SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

Synovial fibroblast (SF) hyperplasia contributes to the pathogenesis of rheumatoid arthritis (RA), but quantitative information on this process is scarce. This study was undertaken to evaluate the fibroblast-specific marker Hsp47 as a quantitative marker for SFs and to analyze its clinicopathologic correlates and evolution after anti–tumor necrosis factor α (anti-TNFα) therapy.

Methods

Synovial biopsy samples were obtained from 48 patients with RA and 20 controls who were healthy or had osteoarthritis (OA). Twenty-five RA patients who had active disease at the time of biopsy underwent a second biopsy after anti-TNFα therapy. Immunolabeling for Hsp47, inflammatory cells, and vascular cell markers was performed. Hsp47-positive lining and sublining fractional areas were quantified, and their correlation with clinicopathologic variables was analyzed.

Results

In normal and diseased synovial tissue, Hsp47 was specifically and uniformly expressed by lining, sublining, and perivascular fibroblasts. Lining SF area was significantly increased in both RA and late OA tissue compared to normal tissue. Sublining SF area was increased in RA tissue but not in late OA tissue compared to normal tissue. Lining SF area was positively correlated with macrophage density, Disease Activity Score in 28 joints, and RA disease duration. In contrast, sublining SF area was negatively correlated with RA disease duration and activity. A significant reduction in lining SF area but not sublining SF area was observed after anti-TNFα therapy.

Conclusion

Our findings indicate that Hsp47 is a reliable marker for quantifying SFs in human synovial tissue. Our data suggest that lining and sublining SFs undergo different dynamics during the course of the disease. Lining SF expansion parallels the activity and temporal progression of RA and can be partially reversed by anti-TNFα therapy.

Synovial fibroblasts (SFs), also termed fibroblast-like or type B synoviocytes, are the most abundant resident cell type in human synovial tissue. In rheumatoid arthritis (RA) synovium, SF expansion may occur by either acquired growth advantages or the recruitment of variably differentiated precursors (1–4). Numerous lines of evidence support the potential contribution of SFs to the pathogenesis of chronic arthritis (5, 6). SFs respond to cytokines, notably tumor necrosis factor α (TNFα), by producing a large variety of mediators of inflammation and tissue destruction. In addition, arthritic SFs display a constitutive proinflammatory phenotype that persists in tissue culture in the absence of exogenous stimuli (7–9). Both cytokine-induced responses and constitutive changes often converge to common pathways that result in increased synthesis of chemokines, cytokines, proangiogenic factors, and factors related to increased invasiveness and tissue destruction (5–9). Therefore, the expansion of an SF pool with an abnormal phenotype could significantly contribute to chronic inflammation and destruction of the joints. Proof of this concept has been generated in animal models of arthritis, where specifically targeting fibroblast TNFα receptors is sufficient to preclude the development of TNFα-mediated arthritis (10). Interruption of cadherin-mediated SF cell–cell adhesions critical to lining and pannus formation has also been shown to reduce the severity of arthritis in mice (11).

The morphologic features of SF hyperplasia in the synovial lining and the cartilage-invasive pannus have been described previously (12, 13). However, since no reliable markers for SFs are available, truly quantitative data on SF hyperplasia are scarce. The expression of different SF markers can be modified by disease status and by the different locations of SFs in the synovium. Increased uridine diphosphoglucose dehydrogenase (UDPGD), CD55 (decay-accelerating factor [DAF]), vascular cell adhesion molecule (VCAM), and cadherin 11 expression in lining versus sublining SFs has been reported, and the expression and distribution of these markers vary between normal synovium, RA synovium, and cultured SFs exposed to cytokines (14–20). Fibroblast lineage markers such as prolyl hydroxylases or Thy1 also show important sensitivity and specificity limitations (21–23). This paucity of specific phenotypic markers explains the lack of quantitative information on the potential changes in SFs in relation to RA inflammatory activity or therapeutic responses.

Most SF markers are preferentially expressed by lining SFs, and no information is available on changes in sublining SFs in RA. However, only sublining SFs are in direct contact with infiltrating lymphocytes and in close proximity to blood vessels. Therefore, their participation in processes in which direct cell–cell contact or proximity to other cell elements is required seems critical (5, 9, 24, 25).

Immunohistochemical detection of the collagen-specific chaperone Hsp47 has recently been demonstrated to be a highly specific and sensitive method to identify fibroblasts in healthy or pathologic human tissue (23, 26). Hsp47 is not expressed by macrophages, endothelial cells, or smooth muscle cells, and in fibroblasts, it is constitutive and not dependent on their activation status. In this study, we confirmed the validity of this marker for quantifying both lining and sublining SFs in synovial tissue and analyzed the correlations between SF expansion and relevant clinicopathologic variables in a series of RA patients. Our data suggest that lining SF expansion but not sublining SF expansion is a dynamic component of RA synovitis that parallels the inflammatory activity and temporal progression of the disease and can be partially reversed in response to anti-TNFα therapy.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Patients and synovial biopsies.

Arthroscopic synovial tissue biopsy samples were obtained from the knees of 48 patients who fulfilled the American College of Rheumatology revised criteria for RA (27). All patients had active disease characterized by inflammation of at least the biopsied knee joint regardless of previous therapy. Arthroscopy of an actively inflamed knee, defined as a painful and swollen joint with inflammatory synovial fluid, was performed with a 2.7-mm arthroscope (Storz) under local anesthesia. In all patients, 6–8 biopsy samples were obtained from the suprapatellar pouch and medial and lateral gutters with a 3-mm grasping forceps (Storz). Tissue specimens were formalin fixed and paraffin embedded.

Patient characteristics at biopsy, including disease duration, previous therapy, Disease Activity Score in 28 joints (DAS28), and the presence of autoantibodies (rheumatoid factor or anti–cyclic citrullinated peptide [anti-CCP]) and erosions were recorded. (These data are available online at http://www.imas12.es/doc/patientsbiopsy.pdf.)

Written informed consent was obtained from all patients. The present study was approved by the institutional ethics committees of both participating centers (Ethics Committee of the Hospital Clínic de Barcelona and Clinical Research Ethics Committee of the Hospital 12 de Octubre).

A subgroup of 25 patients in which anti-TNFα therapy (etanercept, adalimumab, or infliximab) was started after the first biopsy, due to active disease refractory to previous methotrexate therapy (mean ± SD DAS28 6.0 ± 1.4), underwent a second biopsy after a mean ± SD of 10 ± 2 months of anti-TNFα therapy plus methotrexate.

Control synovial tissue samples from 14 patients with osteoarthritis (OA) were obtained by synovectomy at prosthetic joint replacement surgery. The whole synovium was collected, and multiple samples from the best-preserved synovial surface were obtained and similarly processed. In addition, normal uninflamed synovial tissue samples were obtained from 6 individuals without previous joint disease at elective arthroscopic surgery for minor meniscal lesions. Lack of inflammatory changes in normal tissue was confirmed by histologic examination.

Immunolabeling of synovial tissue.

Tissue samples were deparaffinized, rehydrated, and microwaved in EDTA (pH 9) for antigen retrieval. Fibroblasts were immunolabeled with anti-Hsp47 monoclonal antibodies (mAb) (IgG2b M16.10A1 clone; Assay Designs) and IgG2b-specific Alexa 594 (red) or Alexa 488 (green) (Molecular Probes Invitrogen) or biotinylated secondary antibodies. Sections were counterstained with DAPI or hematoxylin. Immunoperoxidase staining and quantification of T cells, B cells, and macrophages were performed as previously described (28).

Double labeling of endothelium was performed with anti-CD31 mAb (IgG1 JC70A clone; Dako) as previously described (29). Double labeling of Hsp47 and mononuclear cells was performed with anti-CD45 (IgG1 2B11/PD7/26 clone; Dako) as a pan-leukocyte marker or anti-CD68 (IgG1 KP1 clone; Dako) as a macrophage marker. After Hsp47 labeling, sequential incubation with the second primary IgG1 mAb and IgG1-specific secondary antibodies Alexa 594 (red) or Alexa 488 (green) was performed.

The entire area of each tissue specimen was photographed and digitized using a SPOT RT CCD camera and SPOT 4.0.4 software (Diagnostic Instruments) on an Axioplan 2 fluorescence microscope (Zeiss). The SF fractional area was quantified on anti-Hsp47 Alexa 594–immunolabeled sections using ImageJ software (NIH Image, National Institutes of Health, Bethesda, MD; online at: http://rsb.info.nih.gov/ij). Since the total lining area increased in parallel to the degree of SF hyperplasia, the Hsp47-positive lining area was adjusted to the linear horizontal length (mm2/mm) of the analyzed lining, whereas the Hsp47-positive sublining area was adjusted to the total area (mm2/mm2). Lining and sublining regions were separated by drawing a basal line under the superficial area where cells were clearly arranged by layers in an epithelial-like manner.

Immunolabeling of cultured SFs.

SFs from the synovial tissue of patients with RA or normal controls were obtained by explant growth from arthroscopic or synovectomy specimens in 10% fetal calf serum/Dulbecco's modified Eagle's medium. For immunofluorescence labeling, cells were grown on glass coverslips (10-mm diameter) after the third passage and immunolabeled for Hsp47 using the same protocol as for synovial tissue. Parallel cultures were left untreated or were treated with TNFα (20 ng/ml) for 24 hours and analyzed.

Different coverslips containing the same SF line, either left untreated or treated with TNFα, were mounted on the same slide and used for Hsp47 labeling. This guaranteed equal exposure to all reagents throughout the labeling procedure. Duplicate slides were processed in parallel for isotype control labeling.

Flow cytometric analysis of Hsp47 expression in SF cultures left untreated or treated with TNFα was performed. Cells were detached by EDTA treatment, fixed with Cytofix/Cytoperm (BD Biosciences), and permeabilized in 0.1% Triton X-100. The same anti-Hsp47 clone used for immunolabeling of synovial tissue and phycoerythrin-conjugated secondary goat anti-mouse IgG (30) (Jackson ImmunoResearch) were used. Cells were analyzed on a BD FACSCalibur instrument (BD Biosciences).

Statistical analysis.

For cross-sectional analyses, quantitative variables were compared by Mann-Whitney U test or one-way analysis of variance with Tukey's post hoc test. Changes in quantitative variables after anti-TNFα therapy were tested by Wilcoxon's signed rank test for paired data.

Correlation between different numerical variables was analyzed by Pearson's test. Disease duration did not follow a normal distribution, and, therefore, its correlations were analyzed by Spearman's test. Bonferroni corrections for multiple comparisons were not applied due to the relatively low number of comparisons and the exploratory nature of these analyses. Data are presented as the mean ± SEM.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Immunolabeling of SFs in synovial tissue and tissue culture.

In hematoxylin-counterstained sections from healthy or diseased synovial tissue, anti-Hsp47 immunolabeling showed a specific pattern that included abundant cells in the lining, fibroblast-shaped cells in the sublining, and some cells with a perivascular distribution, whereas mononuclear cell infiltrates were Hsp47 negative (Figure 1). In RA hyperplasic lining, Hsp47-labeled cells tended to accumulate at the basal rather than the superficial layers, consistent with the previously described distribution of UDPGD activity (14).

thumbnail image

Figure 1. Hsp47 immunolabeling of synovial tissue from patients with rheumatoid arthritis (RA) and controls. Left, Hsp47 and CD31 (top), CD68 (middle), or CD45 (bottom) double immunofluorescence labeling in sections from RA synovial tissue. Merged images are shown. Arrows in the top row indicate perivascular Hsp47-positive cells. Arrows in the middle row indicate the lining surface. DAPI nuclear counterstained; original magnification × 400. Right, Anti-Hsp47 immunoperoxidase (brown) labeling of lining and superficial sublining areas in RA tissue, osteoarthritic (OA) tissue, and normal (N) tissue. Hematoxylin counterstained; original magnification × 400. Bars show the mean ± SEM Hsp47-positive lining or sublining fractional area in tissue samples from 48 patients with RA, 14 patients with OA, and 6 normal controls. ∗ = P = 0.005; # = P = 0.04; ¶ = no significant difference versus normal tissue, by Mann-Whitney U test.

Download figure to PowerPoint

Cultured SFs from either normal or RA tissue displayed a uniform cytoplasmic pattern of Hsp47 labeling. No unlabeled cells were detected by DAPI counterstaining in SF cultures (Figure 2). Pretreatment of SF cultures with TNFα did not modify the pattern or the relative intensity of immunofluorescence labeling compared to that in untreated cells (Figure 2). Flow cytometric analysis confirmed that all cultured SFs were universally labeled by Hsp47. The mean fluorescence intensity (MFI) of Hsp47 fluorescence labeling was not increased by TNFα treatment. The mean ± SEM percentage change in MFI after TNFα treatment was −7 ± 6% (P = 0.17; n = 3).

thumbnail image

Figure 2. Hsp47 immunolabeling of cultured synovial fibroblasts (SFs). Left, Isotype control without counterstaining and with DAPI counterstaining. Middle, Hsp47 immunolabeling of untreated rheumatoid arthritis SFs (RASFs) and normal SFs cultured on glass coverslips and DAPI nuclear counterstained. Right, Hsp47 immunolabeling of RASFs and normal SFs treated with tumor necrosis factor α (TNFα; 20 ng/ml) for 24 hours and DAPI nuclear counterstained. Original magnification × 400.

Download figure to PowerPoint

When double labeling was performed, neither lining nor sublining macrophages (CD68+) were labeled by anti-Hsp47 (Figure 1). In the lining, all Hsp47-negative cells were CD68+, excluding the presence of Hsp47-negative fibroblasts in this area. Similarly, double labeling with the pan-leukocyte marker CD45 and Hsp47 showed that they were mutually exclusive, ruling out Hsp47 labeling of all lymphoid and myeloid cell types. To better identify Hsp47-labeled perivascular cells, double labeling of endothelial cells with anti-CD31 and Hsp47 was performed. Perivascular fibroblasts or pericytes of small vessels, but not endothelial cells, were Hsp47 positive (Figure 1). In larger blood vessels of deeper areas, perivascular fibroblasts, but not smooth muscle cells, were labeled by anti-Hsp47 mAb (results not shown). The intensity of labeling was uniform and similar for lining, sublining, and perivascular fibroblasts.

The density and distribution of Hsp47-labeled SFs in normal or diseased tissue were variable. In healthy synovial tissue, a single layer of alternating SF and nonfibroblast cells was observed in the lining (Figure 1). In OA and RA tissue, an increased proportion of Hsp47-positive SFs was observed, arranged in several layers toward the basal lining area. The Hsp47-positive lining fractional area was significantly increased in both RA and late OA tissue compared to normal tissue (Figure 1). The Hsp47-positive sublining area was significantly increased in RA tissue compared to healthy tissue, whereas no differences between late OA and normal tissue were found (Figure 1).

Correlations between SF hyperplasia and clinicopathologic variables.

Lining and sublining SF areas were highly variable between different patients with RA, who were heterogeneous in terms of disease duration and RA activity or severity. We therefore analyzed whether clinical activity (DAS28), disease duration, density of infiltration of inflammatory cells (CD68, CD3, and CD20), and CD31+ blood vessels were correlated with the observed increase in lining or sublining SF areas (Table 1).

Table 1. Correlations of SF areas in RA tissue with clinicopathologic variables*
 Lining SF areaSublining SF area
PrPr
  • *

    SF = synovial fibroblast; RA = rheumatoid arthritis; CRP = C-reactive protein; DAS28 = Disease Activity Score in 28 joints. Correlations were determined using Pearson's test for all variables except for disease duration, which was analyzed by Spearman's test.

No. of CD3+ cells/mm20.7510.050.695−0.06
No. of CD20+ cells/mm20.890−0.020.675−0.06
No. of CD68+ sublining cells/mm0.0170.380.9220.02
No. of CD31+ vessels/mm20.0780.250.6620.06
CRP0.0880.300.089−0.25
DAS280.0390.240.041−0.30
Disease duration, months0.0130.360.003−0.43

The DAS28 was significantly correlated with the lining SF area. The duration of the disease before arthroscopy was also significantly correlated with the lining SF area. A progressive increase in the lining SF area was observed from the earliest phases of the disease (Figure 3). The mean lining SF area was significantly increased in the group with longer disease durations (>12 months) than in the group with shorter disease durations (Figure 3). The lining SF area was significantly and positively correlated with the density of infiltration by CD68+ macrophages in the sublining, a validated parameter for inflammatory activity (30). No statistically significant correlation with lymphocytic infiltration by T cells or B cells or with vascularity was found.

thumbnail image

Figure 3. Correlations between mean lining or sublining synovial fibroblast (SF) area in rheumatoid arthritis (RA) tissue and clinicopathologic variables. Top, Correlations between lining SF area and Disease Activity Score in 28 joints (DAS28), disease duration, and CD68+ macrophage density. Bottom left and middle, Correlations between sublining SF area and DAS28 and disease duration. Squares represent samples from individual RA patients (n = 48). Bottom right, Hsp47-positive lining and sublining SF areas in groups stratified by disease duration. Bars show the mean ± SEM. ∗ = P values determined by global analysis of variance for all 3 groups. P values were less than 0.05 for disease duration <6 months versus disease duration >12 months in both lining SFs and sublining SFs, by post hoc test.

Download figure to PowerPoint

The sublining SF area was increased during the earliest phases of the disease in RA tissue, compared to OA or healthy control tissue, and decreased significantly with longer disease duration (Figure 3). Contrary to the lining SF area, the sublining SF area was negatively correlated with DAS28 (Table 1). No other significant correlations between sublining SF area and other clinicopathologic characteristics were found.

Significant differences in the lining or sublining SF area in patients stratified by erosive disease at biopsy, or by the presence of rheumatoid factor or anti-CCP autoantibodies, were not found (data not shown).

Changes in SF hyperplasia after anti-TNFα therapy.

In a subgroup of 25 patients, sequential biopsies were performed before initiating anti-TNFα therapy because of an inadequate response to disease-modifying antirheumatic drugs and a mean ± SEM of 10 ± 2 months later, when a good or moderate European League Against Rheumatism (EULAR) response (31) had been achieved in 18 patients (72%) but not in 7 patients (28%). Additional clinical and pathologic changes induced by anti-TNFα therapy in this series have previously been reported (29).

In the entire group of patients treated with anti-TNFα, a significant reduction in lining SF area was observed after the second biopsy, whereas sublining fibroblasts remained unchanged after anti-TNFα therapy (Figure 4). The decrease in lining SFs after anti-TNFα therapy was observed in both the responder group (those patients in whom a good or moderate EULAR response was achieved) and the nonresponder group. A trend toward a greater decrease in lining SFs in responders was observed, but it did not reach statistical significance (Figure 4). Sublining SFs remained similarly unchanged by anti-TNFα therapy in both responders and nonresponders (Figure 4). Neither basal lining nor sublining SF area predicted DAS28 change or EULAR response to anti-TNFα therapy (data not shown).

thumbnail image

Figure 4. Changes in lining and sublining synovial fibroblast (SF) areas in rheumatoid arthritis (RA) tissue after anti–tumor necrosis factor α (anti-TNFα) therapy. Left, Hsp47 immunofluorescence labeling in a representative responder (Resp) and a representative nonresponder (Nonresp) before and after anti-TNFα treatment. Responders (n = 18) were defined as patients in whom a moderate or good European League Against Rheumatism (EULAR) response was achieved at second biopsy; nonresponders (n = 7) were patients in whom a moderate or good EULAR response was not achieved. Arrows indicate the lining surface. DAPI counterstained; original magnification × 400. Right, Hsp47-positive lining and sublining areas in pretreatment biopsy samples (pre) and in posttreatment biopsy samples (post) and changes in Hsp47-positive lining and sublining areas in nonresponders and responders. Bars show the mean ± SEM (n = 25 patients). ∗ = P < 0.05 versus pretreatment biopsy samples, by Wilcoxon's matched pairs signed rank test. NS = not significant.

Download figure to PowerPoint

Analysis of correlations between the decrease in lining SFs and the change in other clinicopathologic parameters also showed a positive but nonsignificant trend for CD3+ T cells, CD68+ sublining macrophages, and C-reactive protein level, but not for CD20+ B cells, vascularity (CD31), or DAS28 (Table 2).

Table 2. Correlations between changes in lining SF area and clinicopathologic changes after anti-TNFα therapy*
 Pr
  • *

    Change was defined as the basal value minus the value after anti–tumor necrosis factor α (anti-TNFα) treatment. Correlations were determined using Pearson's test. See Table 1 for other definitions.

Change in no. of CD3+ cells/mm20.0820.35
Change in no. of CD20+ cells/mm20.6860.08
Change in no. of CD68+ sublining cells/mm0.0710.36
Change in no. of CD31+ vessels/mm20.452−0.15
Change in CRP0.0670.37
Change in DAS280.1770.27

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

During the past decades, numerous studies have pointed to RASFs as active drivers of synovial inflammation and joint destruction (3–8). Most previous studies have described the gene expression and phenotypic changes in RASFs that explain their transition from normal connective tissue cell components with synthetic functions to proinflammatory and destructive cells. Lining hyperplasia has long been described in RA, OA, and other inflammatory conditions as the best evidence of SF expansion (12, 32). Lining thickness variation correlates with activity and can decrease after effective therapy (12, 33–35). However, most cells in this area are macrophages, and, therefore, the reduction may also be due to a decrease in macrophages (32, 35). In the sublining, where interactions between lymphocytes and SFs seem important (24, 25), changes in SFs have not been described. The present study confirms that Hsp47 immunolabeling is a useful fibroblast lineage marker in synovial tissue. This marker permits the gathering of quantitative data on both lining and sublining SF fractions that can be correlated with clinical and therapeutic changes.

Our analysis shows that both SF pools are significantly expanded in RA tissue, even at the earliest stages of the disease, compared to normal synovial tissue. However, we found important differences between the 2 SF subpopulations. Whereas lining SF hyperplasia tended to increase with time and inflammatory activity, the opposite was true for sublining SFs. Furthermore, lining SF hyperplasia was not specific to RA and was similar to what was observed in end-stage OA tissue, consistent with previous observations (32). In contrast, sublining SF hyperplasia was an RA-specific feature that was not observed in OA tissue. Therefore, lining and sublining SFs seem to follow different dynamics in both diseases.

Previous studies have shown that differences between lining and sublining SFs in the expression of some SF factors, such as cadherin 11, VCAM, or DAF (CD55), are lost in RA, where a more homogeneous phenotype is observed (14–17). This may be due to the inducible expression of these markers in SFs by proinflammation mediators, such as TNFα (18–20). This inflammatory transformation of sublining SFs into lining-like SFs, together with the inverse relationship between sublining SF area and lining SF area during the disease course, suggests that spatial accumulation of SFs in the lining and development of a proinflammatory phenotype could be related processes. Alternatively, sublining SF expansion may be a less stable process that stops early under the influence of persistent inflammation. A selective effect of therapies on this SF pool cannot be ruled out. However, the absence of a detectable effect of anti-TNF therapy provides evidence against such a possibility. These observations would also suggest that sublining SFs may be less relevant than lining SFs as proinflammatory or chronic destructive factors.

The possibility of specific targeting of the stromal cell component in arthritis has been explored only in animal models, where all evidence is consistent with a relevant contribution of SFs to the inflammatory and joint destruction process (10, 11). The indirect effects of available RA therapies on SF hyperplasia have not been confirmed, possibly because of the aforementioned limitations to obtaining quantitative data (33, 35). The effects of TNFα antagonists are of particular interest, since TNFα is not only a critical factor in the proinflammatory and tissue destructive response of SFs, but also a proliferative and survival factor for SFs (1–3, 10).

In this study, we demonstrated a significant reduction in lining SFs, but not sublining SFs, in response to anti-TNFα therapy. However, it is unclear from our data whether lining reduction parallels clinical response. This is likely due to the small number of nonresponder patients in our series. The net reduction observed in nonresponders might be explained by subclinical therapeutic effects. Protective effects of anti-TNFα therapy on the joints have previously been demonstrated even in the absence of a clinical response (36). Therefore, the significance of the observed decrease in lining SF hyperplasia after therapy remains unclear, and further, larger studies are needed.

The mechanisms of the decrease in lining SFs after treatment may relate to changes in the balance between proliferation and survival or to a reduction in the recruitment of precursors. We did not detect an increased SF apoptosis rate in biopsy samples from anti-TNFα–treated patients, but this could be due to the very low rate of detectable apoptotic events in SFs compared to lymphoid infiltrates or endothelium (Izquierdo E, et al: unpublished observations). Nevertheless, the observed decrease in lining SFs after therapy was only partial, even in responders. Lining SFs remained significantly increased in this group compared to normal synovium. The prognostic significance of persistent SF hyperplasia should be addressed in further longitudinal studies.

In summary, we provide evidence of the utility of Hsp47 immunolabeling as a fibroblast marker in synovial tissue that may facilitate further studies on this important cell component in arthritis. Our data demonstrate differential changes in the lining and sublining SF compartments, consistent with a different dynamic during the disease course and in response to anti-TNFα therapy.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Pablos 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 conception and design. Cañete, Pablos.

Acquisition of data. Izquierdo, Cañete, Celis, Del Rey, Usategui, Marsal, Sanmartí.

Analysis and interpretation of data. Izquierdo, Cañete, Del Rey, Usategui, Criado, Pablos.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We are grateful to the Servicio de Traumatología y Cirugía Ortopédica (Hospital 12 de Octubre) for providing control synovial tissue. We also thank Vanessa Miranda for excellent technical assistance.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES