon-Rot, Germany) and oligo(dT) primers. Quantitative RT-PCR was performed on a MyiQ single-color real-time PCR detection system (Bio-Rad, Nazareth Eke, Belgium) using SYBR Green detection mix. For each sample, 5 ng cDNA was loaded in triplicate with 1× SYBR Green mix (Applied Biosystems, Foster City, CA) and the following 10-mM primers: for β-actin, 5′-GGCATCGTGATGGACTCCG-3′ and 3′-GCTGGAAGGTGGACAGCGA-3′ (amplification constant 1.96); for IFI44, 5′-GAGAGATGTGAGCCTGTGAGG-3′ and 3′-TTTTCCTTGTGCACAGTTGAT-5′ (amplification constant 1.97); for IFI27, 5′-ACCTCATCAGCAGTGACCAGT-3′ and 3′-ACATCATCTTGGCTGCTATGG-5′ (amplification constant 2.00); and for IFI44L, 5′-GTGGATGATTGCAGTGAGGTT-3′ and 3′-AATATCCTTCATGGGGTCCAG-5′ (amplification constant 2.01). The melting curves obtained after each PCR amplification confirmed the specificity of the SYBR Green assays.
Relative expression of the target genes in the studied samples was obtained using the difference in the comparative threshold (ΔΔCt) method. Briefly, for each sample, a value for the cycle threshold (Ct) was determined, defined as the mean cycle at which the fluorescence curve reached an arbitrary threshold. The ΔCt for each sample was then calculated according to the formula Ct target gene − Ct actin; ΔΔCt values were then obtained by subtracting the ΔCt of a reference sample from the ΔCt of the studied samples. Finally, the levels of expression of the target genes in the studied samples as compared with the reference sample were calculated as 2.
Histopathologic and immunohistochemical analyses of paraffin-embedded biopsy tissue.
Fresh synovial biopsy tissue samples were fixed overnight in 10% formalin buffer at pH 7.0, and embedded in paraffin for histologic and immunohistochemical analyses. Serial histologic sections were stained with hematoxylin and eosin and analyzed by an observer (IT) who was blinded to the diagnoses. The following parameters were evaluated: perivascular lymphoplasmocytic cell infiltrates, diffuse lymphoplasmocytic cell infitrates, and the thickness of the synovial lining layer. A global semiquantitative score that included the whole area of the biopsy sample was assigned as a measure of the lymphoplasmocytic cell infitrates, with scores ranging from 0 to 3 (with 0 indicating the absence of infiltrates and 3 indicating a high level of infiltrates). A specific score was also assigned for hyperplasia of the synovial lining layer, with 0 indicating 1–2 cell layers, 1 indicating 3–4 cell layers, 2 indicating 5–6 cell layers, and ≥3 indicating at least 7 cell layers.
Immunolabeling experiments were performed using a standard protocol. After removal of paraffin and inactivation of endogenous peroxidases with 0.3% H2O2 for 30 minutes at room temperature, sections were incubated in 10 mM sodium citrate buffer, pH 5.8, and heated in a bain-marie at 98°C for 75 minutes to retrieve the antigenic sites. Nonspecific binding was blocked by a 30-minute incubation with 50 mM Tris HCl, pH 7.4, containing 10% (volume/volume) normal goat serum and 1% (weight/volume) bovine serum albumin. Sections were then incubated overnight at 4°C with the primary antibody. The following monoclonal antibodies were used: CD3 (Neomarkers, Westinghouse, CA), CD8 (DakoCytomation, Glostrup, Denmark), CD20 (Biocare Medica, Walnut Creek, CA), CD68 (DakoCytomation), and CD138 (DakoCytomation). After 3 washes in 50 mM Tris HCl, pH 7.4, specifically bound antibodies were labeled for 1 hour at room temperature with Envision (DakoCytomation), and the activity of the peroxidases was revealed by a 10-minute incubation with 0.5 mg/ml diaminobenzidine in Tris HCl buffer. As a final step, sections were washed in tap water and lightly counterstained with hematoxylin. In each case, negative control experiments were performed by replacing the primary antibody with a nonrelevant antibody of the same isotype; all controls consistently produced no signal. Immunohistochemical evaluation of CD3, CD8, CD20, CD68, and CD138 was performed using a semiquantitative score on a 0–3 scale, with 0 indicating absence of expression and 3 indicating the highest expression.
Immunohistochemical analyses of frozen sections.
After initial blocking of endogenous peroxidases with a peroxidase-blocking reagent (DakoCytomation), frozen sections of the synovial biopsy samples were stained with primary antibodies for the following genes: Toll-like receptor 4 (TLR-4) (AMS Biotechnology, Oxon, UK), IFI27 (Abnova, Taipei City, Taiwan), STAT-1 (Cell Signaling Technology, Danvers, MA), latent transforming growth factor β binding protein 2 (LTBP-2) (Abnova), chondroitin sulfate proteoglycan 2 (CSPG-2) (Sigma-Aldrich, Bornem, Belgium), fibroblast activation protein α (FAP) (Bender MedSystems, Vienna, Austria), and interleukin-7 receptor (IL-7R) (Sigma-Aldrich). After incubation with the primary antibody, slides were sequentially incubated with an EnVision horseradish peroxidase (HRP) rabbit or mouse secondary antibody conjugated to an HRP-labeled polymer (Dako EnVision+ System; DakoCytomation) and diaminobenzidene-positive chromagen (DakoCytomation). The slides were subsequently counterstained with hematoxylin and eosin for further analyses. Semiquantitative analyses were performed using a semiquantitative score on a scale of 0–3.
Quantitative analysis of the immunostained sections was performed using ImageJ software (http://rsb.info.nih.gov/ij/). Six digitalized pictures (400× magnification) were obtained for each slide. The surface staining and staining of the surface of the nuclei were determined for each image, and the area of staining was then normalized by calculating the ratio of surface staining to nuclei staining.
Statistical analyses of the microarray data were performed using GeneSpring software (Agilent, Palo Alto, CA). For each slide, scaled data were normalized to the 50th percentile value for each chip and to the median value for each gene. The data were assessed by analysis of variance (ANOVA) with or without Benjamini-Hochberg corrections for multiple comparisons, with the minimal level of differential expression between RA or SLE and OA synovial tissue samples set at a limit of 2-fold. Gene clustering studies were performed using the Genesis program (http://genome.tugraz.at/genesisserver/), after processing of the scaled data for selection of the 8,000 genes that displayed the widest interindividual variations. Complete linkage clustering was performed based on Pearson's correlation analyses between the selected genes. Pathway analyses were performed on all genes identified by ANOVA using Gostat (http://gostat.wehi.edu.au/), which is an application that searches for statistically overrepresented gene ontology (GO) terms within a group of genes (8).
Microarray hybridizations and results of unsupervised clustering.
In the first set of experiments, RNA was extracted from snap-frozen samples of synovial biopsy tissue obtained from untreated OA, RA, and SLE patients. The RNA was labeled and hybridized in monoplicate on GeneChip Human Genome U133 Plus 2.0 slides. To investigate whether the gene expression profiles were globally different among the 3 patient populations, we first analyzed the samples using unsupervised hierarchical clustering. Strikingly, this procedure resulted in an initial distribution of the samples into 2 clusters, corresponding to nonautoimmune (OA) and autoimmune (SLE and RA) synovial samples. In the autoimmune group, the samples also correctly clustered according to the patient's diagnosis, thereby indicating that the 3 disorders are characterized by distinct synovial molecular signatures (Figure 1A).
Differential gene expression profiles and pathways.
We next used ANOVA to identify the genes that displayed the most significant differences in expression among the 3 groups. In SLE synovial biopsy tissue, 590 of the 54,675 transcripts present on the slides were found to be specifically up-regulated and 2,076 were down-regulated as compared with the other groups; after correction for multiple comparisons, the numbers of transcripts that were up- and down-regulated in SLE synovial tissue dropped to 40 and 34, respectively (Figure 1B). In RA synovial tissue, 310 transcripts were specifically up-regulated and 427 were down-regulated as compared with the other groups; after correction for multiple comparisons, 27 transcripts were found to be up-regulated and 13 were down-regulated (Table 1). Of the genes up-regulated in SLE synovial biopsy tissue (Figure 1B), 47% were type I IFI genes, according to the literature and database annotations. Interestingly, 35% of the genes that were down-regulated in SLE synovial tissue (Figure 1B) are involved in ECM homeostasis. In RA synovial tissue, 48% of the genes that were up-regulated (Table 1) are involved in regulation of T cell and B cell responses.
Table 1. Genes up- and down-regulated in RA synovial tissue as compared with OA and SLE synovial tissue*
Gene in RA synovial tissue
GenBank accession number
Genes that are up- or down-regulated in rheumatoid arthritis (RA) as compared with systemic lupus erythematosus (SLE) and osteoarthritis synovial tissue were identified by analysis of variance with Benjamini-Hochberg corrections for multiple comparisons. The values for fold change are the mean level of differential expression in RA as compared with SLE synovial biopsy tissue, with a defined limit of more than or less than 2. IL-7 = interleukin-7; HMG = high mobility group; RAS = rat sarcoma; FMR-1 = fragile x mental retardation protein gene; ZNF = zinc finger.
T cell receptor β-chain BV20S1, 5 BC1 mRNA expressed sequence tag
IL-2–inducible T cell kinase
Chromosome 22 open reading frame 5
Human mRNA for T cell receptor α-chain
Transcribed sequences of expressed sequence tag
Intercellular adhesion molecule 3
Homo sapiens cDNA clone IMAGE:2129397 3′
Selectin L (lymphocyte adhesion molecule 1)
Lymphocyte enhancer–binding factor 1
CD2 antigen (p50), sheep red blood cell receptor
Spermatid perinuclear RNA binding protein
Sterile α-motif domain–containing 3
Runt-related transcription factor 3
Positive regulatory domain–containing 1, with ZNF domain
Chemokine (CC motif) receptor 2
Human DNA sequence from clone RP5-991C6 on chromosome 6q14.1–15
Echinoderm microtubule–associated protein–like 1
β-amyloid (A4) precursor–like protein 2
Tissue factor pathway inhibitor
G protein–coupled receptor 64
Expressed sequence tag
Homo sapiens histone 3, H2a mRNA
Hypothetical protein FLJ90724
Glioma pathogenesis–related 1
ADAM-9 (meltrin γ)
Leukocyte immunoglobulin–like receptor, subfamily B, member 1
To perform a more accurate evaluation of the gene families enriched in the different lists of genes, we analyzed the frequency of the available GO annotations in each list, using online data-mining software. As shown in Table 2, the genes up-regulated in SLE synovial tissue were significantly enriched in genes belonging to the GO families “response to virus,” “response to pest, pathogen, or parasite,” or “response to other organism,” which are groups of GO annotations that include the IFI genes. In contrast, the genes down-regulated in SLE synovial biopsy tissue were significantly enriched in genes belonging to the “collagen,” “extracellular matrix,” and “fibrillar collagen” GO families, thereby confirming that SLE synovial tissue displays a significant down-regulation of genes involved in ECM homeostasis. Finally, the genes up-regulated in RA synovial tissue belonged to the groups of GO families that include genes involved in T cell and B cell activation (Table 2).
Table 2. Gene ontology (GO) annotations overrepresented in the lists of genes up- or down-regulated in SLE and RA synovial tissue*
Disease, gene group, GO annotation number
No. of positive genes in list/total no. in list
Total no. of positive genes/total no. of genes
P, enrichment in gene list
The frequency of available GO terms in the genes up- or down-regulated in systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) synovial biopsy tissue was calculated using online data-mining software and compared with the frequency of the same GO terms in the total list of GO annotated genes (n = 14,367). P values were calculated by chi-square test with corrections for multiple comparisons.
For the SLE samples, of 590 up-regulated genes submitted, there were 579 unique genes and 230 annotated; of 2,076 down-regulated genes submitted, there were 2,056 unique genes and 813 annotated. For the RA samples, of 310 up-regulated genes submitted, there were 304 unique genes and 134 annotated; of 427 down-regulated genes submitted, there were 422 unique genes and 156 annotated.
Real-time PCR experiments were run to confirm the microarray data on selected IFI genes. As shown in Figure 2, IFI27, IFI44, and IFI44L were significantly overexpressed in SLE synovial tissue as compared with OA and RA synovial tissue, thereby validating the differences observed in these genes in the global gene expression analyses.
In a second set of experiments, we studied the histopathologic characteristics of the synovial material obtained from the 3 groups of patients. All samples displayed a common pattern of cell membrane thickness in the synovial lining and infiltration of mononuclear cells. Semiquantitative evaluation of the stained sections on the slides indicated that RA synovial biopsy tissue was characterized by significantly more synovial hyperplasia and diffuse and perivascular mononuclear inflammatory cell infiltrates than was observed in OA and SLE synovial biopsy tissue (Figure 3).
We also performed immunohistochemical staining for specific cell populations. Semiquantitative analyses revealed that scores for CD3, CD8, CD20, and CD138 were significantly higher in RA synovial biopsy tissue than in OA and SLE synovial biopsy tissue. Semiquantitative evaluation of CD68 also yielded a higher score in RA synovial samples compared with the other groups, but the difference was not statistically significant (Figure 3).
In the final set of experiments, we performed immunostaining experiments in order to confirm, at the protein level, the results obtained on selected target genes identified by ANOVA and the results of the pathway analyses in the gene expression arrays. LTBP-2, CSPG-2 (versican), and FAP are molecules involved in ECM homeostasis. Both semiquantitative (data not shown) and quantitative (Figure 4) evaluations showed that these 3 molecules were significantly down-regulated in SLE synovial tissue as compared with RA synovial tissue, thereby confirming the results of the gene expression arrays.
Semiquantitative evaluation of IFI proteins IFI27, TLR-4, and STAT-1 revealed no difference between RA and SLE synovial biopsy tissue for any of these stains. However, in quantitative evaluations in which the surface of the staining was normalized to the cellularity of the samples, the percentage of cells positive for these 3 molecules was significantly higher in SLE synovial biopsy samples (Figure 4). Finally, both semiquantitative (data not shown) and quantitative (Figure 4) evaluations of IL-7R immunostaining confirmed that IL-7R was significantly up-regulated in RA synovial tissue as compared with SLE and OA synovial tissue.
In this study, we performed global gene expression analyses of synovial biopsy tissue from untreated SLE, RA, and OA patients with active disease. We found that the 3 diseases are characterized by distinct molecular signatures. In particular, genes involved in T cell and B cell regulation are up-regulated in RA synovial tissue. In contrast, in SLE synovial tissue, IFI genes are up-regulated while genes involved in ECM homeostasis are down-regulated. Immunostaining experiments were performed on a few representative genes, and the quantitative findings paralleled the microarray data.
In previous studies (9, 10), IFI genes were found to be up-regulated in peripheral blood mononuclear cells (PBMCs) from SLE patients as compared with healthy and disease control subjects. Our results indicate that the same molecular signature is found in SLE synovial tissue. The imbalance between IFI molecules and other molecules in SLE synovial tissue might be of interest pathophysiologically in the course of SLE arthritis. Several lines of evidence indicate that cross-regulation of type I IFNs and TNFα plays an important role in the regulation of pathologic inflammatory responses. Thus, the addition of IFNβ to stimulated human PBMCs inhibits the production of both IL-1β and TNFα and induces the production of IL-1R antagonist (11). In another study, in vivo administration of IFNβ to patients with multiple sclerosis resulted in decreased numbers of TNFα-secreting PBMCs as compared with that in patients who were untreated (12). Conversely, Palucka et al found that the addition of TNFα inhibited virus-induced IFNα production by PBMCs, whereas in vitro blockade of TNFα induced IFNα production by virus-exposed immature plasmacytoid dendritic cells (13). In that study, in vivo administration of TNF-blocking agents resulted in increased transcription of IFI genes in the PBMCs (13).
In the present study, we did not observe decreased TNF gene expression in SLE synovial tissue as compared with RA and OA synovial tissue (data not shown). However, a group of genes typically produced by synovial fibroblasts and involved in ECM homeostasis was found to be significantly down-regulated in SLE biopsy tissue. Our immunostaining experiments confirmed that synovial tissue from SLE patients is characterized by a significant down-regulation of LTBP-2, CSPG-2, and FAP. Interestingly, FAP is a serine protease that is produced by activated, but not resting, stromal fibroblasts and plays an important role in both the growth and migration of fibroblasts and the interaction of fibroblasts with the ECM environment. Recent studies have shown an increased expression of FAP by activated chondrocytes and chondrocytes from patients with OA (14). The differential expression of FAP in SLE arthritis as compared with arthritis in OA or RA is further evidence that the molecule might play an important role in the cartilage and bone damage associated with these latter diseases.
However, important methodologic issues need to be addressed when considering the potential pathophysiologic relevance of our observations. Synovial tissue from patients with arthritis is composed of heterogeneous cell populations, including resident synovial fibroblasts or macrophages and infiltrating inflammatory cells. Our results indicated that the distribution of these cell populations was different among the 3 groups of patients; in particular, RA synovium was characterized by a higher amount of infiltrating T cells and B cells as compared with that in SLE and OA synovium. It therefore remains to be elucidated whether the differences observed in global gene expression among the different groups of patients might be related to these differences in cell subpopulations.
In accordance with this perspective, validation of the microarray data at the protein level is a necessary step. This could be accomplished using immunostaining experiments aimed at visualizing selected IFI proteins. In the present study, semiquantitative evaluation of the immunostained samples showed no difference between RA and SLE synovial tissue with regard to the expression of STAT-1, IFI27, and TLR-4. However, this evaluation procedure did not include a normalization step. When the staining results were normalized to the number of cells during the process of digitalized quantification in SLE synovial tissue, the 3 molecules were found to be significantly enriched. Of note, microarray-based generation of data includes a similar process of normalization, in which individual gene expression is normalized to the total amount of material hybridized on the slides. Thus, the results generated by the 2 techniques are comparable.
In this context, our data should be interpreted with caution. The present findings indicate that the increased expression of IFI genes in SLE synovial tissue may, in fact, be attributable to the presence of a large amount of infiltrating inflammatory cells that are negative for these genes in RA synovial tissue. As yet, no information regarding the production of IFI genes at the level of the single cell is available. This information would help confirm that SLE synovial cells are characterized by an increased production of IFI genes as compared with RA synovial cells.
Further studies are therefore needed to elucidate the pathophysiologic mechanisms of arthritis in SLE. In particular, molecular and pathologic studies of the purified cell populations within the synovium will provide a more reliable map of the molecular changes associated with synovitis in SLE. In addition, in vitro mechanistic studies of cultured synovial cells involving the use of type I IFNs to modulate production of cartilage-degradating enzymes would also be of interest.
Dr. Lauwerys had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Nzeusseu Toukap, Galant, Theate, Houssiau, Lauwerys.
Acquisition of data. Nzeusseu Toukap, Galant, Theate, Maudoux, Houssiau, Lauwerys.
Analysis and interpretation of data. Nzeusseu Toukap, Galant, Theate, Maudoux, Lories, Houssiau, Lauwerys.