Decreased levels of the gelsolin plasma isoform in patients with rheumatoid arthritis
Gelsolin is an intracellular actin-binding protein involved in cell shape changes, cell motility, and apoptosis. An extracellular gelsolin isoform, plasma gelsolin circulates in the blood of healthy individuals at a concentration of 200 ± 50 mg/L and has been suggested to be a key component of an extracellular actin-scavenging system during tissue damage. Levels of plasma gelsolin decrease during acute injury and inflammation, and administration of recombinant plasma gelsolin to animals improves outcomes following sepsis or burn injuries. In the present study, we investigated plasma gelsolin in patients with rheumatoid arthritis.
Circulating and intra-articular levels of plasma gelsolin were measured in 78 patients with rheumatoid arthritis using a functional (pyrene-actin nucleation) assay and compared with 62 age- and gender-matched healthy controls.
Circulating plasma gelsolin levels were significantly lower in patients with rheumatoid arthritis compared with healthy controls (141 ± 32 versus 196 ± 40 mg/L, P = 0.0002). The patients' intra-articular plasma gelsolin levels were significantly lower than in the paired plasma samples (94 ± 24 versus 141 ± 32 mg/L, P = 0.0001). Actin was detected in the synovial fluids of all but four of the patients, and immunoprecipitation experiments identified gelsolin-actin complexes.
The plasma isoform of gelsolin is decreased in the plasma of patients with rheumatoid arthritis compared with healthy controls. The reduced plasma concentrations in combination with the presence of actin and gelsolin-actin complexes in synovial fluids suggest a local consumption of this potentially anti-inflammatory protein in the inflamed joint.
KeywordsRheumatoid Arthritis Rheumatoid Arthritis Patient Synovial Fluid DMARD Treatment Plasma Gelsolin
catalytic domain of matrix metalloproteinase-3
disease-modifying anti-rheumatic drug
recombinant human plasma gelsolin
Plasma gelsolin (pGSN) is the extracellular isoform of a ubiquitous cytoplasmic actin-binding protein, gelsolin (GSN), that mediates cell shape changes and motility . Differentially processed mRNA transcripts present in various cell types [2, 3] and originating from a gene on chromosome 9 program the synthesis of intracellular gelsolin (cGSN) and of its secreted counterpart. The two isoforms are structurally and functionally identical except for 25 additional amino acids at the N terminus of pGSN . pGSN circulates in the plasma of healthy humans and other mammals at average levels of 200 ± 50 mg/L. In every acute tissue injury setting examined, including toxic, hyperoxic, and idiopathic lung injury, adult respiratory distress syndrome, acute liver injury, myonecrosis, pancreatitis, trauma, burns, and bacterial and protozoal sepsis, pGSN levels are subnormal [5, 6, 7, 8, 9, 10, 11, 12, 13, 14].
The unifying explanation for low pGSN concentrations in acute inflammatory conditions is the exposure by injury to plasma of the GSN-binding ligand, actin, a major cellular constituent ordinarily separated from the extracellular environment by intact plasma membranes. In some but not all such cases of pGSN depletion, GSN-actin complexes are detectable in the circulation. pGSN together with Gc-globulin, another extracellular actin-binding protein, is proposed to function as an 'extracellular actin scavenger system' responsible for the removal of actin released during tissue injury . Actin exposed to the extracellular environment polymerizes into filaments (F-actin) that stimulate downstream inflammatory reactions . pGSN has the capacity to sever and depolymerize F-actin into monomeric subunits (G-actin) that are then sequestered by Gc-globulin  and cleared in the liver [18, 19]. Administration of pGSN to animals subjected to systemic inflammation can prolong survival and prevent complications of acute injury [12, 14, 20]. The beneficial effect of pGSN in these settings is unclear but may reside in its binding and/or inactivation of inflammatory mediators such as lysophosphatidic acid, amyloid β protein, diadenosine 5',5"'-P1,P3-triphosphate, endotoxin, and platelet-activating factor) [21, 22, 23, 24, 25, 26]. These findings suggest that pGSN is a broad-spectrum anti-inflammatory buffer and that local pGSN depletion by a shift of binding toward actin during actin exposure following injury allows mediators to promote appropriate defense and repair functions. Catastrophic or prolonged pGSN depletion, however, hypothetically accommodates dysfunctional and destructive actions of the mediators, leading to secondary organ damage and even death.
This set of events is theoretically also applicable to chronic inflammatory conditions in which cellular damage and mediator release occur, but no studies have hitherto examined pGSN levels in such states. Rheumatoid arthritis (RA) is a chronic autoimmune disease of unknown etiology that most prominently affects the synovial lining, resulting in a persistent and progressive diarthrodial joint inflammation and destruction. We report here that pGSN levels are diminished in the blood of RA patients and that analysis of synovial fluids (SFs) suggests that pGSN is consumed in the inflamed joint. Our findings suggest that the reason for the decreased pGSN levels is local exposure of actin to the extracellular environment in these joints.
Materials and methods
Clinical characteristics of patients with rheumatoid arthritis
Erosive RA (n = 47)
Non-erosive RA (n = 31)
61.2 ± 14.1
54.2 ± 7.4
Rheumatoid factor, +/-
Disease duration, years
12.1 ± 9.2
7.7 ± 7.4
Treated with DMARDs
42 ± 56.6
38 ± 39.6
Systemic inflammation, CRP >20 mg/L
White blood cell count, × 109/L
8.2 ± 2.8
7.34 ± 1.1
10.6 ± 17.7
13.0 ± 14.1
Collection and preparation of samples
SF was obtained from knee joints by arthrocentesis, aseptically aspirated, and transferred into tubes containing sodium citrate (0.129 mol/L, pH 7.4). Blood samples were simultaneously obtained from the antecubital vein and collected into sodium citrate anti-coagulant. Blood and SF samples were centrifuged at 800 g for 15 minutes. Supernatants were collected, separated into lots, and stored frozen at -70°C until use.
Measurements of plasma gelsolin concentrations in plasma and synovial fluid
pGSN was quantified functionally by its ability to promote the nucleation of actin filament assembly using a fluorometric assay as previously described . The assay is based on the principle that calcium-activated pGSN binds pyrene-labeled actin monomers to form a nucleus from which actin polymerizes in the pointed (slowest-growing) end direction. Pyrene-labeled actin fluoresces with higher intensity as a polymer than as a monomer. Pyrene actin was prepared by derivatizing actin with N-pyrenyliodoacetamide (Molecular Probes, now part of Invitrogen Corporation, Carlsbad, CA, USA) using the procedure of Kouyama and Mihashi , exchanging CaCl2 for MgCl2. Before use, pyrene actin was diluted in depolymerization buffer (buffer A: 0.5 mM ATP, 0.5 mM β-mercaptoethanol, 2 mM Tris, 0.2 mM CaCl2, pH 7.4) to 20 μM, stored 1 hour at 37°C to reach monomer equilibrium, and centrifuged at 250,000 g and 4°C for 30 minutes in an Optima™ TL Ultracentrifuge (Beckman Coulter, Inc., Fullerton, CA, USA) to pellet any remnant F-actin. The supernatant was withdrawn and stored in an ice water bath until use. Plasma or SF to be analyzed was diluted 1:5 in polymerization buffer (buffer B: 0.1 M KCl, 0.2 mM MgCl2, 1.5 mM CaCl2, 0.5 mM ATP, 10 mM Tris, 0.5 mM β-mercaptoethanol, pH 7.4). Pyrene-actin fluorescence was recorded using a spectrofluorometer (FluoroMax-2®; JobinYvon-Spex Instruments S.A., Inc, now HORIBA Jobin Yvon Inc, Edison, NJ, USA). Excitation and emission wavelengths were 366 and 386 nm, respectively. Pyrene actin was added to a final concentration of 1 μM in 280 μL of buffer B containing 0.4 μM phallacidin and 5 μL of diluted sample in 6 × 50 mm borosilicate glass culture tubes (Kimble, Glass Inc, Vineland, NJ, USA). Nucleation was monitored for 240 seconds in the fluorometer following a fast vortex. The linear slope of the fluorescence increase was calculated between 100 and 200 seconds. All of the samples were run in duplicates. Polymerization rate in each sample was converted to pGSN concentration by use of a standard curve of recombinant human pGSN (rhpGSN).
Measurements of interleukin-6 levels in synovial fluid
The levels of IL-6 in SF were determined by a bioassay with a cell clone B13.29, subclone B9, which is dependent on IL-6 for growth, as described previously . The samples were tested in 250-fold dilutions and compared with a standard curve obtained using human recombinant IL-6 (Genzyme, Kent, UK).
Measurements of albumin in plasma and synovial fluid
Albumin was measured by use of a kit (QuantiChrom™ BCG Albumin Assay Kit; BioAssay Systems, Hayward, CA, USA) according to the manufacturer's instructions.
Immunoblotting for gelsolin isoform in synovial fluid and gelsolin fragments in plasma
Platelet-poor plasmas or SFs were diluted 1:100 in 1× sample buffer (SB) (10% glycerol, 2% SDS, 62.5 mM Tris-HCl, 0.03% Bromphenol blue, 5% β-mercaptoethanol, pH 6.8) to detect GSN isoforms and 1:40 to document pGSN fragments, vortexed briefly, and incubated at 97°C for 5 minutes. Samples (10 μL for GSN isoform and 20 μL for GSN fragments) were run on 10% SDS-PAGE gels in a modified Laemmli system  and transferred to Immobilon P membranes (polyvinylidene difluoride [PVDF]) (0.45 μm) (Millipore Corporation, Billerica, MA, USA). Platelet lysate (2 × 108/mL, 5 μL) and rhpGSN served as negative and positive controls for pGSN, respectively. For determination of GSN isoform, a polyclonal antibody recognizing an epitope in the plasma extension of human pGSN was used (1:2,000, 2 hours, 22°C). The antibody was designed using a peptide from the plasma extension sequence and produced by Invitrogen Corporation in rabbit using a KLH (keyhole limpet hemocyanin) carrier. The antibody titer was checked at 4, 8, and 10 weeks, and the antibody was affinity-purified. For total GSN, the mouse monoclonal anti-GSN antibody 2c4 was used [31, 32] (1:2,500, 2 hours, 22°C). Secondary antibodies used were goat anti-rabbit IgG (H+L)-horseradish peroxidase (HRP) (1:5,000, 80 minutes, 22°C) and goat anti-mouse IgG (H+L)-HRP conjugate, respectively (1:3,300, 80 minutes, 22°C) (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Chemiluminescence detection was done using SuperSignal® West Pico Chemiluminescent Substrate for detection of HRP (Pierce, Rockford, IL, USA). HyBlot CL autoradiography film (Denville Scientific Inc., Metuchen, NJ, USA) was exposed to the membrane for 1 minute (isoform detection) or overnight (matrix metalloproteinase [MMP]-cleavage product detection). The film was developed using an M35A X-OMAT Processor (Eastman Kodak Company, Rochester, NY, USA).
The 2c4 antibody recognizes the C-terminal half of the GSN molecule and was used for detection of approximately 42- to 46-kDa fragments by immunoblotting as previously reported . To confirm that the 2c4 antibody recognizes pGSN cleaved by MMPs into fragments and that cleavage can occur in plasma, rhpGSN (115 nM) or dilute human plasma (approximately 115 nM pGSN) was incubated with catalytic domain of MMP-3 (cMMP-3) (230 nM) (Sigma-Aldrich, St. Louis, MO, USA) in 50 mM Tris-HCl, 150 mM NaCl, 5 mM CaCl2, and 0.5 mM ZnCl2, pH 7.5, for various time points. SDS-PAGE and Western blots were performed as described above.
Identification of actin in synovial fluid and plasma
SFs and plasmas were pre-cleared by centrifugation at 2,500 g for 5 minutes, diluted 1:20 and 1:10 in 1 × SB, respectively, and boiled for 10 minutes at 97°C. Samples (20 μL) were analyzed by 12% SDS-PAGE and transferred to PVDF membranes as described above. Actin was identified using a primary mouse monoclonal anti-β-actin antibody (Clone AC-15 Mouse Ascites Fluid, 1:1,000, 2 hours, 22°C) (Sigma-Aldrich) and a secondary (H+L)-HRP conjugated goat anti-mouse IgG (1:3,300, 80 minutes, 22°C) (Bio-Rad Laboratories, Inc.). Chemiluminescence detection was performed as described above, exposing the film to the membrane for less than 5 minutes. Quantification of actin in SFs and plasma was performed by densitometry using a human actin standard (actin protein, non-muscle) (Cytoskeleton, Inc., Denver, CO, USA) and Scion Image 1.62a software (Scion Corporation, Frederick, MD, USA).
Detection of gelsolin-actin complexes in synovial fluid by immunoprecipitation
Eleven SFs that were strongly positive for actin were centrifuged at 2,500 g to pellet any cellular debris. Fifty microliters of supernatant was withdrawn and diluted 1:8 in binding buffer (20 mM Tris, 100 mM NaCl, 1 mM CaCl2, 0.01% Tween 20, pH 7.4). Samples were pre-cleared by incubation with 20 μL of GammaBind Plus Sepharose (GE Healthcare, Little Chalfont, Buckinghamshire, UK) 50/50 bead slurry for 1 hour end-over-end at 4°C to minimize unspecific interactions with the beads. After centrifugation to pellet beads, pre-cleared supernatants were removed and incubated end-over-end for 1 hour at 22°C with affinity-purified mouse anti-GSN IgG (2c4, 5 μg/mL). An unspecific mouse IgG was used as a control (normal mouse control IgG, sc-2025) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) (5 μg/mL). Twenty microliters 50/50 bead slurry was added and incubation continued for 1 hour end-over-end. Beads were pelleted and washed four times in binding buffer before being resuspended in 50 μL of 1 × SB, boiled for 10 minutes, and analyzed by SDS-PAGE as described in the section above. Actin was identified using a rabbit polyclonal IgG (anti-actin N-terminal antibody produced in rabbit 1:750, 12 hours, 4°C) (Sigma-Aldrich) and a secondary (H+L)-HRP conjugated goat anti-rabbit IgG (1:2,000, 80 minutes, 22°C) (Bio-Rad Laboratories, Inc.). Chemiluminescence detection was performed as described above, exposing the film to the membrane for 2 minutes.
The levels of pGSN in the blood and SF samples were expressed as mean ± standard deviation and as median with interquartile range. Comparisons between the matched blood and SF samples were analyzed by paired t test. Comparison of pGSN and albumin levels was also performed between the patient blood samples and the healthy controls. For further comparison, patient material was stratified according to radiological findings (erosive RA versus non-erosive RA). For the evaluation of a possible influence of therapeutic interventions on pGSN levels, patients were stratified according to DMARD treatment (treated versus untreated). Differences in pGSN levels in the blood and SF between the groups were calculated separately employing the Mann-Whitney U test. Spearman correlation was used to determine the association between pGSN and albumin levels in blood and SF (GraphPad Prism software; GraphPad Software, Inc., San Diego, CA, USA). For all of the statistical evaluation of the results, P values of below 0.05 were considered statistically significant.
Plasma gelsolin levels are significantly lower in plasma of patients with rheumatoid arthritis compared with matched healthy controls
To assess the impact of systemic inflammation on circulating pGSN levels, patients were stratified according to CRP levels, where CRP of above 20 mg/L indicated the presence of systemic inflammation. The pGSN levels had a tendency to be lower in circulation (135.8 ± 32.2 versus 148.9 ± 29.9 mg/L, P = not significant) and in SF (92.2 ± 24.4 versus 96.3 ± 23.7 mg/L, P = not significant) of patients having systemic inflammation as compared with those without. To evaluate a relationship between intra-articular levels of pGSN and local inflammation, we measured levels of IL-6 in SFs. The mean IL-6 level was 1.59 ± 0.11 ng/mL (range of 0.03 to 4.52 ng/mL), indicating some degree of local inflammation in most of the samples. However, pGSN levels showed no correlation with IL-6 levels.
Characteristics of gelsolin present in plasma and synovial fluid of patients with rheumatoid arthritis
Actin and gelsolin-actin complexes are present in synovial fluids of rheumatoid arthritis patients
This is the first study to show that pGSN levels are reduced in plasma of patients with RA. The decrease is inversely related to the intensity of systemic inflammation, determined by an inverse correlation with levels of acute-phase protein (CRP). In contrast, local inflammation measured by IL-6 levels in SF has no direct correlation with pGSN levels. The finding that circulating pGSN levels decrease during chronic joint inflammation is consistent with observations of acute inflammatory disease states, such as sepsis and acute respiratory distress syndrome, in which a drop in pGSN concentration precedes more severe injuries [5, 6, 7, 8, 9, 10, 11, 12, 13, 14] and the beneficial action of pGSN during re-administration in traumatized animals suggests that it has a protective role in inflammation [12, 14, 20]. The mechanism behind this demonstrated protective effect during acute inflammation is unclear but may reside in its binding and inactivation of inflammatory mediators) [21, 22, 23, 24, 25, 26].
Furthermore, MMPs, considered the major enzymes responsible for the extracellular matrix degradation, cleave pGSN in vitro [33, 35] and are elevated during inflammation. It is possible that pGSN is cleaved by MMPs in vivo during RA and that this cleavage contributes to lower levels being detected. This is further supported by observations of MMP-mediated cleavage of 68-kDa secreted peptides from mutant (D187Y/N) pGSN [45, 46]. MMP-induced pGSN cleavage would most likely be interpreted as a decrease in pGSN concentration in our functional assay since the nucleating activity of pGSN cut in half is weaker than for the full-length protein [47, 48, 49]. Although we did not detect proteolytic pGSN fragments by immunoblotting, it is possible that they are further degraded or rapidly cleared from the circulation.
We have documented a decrease in levels of circulating pGSN in RA, the first example of such depletion during chronic inflammation. pGSN levels are even lower intra-articularly in affected joints. Although the cause of the decrease is unclear and enzymatic degradation or decreased production cannot be definitively ruled out, the capacity of pGSN to bind several factors present in the inflamed joint space and the lack of detection of proteolyic fragments suggest a local consumption. The proportion of pGSN levels in SF versus plasma, in combination with the observations that actin is exposed to the extracellular environment during RA and that GSN-actin complexes are present in SF, supports this hypothesis. However, the lack of a direct relation between the degree of local inflammation as measured by IL-6 and pGSN levels suggests that the regulation of pGSN in RA is complex. Further studies on the involvement of pGSN in RA might help in understanding the pathogenesis and possibly aid in diagnosis and future treatments.
The authors thank Po-shun Lee for constructing the human pGSN antibody, Eric Osborn for reviewing the manuscript, and Claes Dahlgren for mentorship. The work was supported by grants from the 80-Year Foundation of King Gustav V, the Swedish Association Against Rheumatism, the Swedish Research Council, the Swedish National Inflammation Network, the Nanna Svartz Foundation, the Gothenburg Medical Society, and University of Gothenburg. TMO was a recipient of the University of Gothenburg Jubileumsfond's stipend.
- 16.Erukhimov JA, Tang ZL, Johnson BA, Donahoe MP, Razzack JA, Gibson KF, Lee WM, Wasserloos KJ, Watkins SA, Pitt BR: Actin-containing sera from patients with adult respiratory distress syndrome are toxic to sheep pulmonary endothelial cells. Am J Respir Crit Care Med. 2000, 162: 288-294.CrossRefPubMedGoogle Scholar
- 27.Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS, Medsger TA, Mitchell DM, Neustadt DH, Pinals RS, Schaller JG, Sharp JT, Wilder RL, Hunder GG: The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum. 1988, 31: 315-324. 10.1002/art.1780310302.CrossRefPubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.