Effects of carrageenan induced synovitis on joint damage and pain in a rat model of knee osteoarthritis.

OBJECTIVE: Knee osteoarthritis (OA) is associated with ongoing pain and joint damage that can be punctuated by acute flares of pain and inflammation. Synovitis in normal knees might resolve without long-term detriment to joint function. We hypothesised that osteoarthritis is associated with impaired resilience to inflammatory flares. DESIGN: We induced synovitis by injecting carrageenan into rat knees with or without meniscal transection (MNX)-induced OA, and measured synovitis, weightbearing asymmetry (pain behaviour), and joint damage up to 35 days after OA induction (23 days after carrageenan-injection). RESULTS: Carrageenan injection induced weightbearing asymmetry for 1 week, transient increase in knee diameter for 2 days, and a sustained increase in synovial macrophages, endothelial cell proliferation and vascular density compared with naive vehicle-injected controls. MNX surgery induced weightbearing asymmetry and histological evidence of OA. Carrageenan-injection in MNX-operated knees was followed for 2 days by increased weightbearing asymmetry compared either to MNX+vehicle or to sham+carrageenan groups. OA structural damage and synovitis at day 35 were greater in MNX+carrageenan compared to MNX+vehicle and sham+carrageenan groups. Carrageenan injection did not induce OA in Sham-operated knees. CONCLUSION: Intra-articular injection of the pro-inflammatory compound carrageenan in OA and sham-operated control knees induced a short term increase in joint pain. Even though pain flares resolved in both groups and damage was not induced in sham-operated knees, carrageen injection exacerbated long-term joint damage in OA knees. OA knees display less resilience to inflammatory episodes. Preventing inflammatory flares may be particularly important in preventing symptoms and long term joint damage in OA.

Histopathological subgroups in knee osteoarthritis.

OBJECTIVE: Osteoarthritis (OA) is a heterogeneous, multi-tissue disease. We hypothesised that different histopathological features characterise different stages during knee OA progression, and that discrete subgroups can be defined based on validated measures of OA histopathological features. DESIGN: Medial tibial plateaux and synovium were from 343 post-mortem (PM) and 143 OA arthroplasty donations. A 'chondropathy/osteophyte' group (n = 217) was classified as PM cases with osteophytes or macroscopic medial tibiofemoral chondropathy lesions ≥grade 3 to represent pre-surgical (early) OA. 'Non-arthritic' controls (n = 48) were identified from the remaining PM cases. Mankin histopathological scores were subjected to Rasch analysis and supplemented with histopathological scores for subchondral bone marrow replacement and synovitis. Item weightings were derived by principle components analysis (PCA). Histopathological subgroups were sought using latent class analysis (LCA). RESULTS: Chondropathy, synovitis and osteochondral pathology were each associated with OA at arthroplasty, but each was also identified in some 'non-arthritic' controls. Tidemark breaching in the chondropathy/osteophyte group was greater than in non-arthritic controls. Three histopathological subgroups were identified, characterised as 'mild OA', or 'severe OA' with mild or moderate/severe synovitis. CONCLUSIONS: Presence and severity of synovitis helps define distinct histopathological OA subgroups. The absence of a discrete 'normal' subgroup indicates a pathological continuum between normality and OA status. Identifying specific pathological processes and their clinical correlates in OA subgroups has potential to accelerate the development of more effective therapies.

Relationship between structural pathology and pain behaviour in a model of osteoarthritis (OA).

OBJECTIVES: To address the hypothesis that different types of established osteoarthritis (OA) pain behaviours have associations with different aspects of articular pathology, we investigated the relationship between structural knee joint pathology and pain behaviour following injection of a low vs a high dose of monosodium iodoacetate (MIA) in the rat. METHODS: Rats received a single intra-articular injection of 0.1 mg or 1 mg MIA or saline (control). Pain behaviour (hind limb weight bearing asymmetry (WB) and hindpaw withdrawal threshold (PWT) to punctate stimulation) was assessed. Cartilage and synovium were examined by macroscopic visualisation of articular surfaces and histopathology. RESULTS: Both doses of MIA lowered PWTs, 1 mg MIA also resulted in WB asymmetry. Both doses were associated with cartilage macroscopic appearance, proteoglycan loss, abnormal chondrocyte morphology, increased numbers of vessels crossing the osteochondral junction, synovitis and macrophage infiltration into the synovium. PWTs were more strongly associated with chondrocyte morphology, synovitis and macrophage infiltration than with loss of cartilage surface integrity. CONCLUSIONS: Both pain behaviours were associated with OA structural severity and synovitis. Differences in pain phenotype following low vs higher dose of MIA were identified despite similar structural pathology. OA structural pathology as traditionally measured only partially explains the MIA-induced pain phenotype.

The anti-NGF antibody muMab 911 both prevents and reverses pain behaviour and subchondral osteoclast numbers in a rat model of osteoarthritis pain.

OBJECTIVE: Nerve growth factor (NGF) has a pivotal role in peripheral hyperalgesia and inflammation; anti-NGF antibodies attenuate pain responses in inflammatory pain models, and in people with osteoarthritis (OA) or low back pain. The aim of this study was to characterise the peripheral mechanisms contributing to the analgesic effects of anti-NGF antibody treatment in an established model of joint pain, which mimics key clinical features of OA. DESIGN: Effects of preventative vs therapeutic treatment with an anti-NGF antibody (monoclonal antibody 911: muMab 911 (10 mg/kg, s.c.)) on pain behaviour (weight bearing asymmetry and hindpaw withdrawal thresholds (PWT)), cartilage damage, synovitis and numbers of subchondral osteoclasts were investigated in the monosodium iodoacetate (MIA) model. Potential direct effects of NGF on receptor activator of nuclear factor kappa-B ligand (RANKL) mediated osteoclastogenesis were investigated in cultured human osteoclasts. RESULTS: Intra-articular MIA injection resulted in significant pain behaviour, cartilage damage, synovitis and increased numbers of subchondral osteoclasts. Both preventative and therapeutic treatment with muMab 911 significantly prevented, or reversed, MIA-induced pain behaviour, but did not alter cartilage or synovial pathology quantified at the end of the treatment period. NGF did not facilitate RANKL driven osteoclast differentiation in vitro, but preventative or therapeutic muMab 911 reduced numbers of TRAP positive osteoclasts in the subchondral bone. CONCLUSIONS: We demonstrate that anti-NGF antibody treatment attenuates OA pain behaviour despite permitting cartilage damage and synovitis. Indirect effects on subchondral bone remodelling may contribute to the analgesic effects of NGF blockade.

Design, study quality and evidence of analgesic efficacy in studies of drugs in models of OA pain: a systematic review and a meta-analysis.

OBJECTIVE: Studies using animal models are important in drug development, but often poorly predict treatment results in man. We investigated factors that may impact on the magnitude of the analgesic treatment effect in animal models of osteoarthritis (OA) pain. DESIGN: Systematic review of studies that measured behavioural pain outcomes in small animal models of OA, and tested drugs which reduce OA pain in man. Standardised mean difference (SMD) and 95% confidence intervals (CIs) were calculated using random effects meta-analysis for selected models and drugs. RESULTS: Most studies used rat models (42/50) and chemical methods of OA induction (39/50). Analgesic treatment effect (SMD) was most commonly measured between drug- and vehicle treated rats with knee OA. Meta-analysis was carried out for 102 such comparisons from 26 studies. The pooled SMD was 1.36 (95% CI = 1.15-1.57). Non-steroidal anti-inflammatory drugs (NSAIDs) were associated with smaller SMDs than opioids (z = -3.25, P = 0.001). Grip strength gave larger SMDs than assessment of static weight bearing (z = -4.60, P < 0.001), mechanically-evoked pain (z = -3.83, P = 0.001) and movement-evoked pain (z = -5.23, P < 0.001), and SMDs for mechanically-evoked pain were larger than for movement-evoked pain (z = -2.78, P = 0.006). Studies that reported structural evaluation of OA phenotype were associated with smaller SMDs (z = -2.45, P = 0.014). Publication was significantly biased towards positive findings. CONCLUSION: Attention to study-level moderators and publication bias may improve the ability of research using animal models to predict whether analgesic agents will reduce arthritis pain in man.

Differences in structural and pain phenotypes in the sodium monoiodoacetate and meniscal transection models of osteoarthritis.

OBJECTIVES: To characterize differences in joint pathology and pain behavior between two rat models of osteoarthritis (OA) in order to inform selection of animal models for interventional studies. METHOD: Knee OA was induced in Sprague Dawley rats by either meniscal transection (MNX) or intra-articular injection of monosodium iodoacetate (MIA). Controls were subjected to sham surgery or saline-injection. In a separate experiment, a single intra-articular injection of triamcinolone acetonide was administered 14 days after MNX or MIA arthritis induction. Pain behavior and joint pathology were quantified. RESULTS: Both models displayed synovial inflammation, chondropathy and osteophytosis. Chondropathy scores increased with time similarly in the two models. Inflammation and osteophyte scores were greater in MNX model compared to the MIA model. At day 49, the MNX model exhibited a greater number of channels crossing the osteochondral junction compared to all other groups. The MNX model exhibited greater weight bearing asymmetry compared to the MIA model, whereas the MIA model displayed more consistent hindpaw allodynia. Triamcinolone attenuated weight bearing asymmetry and distal allodynia to control levels in the MNX model, but distal allodynia was unaltered in the MIA model. CONCLUSIONS: The comparison of the two models of OA in rats, using identical assessment tools has demonstrated that although both models display features of OA, there are differences between the models which may represent different aspects of human OA. Thus, model selection should be based on the pathological aspects of OA under investigation.

Lymphatic vessels in osteoarthritic human knees.

OBJECTIVES: The distribution and function of lymphatic vessels in normal and diseased human knees are understood incompletely. This study aimed to investigate whether lymphatic density is associated with clinical, histological or radiographic parameters in osteoarthritis (OA). METHODS: Sections of synovium from 60 knees from patients with OA were compared with 60 post mortem control knees (from 37 individuals). Lymphatic vessels were identified using immunohistochemistry for podoplanin, and quantified as lymphatic vessel density (LVD) and lymphatic endothelial cell (LEC) fractional area. Effusion status was determined by clinical examination, radiographs were scored for OA changes, and inflammation grading used haematoxylin and eosin stained sections of synovium. RESULTS: Lymphatic vessels were present in synovia from both disease groups, but were not identified in subchondral bone. Synovial lymphatic densities were independent of radiological severity and age. Synovia from patients with OA displayed lower LVD (z=-3.4, P=0.001) and lower LEC fractional areas (z=-4.5, P<0.0005) than non-arthritic controls. In patients with OA, low LVD was associated with clinically detectable effusion (z=-2.2, P=0.027), but not with histological evidence of synovitis. The negative associations between lymphatics and OA/effusion appeared to be independent of other measured confounders. CONCLUSION: Lymphatic vessels are present in lower densities in OA synovia. Abnormalities of synovial fluid drainage may confound the value of effusion as a clinical sign of synovitis in OA.

Effects of a metalloproteinase inhibitor on osteochondral angiogenesis, chondropathy and pain behavior in a rat model of osteoarthritis.

OBJECTIVE: To investigate the effects of a matrix metalloproteinase (MMP) inhibitor on joint pathology and pain behavior in the rat meniscal transection (MNX) model of osteoarthritis (OA) and evaluate which aspects of structural disease modification contribute to symptom improvement. METHODS: OA pathology was induced in male Lewis rats, by transecting the medial collateral ligament with (MNX) or without (SHAM) a full thickness cut through the meniscus. MNX animals were orally administered an equipotent MMP 2, 8, 9, 12, 13 inhibitor (0.25, 1 and 5 mg/kg/day) or vehicle from day 1. Chondropathy, osteophytosis, osteochondral vascularity were assessed from toluidine blue stained coronal sections of the total knee joint and weight-bearing asymmetry by incapacitance. Group differences were evaluated using 1-way analysis of variance (ANOVA) and associations as Spearman's correlation coefficients. RESULTS: Treatment with the MMP inhibitor reduced weight-bearing asymmetry from day 14 onwards, and attenuated chondropathy (both P<0.05). Osteochondral vascularity was elevated in MNX compared with SHAM-operated animals (P<0.001) and reduced, dose-dependently, by MMP inhibitor treatment (r=-0.89, P<0.05). Reduced osteochondral vascularity and chondropathy were associated with the amelioration of weight-bearing asymmetry (both P<0.05). CONCLUSION: Here we show that treatment with a MMP inhibitor reduces joint damage, osteochondral angiogenesis and behavioral evidence of pain. The association between osteochondral angiogenesis and pain behavior may be explained by perivascular nerve growth or stimulation of subchondral nerves following loss of osteochondral integrity. Our data suggest that targeting angiogenesis may have utility in the treatment of pain associated with structural damage in OA.

Osteochondral angiogenesis and increased protease inhibitor expression in OA.

OBJECTIVE: Normal cartilage is resistant to vascular invasion and anti-angiogenic protease inhibitors may contribute to its avascular status. We hypothesized that dysregulated expression of four key anti-angiogenic protease inhibitors may contribute to increased osteochondral vascularity in osteoarthritis (OA). DESIGN: Medial tibial plateaux from OA patients (n=40) were compared with those from non-arthritic controls collected post-mortem (PM, n=10). Immunohistochemistry was performed for protease inhibitors TIMP-1, TIMP-3, PAI-1 and SLPI and the pro-angiogenic factor vascular endothelial growth factor (VEGF). Immunoreactivity in articular chondrocytes was scored. Chondropathy was measured as a modified Mankin score, and osteochondral vascular density as number of channels crossing each mm of tidemark. Non-parametric analyses were used for all data. RESULTS: All protease inhibitors and VEGF were localised to chondrocytes near the articular surface, less often in the middle zone, and rarely to deep chondrocytes. Scores for VEGF, TIMP-1, TIMP-3, SLPI and PAI-1 were all increased in OA compared with PM, and higher scores were associated with greater chondropathy. Chondrocyte expression of VEGF was associated with higher osteochondral vascular density (r=0.32, P<0.05), whereas protease inhibitors were not. CONCLUSIONS: The resistance of normal articular cartilage to vascular invasion may be more due to its matrix environment than ongoing protease inhibitor expression. Upregulation of protease inhibitors by superficial chondrocytes in OA may moderate the angiogenic effects of growth factors such as VEGF. However, failure of deep chondrocytes to express anti-angiogenic protease inhibitors may permit vascular invasion into the articular cartilage.

Contralateral hyperalgesia and allodynia following intradermal capsaicin injection in man.

OBJECTIVES: Contralateral responses to unilateral stimuli have been well described in animal models. These range from central sensitization to peripheral inflammatory responses. Our aim was to test for contralateral responses following unilateral intradermal capsaicin injection in man. METHODS: Three groups were investigated. A healthy volunteer group (1) was injected with capsaicin into the volar aspect of one forearm. A group of patients with RA (2) was also injected with capsaicin. A control group of healthy volunteers (3) was not injected with capsaicin. All groups were tested for hyperalgesia and allodynia every 10 min for 1 h following the injection using quantitative sensory testing. RESULTS: A total of 9/14 healthy volunteers (Group 1) and 10/14 patients with RA (Group 2) demonstrated contralateral sensitization that subsided within 1 h following intradermal capsaicin injection. A total of 2/23 control subjects (Group 3) demonstrated positive responses with the monofilaments. The frequency of the contralateral responses in the experimental groups compared with the control group is significant (P < 0.05). The peak hyperalgesia was relatively delayed contralaterally compared with the ipsilateral side (35 min vs 15 min). The area of sensitization, where present, was reduced compared with the ipsilateral side (5-50%). CONCLUSIONS: This is the first demonstration of a contralateral response following a unilateral stimulus in man. Bilateral neural pathways mediating contralateral responses may have a role in the pathophysiology of chronically painful or inflammatory diseases and a confounding influence on using the contralateral limb as a control experimentally. We did not find that a systemic inflammatory disease sensitized for this phenomenon.

Angiogenesis in two animal models of osteoarthritis.

OBJECTIVE: We have previously described angiogenesis at the osteochondral junction and in synovium of knees from patients with osteoarthritis (OA), but little is known about how closely animal models of OA resemble human disease with respect to vascular growth. This study aimed to characterise two animal models of knee OA with particular respect to osteochondral and synovial angiogenesis. METHOD: We examined the spontaneous Dunkin-Hartley (DH) guinea pig and medial meniscal transection (MNX) rat models of OA. Vessels at the osteochondral junction and in the synovium were identified by lectin immunohistochemistry and quantified by computer-assisted image analysis. Disease severity was assessed using a scoring system. RESULTS: Blood vessels crossed the osteochondral junction in juvenile rats and guinea pigs, with higher densities in the lateral than medial tibial plateau, the number decreasing with maturation in the absence of other OA changes. In the rat model, increased vascular density was observed both at the osteochondral junction and in the synovium, whilst osteochondral vascularity in control rats decreased with maturation, OA rats showed a persistence of blood vessels at the osteochondral junction. In rat synovium, blood vessel fractional area was increased in the hypertrophied synovium 14 days after surgery, then decreased to control levels by day 28. Significant differences in vascularity were not observed between affected (medial) and spared (lateral) compartments of guinea pig knees. CONCLUSION: The rat meniscal transection model of OA reproducibly displays both osteochondral and synovial angiogenesis comparable to our previous observations in human knee OA. DH guinea pigs, by contrast, display low vascularity throughout their protracted course of OA development. Changes in vascularisation occur early during the development of OA in the rat, and may contribute to the pathogenesis of OA.

Inactivation of xanthine oxidoreductase is associated with increased joint swelling and nitrotyrosine formation in acute antigen-induced arthritis.

OBJECTIVE: Arthritis is associated with increased articular formation of nitrotyrosine, which may contribute to injury. Nitrotyrosine is formed by nitration of tyrosine by reactive nitrogen species such as peroxynitrite, the formation of which may be enhanced by xanthine oxidoreductase (XOR), since it can generate nitric oxide from nitrite/nitrate, and superoxide during xanthine metabolism. We hypothesized that inactivation of XOR would protect against antigen-induced arthritis (AIA) and decrease nitrotyrosine formation. METHODS: AIA was induced with methylated bovine serum albumin (mBSA) in three groups of Wistar rats: animals fed on (1) tungsten-enriched chow (0.7 g/kg) (TG), which inactivates XOR, (2) standard chow (SG), and (3) rats treated with allopurinol (50 mg/kg/day; p.o.) (AG). Nitrotyrosine in patella-synovium was quantified by mass spectrometry three weeks after intra-articular (i.a.) antigen injection. RESULTS: Treatment with tungsten, but not allopurinol, suppressed plasma and articular XOR activity at < or = 0.9% of normal levels. XOR inactivation was associated with increased knee swelling 24-48 hrs post i.a. mBSA, compared with controls (mean increase +/- SEM of knee diameter from baseline of 3.3 +/- 0.5, 2.0 +/- 0.3 and 1.9 +/- 0.2 mm in TG, SG and AG (n = 14 each group), respectively; p < 0.05, TG vs SG, ANOVA). Mean ratio of articular nitrotyrosine-tyrosine (+/- SEM) was increased in the XOR-inactivated group, compared with controls: 12.3 +/- 0.7, 9.6 +/- 0.8 and 10.4 +/- 0.5 pg/microg in TG, SG and AG, respectively; p < 0.05, TG vs SG. CONCLUSION: Contrary to expectation, XOR inactivation was associated with increased joint swelling and articular tyrosine nitration in acute AIA, suggesting a novel, protective role for XOR in inflammatory arthritis.

Isolectin B4 binding neurons are not present in the rat knee joint.

Small-diameter sensory neurons are key contributors in joint pain and have been implicated in the pathogenesis of rheumatoid arthritis (RA). Small-diameter sensory neurons can be separated into at least two distinct populations, which include isolectin B4 (IB4)-binding and tyrosine receptor kinase (trk) A-expressing. While trkA-expressing neurons have been identified in the rat knee joint there are no data, we are aware of, to suggest that IB4-binding neurons are also present. We aimed to determine whether or not there exists a population of IB4-binding neurons in the rat knee joint. Retrograde nerve tracing with fluoro-gold (FG) was used to identify the complete population of knee joint afferents in the lumbar dorsal root ganglia (DRG) L3 and L4 of female Wistar rats. IB4 conjugated to fluorescein isothiocyanate (FITC) was used to identify the cell bodies of IB4-binding neurons in the DRG. Of 1096 FG-labeled cell bodies in the DRG of knee joint injected animals (n=4), none were double labeled with FITC. Injection of FG into skin over the medial aspect of the rat knee (n=3) showed 48% of these cutaneous afferents in L3 and L4 DRG were double-labeled with FG and FITC. A complete absence of IB4-binding neurons in the rat knee joint makes it unlikely that this predominantly cutaneous, IB4-binding population of afferent neurons could have any significant influence in chronic inflammatory joint disease. This suggests that trkA-expressing neurons are the sole population of small-diameter sensory neurons in the knee joint and implies a significant role for these afferents in the progression of RA.

Symmetry, T cells and neurogenic arthritis.

Symmetry in clinical disease occurs more commonly than expected by chance and is unexplained. In this paper we focus on symmetry in arthritis and describe the neurogenic hypothesis. Neuropeptides are anatomically relevant to systemic arthritis and have been shown to have modulating effects on both the immune and circulatory systems. Neural networks project bilaterally and are involved in the development and propagation of inflammatory disease. These putative pathological neuro-feedback loops may derive from the existence of biologically protective symmetrical mechanisms.

The transcription factors hypoxia-inducible factor 1alpha and Ets-1 colocalize in the hypoxic synovium of inflamed joints in adjuvant-induced arthritis.

OBJECTIVE: To determine the relationship between hypoxia and the expression of Ets-1 and hypoxia-inducible factor 1alpha (HIF-1alpha) in both normal and inflamed joints. Adjuvant-induced arthritis (AIA) was used as the model system, since it mirrors many aspects of the pathology of rheumatoid arthritis. METHODS: Adjuvant arthritis was induced in a group of 10 female Lewis rats. A second group of 10 uninjected female Lewis rats served as naive controls. When a maximum clinical joint score was achieved in the AIA group, all 20 rats were injected with the specific hypoxic cell marker Hypoxyprobe-1 and subsequently killed. Hypoxyprobe-1 adducts, Ets-1, and HIF-1alpha were localized in the joints of the hind feet from these groups using immunohistochemistry. RESULTS: Compared with the joints from control rats, inflamed joints contained markedly more cells with Hypoxyprobe-1 adduct immunoreactivity, Ets-1-immunoreactive nuclei, and nuclear immunoreactivity for both Ets-1 and HIF-1alpha. CONCLUSION: Our results demonstrate the presence of hypoxia in inflamed joints in this experimental model of arthritis. The colocalization of Ets-1 and HIF-1alpha in these hypoxic areas suggests that hypoxia may induce Ets-1 and HIF-1alpha expression during joint inflammation.

Localization of 3-nitrotyrosine to rheumatoid and normal synovium.

OBJECTIVE: To determine the localization of 3-nitrotyrosine (3-NT), a footprint marker of peroxynitrite (ONOO-) and other reactive nitrogen species, to the inflamed human synovium and to compare this with normal synovial and nonsynovial tissue of human and animal origin. METHODS: Monoclonal and polyclonal antibodies were used to investigate for 3-NT, inducible nitric oxide synthase (iNOS), macrophage marker CD68, and the vascular smooth muscle marker alpha-actin by avidin-biotin immunocytochemistry. RESULTS: In the inflamed synovium, 3-NT was found in the vascular smooth muscle and macrophages. In normal human synovium, 3-NT was present in the vascular smooth muscle and some lining cells and was not associated with immunoreactivity for iNOS. Similarly, 3-NT could be demonstrated in the vascular smooth muscle cells of normal rats and iNOS knockout mice. It was not present in the vascular smooth muscle of healthy, nonsynovial tissue. CONCLUSION: The synovial vasculature in histologically normal human and naive rodent synovium was alone among the normal tissues studied in exhibiting iNOS-independent immunoreactivity for 3-NT. These findings suggest a physiologic role for ONOO- in normal synovial vascular function.

Focally regulated endothelial proliferation and cell death in human synovium.

Angiogenesis and vascular insufficiency each may support the chronic synovial inflammation of rheumatoid arthritis. We have shown by quantitative immunohistochemistry and terminal uridyl deoxynucleotide nick end labeling that endothelial proliferation and cell death indices were each increased in synovia from patients with rheumatoid arthritis compared with osteoarthritic and noninflamed controls, whereas endothelial fractional areas did not differ significantly among disease groups. Markers of proliferation were associated with foci immunoreactive for vascular endothelial growth factor and integrin alpha(v)beta3, whereas cell death was observed in foci in which immunoreactivities for these factors were weak or absent. No association was found with thrombospondin immunoreactivity. The balance between angiogenesis and vascular regression in rheumatoid synovitis may be determined by the focal expression of angiogenic and endothelial survival factors. Increased endothelial cell turnover may contribute to microvascular dysfunction and thereby facilitate persistent synovitis.