Search

Animal models for inflammatory arthritides such as rheumatoid arthritis (RA) and psoriatic arthritis are widely accepted and frequently used to identify pathological mechanisms and validate novel therapeutic strategies. Unfortunately, many publications reporting on these animal studies lack detailed description and appropriate assessment of the distinct histopathological features of arthritis: joint inflammation, cartilage damage and bone erosion. Therefore, the European consortium BeTheCure, consisting of 38 academic and industrial partners from 15 countries, set as goal to standardise the histological evaluation of joint sections from animal models of inflammatory arthritis. The consensual approach of a task force including 16 academic and industrial scientists as well as laboratory technicians has resulted in the development of the Standardised Microscopic Arthritis Scoring of Histological sections (‘SMASH’) recommendations for a standardised processing and microscopic scoring of the characteristic histopathological features of arthritis, exemplified by four different rodent models for arthritis: murine collagen-induced arthritis, collagen–antibody-induced arthritis, human tumour necrosis factor transgenic Tg197 mice and rat pristane-induced arthritis, applicable to any other inflammatory arthritis model. Through standardisation, the SMASH recommendations are designed to improve and maximise the information derived from in vivo arthritis experiments and to promote reproducibility and transparent reporting on such studies. In this manuscript, we will discuss and provide recommendations for analysis of histological joint sections: identification of the regions of interest, sample preparation, staining procedures and quantitative scoring methods. In conclusion, awareness of the different features of the arthritis pathology in animal models of inflammatory arthritis is of utmost importance for reliable research outcome, and the standardised histological processing and scoring methods in these SMASH recommendations will help increase uniformity and reproducibility in preclinical research on inflammatory arthritis.

Background New medications for Rheumatoid Arthritis (RA) have emerged in the last decades, including Disease Modifying Antirheumatic Drugs (DMARDs) and biologics. However, there is no known cure, since a significant proportion of patients remain or become non-responders to current therapies. The development of new mode-of-action treatment schemes involving combination therapies could prove successful for the treatment of a greater number of RA patients. Methods We investigated the effect of the Tyrosine Kinase inhibitors (TKIs) dasatinib and bosutinib, on the human TNF-dependent Tg197 arthritis mouse model. The inhibitors were administered either as a monotherapy or in combination with a subtherapeutic dose of anti-hTNF biologics and their therapeutic effect was assessed clinically, histopathologically as well as via gene expression analysis and was compared to that of an efficient TNF monotherapy. Results Dasatinib and, to a lesser extent, bosutinib inhibited the production of TNF and proinflammatory chemokines from arthritogenic synovial fibroblasts. Dasatinib, but not bosutinib, also ameliorated significantly and in a dose-dependent manner both the clinical and histopathological signs of Tg197 arthritis. Combination of dasatinib with a subtherapeutic dose of anti-hTNF biologic agents, resulted in a synergistic inhibitory effect abolishing all arthritis symptoms. Gene expression analysis of whole joint tissue of Tg197 mice revealed that the combination of dasatinib with a low subtherapeutic dose of Infliximab most efficiently restores the pathogenic gene expression profile to that of the healthy state compared to either treatment administered as a monotherapy. Conclusion Our findings show that dasatinib exhibits a therapeutic effect in TNF-driven arthritis and can act in synergy with a subtherapeutic anti-hTNF dose to effectively treat the clinical and histopathological signs of the pathology. The combination of dasatinib and anti-hTNF exhibits a distinct mode of action in restoring the arthritogenic gene signature to that of a healthy profile. Potential clinical applications of combination therapies with kinase inhibitors and anti-TNF agents may provide an interesting alternative to high-dose anti-hTNF monotherapy and increase the number of patients responding to treatment.

Background The transmembrane-TNF transgenic mouse, TgA86, has been shown to develop spontaneously peripheral arthritis with signs of axial involvement. To assess similarity to human spondyloarthritis, we performed detailed characterization of the axial, peripheral, and comorbid pathologies of this model. Methods TgA86 bone pathologies were assessed at different ages using CT imaging of the spine, tail vertebrae, and hind limbs and characterized in detail by histopathological and immunohistochemical analysis. Cardiac function was examined by echocardiography and electrocardiography and bone structural parameters by μCT analysis. The response of TgA86 mice to either early or late anti-TNF treatment was evaluated clinically, histopathologically, and by μCT analysis. Results TgA86 mice developed with 100% penetrance spontaneous axial and peripheral pathology which progressed with time and manifested as reduced body weight and body length, kyphosis, tail bendings, as well as swollen and distorted hind joints. Whole-body CT analysis at advanced ages revealed bone erosions of sacral and caudal vertebrae as well as of sacroiliac joints and hind limbs and, also, new ectopic bone formation and eventually vertebral fusion. The pathology of these mice highly resembled that of SpA patients, as it evolved through an early inflammatory phase, evident as enthesitis and synovitis in the affected joints, characterized by mesenchymal cell accumulation, and neutrophilic infiltration. Subsequently, regression of inflammation was accompanied by ectopic bone formation, leading to ankylosis. In addition, both systemic bone loss and comorbid heart valve pathology were evident. Importantly, early anti-TNF treatment, similar to clinical treatment protocols, significantly reduced the inflammatory phase of both the axial and peripheral pathology of TgA86 mice. Conclusions The TgA86 mice develop a spontaneous peripheral and axial biphasic pathology accompanied by comorbid heart valvular dysfunction and osteoporosis, overall reproducing the progression of pathognomonic features of human spondyloarthritis. Therefore, the TgA86 mouse represents a valuable model for deciphering the role of transmembrane TNF in the pathogenic mechanisms of spondyloarthritis and for assessing the efficacy of human therapeutics targeting different phases of the disease.

Background Spondyloarthritis (SpA) is a complex disease characterised by chronic inflammation, bone erosion and pathological new bone formation. The TgA86 transmembrane TNF (tmTNF) transgenic mouse is a unique model of SpA, developing spontaneously and with 100% incidence early progressive SpA characterised by peripheral inflammatory arthritis and axial ankylosing spondylitis with cardiovascular involvement. This closely recapitulates the pathological findings and comorbid conditions described in human patients. Objectives To characterise in greater detail the development and progression of theTgA86 pathology and its similarities to human disease and to standardise reliable preclinical protocols and specific readouts for the assessment of the efficacy of human therapeutics. Methods TgA86 peripheral and axial pathology was assessed at different time points from 2.5 to 28 weeks of age. Disease severity was evaluated using clinical parameters and histopathological analysis of ankle and sacroiliac joints, lumbar and caudal vertebrae, as well as whole mount skeletal staining. Clinical and histopathological readouts were used to assess the therapeutic effect of Etanercept that was administered thrice weekly at 30 mg/Kg starting either from 2.5–5 weeks of age (prophylactic protocol) or from 9 weeks of age (therapeutic protocol). Results Clinical pathology in TgA86 mice appears already from 2.5 weeks of age, with signs of paw swelling, digit deformation and tail crinkling, while by 9 weeks of age pathology is fully established, with severe peripheral arthritis and tail and spine ankylosis. Pathology progression was also evident histopathologically, characterised by the originally described features of progressive inflammation, cartilage destruction and bone erosion observed in sacroiliac and ankle joints as well as in lumbar and caudal vertebra. Additionally, new pathology features were detected by identifying signs of enthesitis, new bone formation appearing as cartilaginous structures at the edges of vertebrae endplates, presence of red bone marrow during all stages of disease progression as well as signs of intervertebral disc (IVD) degeneration. Prophylactic treatment with Etanercept ameliorated effectively all clinical and histopathological features of the peripheral and axial pathology. Therapeutic treatment while affecting only minimally the clinical signs of both peripheral and axial pathology, it was found to reduce the peripheral arthritis histopathological score by at least 50%. Finally, treatment with Etanercept was also efficient in ameliorating the comorbid heart valve pathology observed in these animals. Conclusions We have shown that TgA86 pathology includes features of sacroiliitis, enthesitis, new bone formation, persisting red bone marrow and intervertebral disk degeneration, further strengthening the similarities of this model to human pathology. Based on the assessment of all pathology features during prophylactic anti-TNF treatment we suggest that early on in disease there may be a therapeutic window during which optimal treatment of the pathology can be achieved. Reference [1] Alexopoulou L, Pasparakis M, Kollias G. A murine transmembrane tumor necrosis factor (TNF) transgene induces arthritis by cooperative p55/p75 TNF receptor signaling. Eur J Immunol1997;27(10):2588–92. Disclosure of Interest None declared

Mesenchymal TNF signaling is etiopathogenic for inflammatory diseases such as rheumatoid arthritis and spondyloarthritis (SpA). The role of Tnfr1 in arthritis has been documented; however, Tnfr2 functions are unknown. Here, we investigate the mesenchymal-specific role of Tnfr2 in the TnfΔARE mouse model of SpA in arthritis and heart valve stenosis comorbidity by cell-specific, Col6a1-cre-driven gene targeting. We find that TNF/Tnfr2 signaling in resident synovial fibroblasts (SFs) and valvular interstitial cells (VICs) is detrimental for both pathologies, pointing to common cellular mechanisms. In contrast, systemic Tnfr2 provides protective signaling, since its complete deletion leads to severe deterioration of both pathologies. SFs and VICs lacking Tnfr2 fail to acquire pathogenic activated phenotypes and display increased expression of antiinflammatory cytokines associated with decreased Akt signaling. Comparative RNA sequencing experiments showed that the majority of the deregulated pathways in TnfΔARE mesenchymal-origin SFs and VICs, including proliferation, inflammation, migration, and disease-specific genes, are regulated by Tnfr2; thus, in its absence, they are maintained in a quiescent nonpathogenic state. Our data indicate a pleiotropy of Tnfr2 functions, with mesenchymal Tnfr2 driving cell activation and arthritis/valve stenosis pathogenesis only in the presence of systemic Tnfr2, whereas nonmesenchymal Tnfr2 overcomes this function, providing protective signals and, thus, containing both pathologies.

Objectives Rheumatoid arthritis (RA) is a chronic condition characterised by prolonged inflammation of the joints leading to bone and cartilage destruction. The key molecular and cellular regulators of RA pathology are the Tumour Necrosis Factor (TNF) and the mesenchymal-origin Synovial Fibroblasts (SFs), respectively. Apart from the joint pathology, RA patients also show higher morbidity rates, mainly due to the development of extraarticular conditions including cardiovascular, gut and skin disease manifestations. The Tg197 arthritis model develops TNF-driven and mesenchymal synovial fibroblasts (SFs)-dependent polyarthritis. Here, we investigate whether this model develops, similarly to human patients, co-morbid heart pathology and we explore cellular and molecular mechanisms linking arthritis to cardiovascular comorbidities. Methods The established human TNF-transgenic model of arthritis (Tg197) was used to evaluate possible arthritis-related cardiovascular disease. We further studied the role of Valve Interstitial Cells (VICs) in the pathology, using the ColVI-Cre mouse which specifically targets mesenchymal cells. Tg197 ColVI-Cre Tnfr1fl/fl and Tg197 ColVI-Cre Tnfr1cneo/cneo mice were used to explore the role of mesenchymal TNF signalling in the developing heart valve disease. Similarities of pathogenic VICs and SFs were analysed by RNA-sequencing. Results Tg197 mice develop left-sided heart valve disease, characterised by valvular fibrosis with minimal signs of inflammatory cell infiltration and thickened areas, consisting almost entirely of ColVI-expressing mesenchymal VICs, indicating proliferation of this cell type as a hallmark of the observed phenotype. Development of the pathology results in valve stenosis and left ventricular dysfunction, accompanied by arrhythmic episodes and, occasionally, valvular regurgitation. TNF-dependency of the pathology was indicated by the disease amelioration following pharmacological inhibition or genetic ablation of TNF-signalling. Interestingly, Tg197-derived VICs exhibited an activated phenotype ex vivo, resembling the activation of pathogenic Tg197-derived SFs. A significant functional correlation between these two mesenchymal cells was further supported by RNA-seq analysis, suggesting common pathogenic cellular mechanisms operating in arthritis and cardiovascular comorbidities. Conclusions TNF-mediated joint pathologies, commonly associated with comorbid heart valve disease, are efficiently modelled in Tg197 arthritis model and are commonly underlined by mesenchymal cell-specific pathogenic mechanisms. Disclosure of interest None declared

Objectives Patients with rheumatoid arthritis and spondyloarthritisshow higher mortality rates, mainly caused by cardiac comorbidities. The TghuTNF (Tg197) arthritis model develops tumour necrosis factor (TNF)-driven and mesenchymalsynovial fibroblast (SF)-dependent polyarthritis. Here, we investigate whether this model develops, similarly to human patients, comorbid heart pathology and explore cellular and molecular mechanisms linking arthritis to cardiac comorbidities. Methods Histopathological analysis and echocardiographic evaluation of cardiac function were performed in the Tg197 model. Valve interstitial cells (VICs) were targeted by mice carrying the ColVI-Cre transgene. Tg197 ColVI-Cre Tnfr1 fl/fl and Tg197 ColVI-Cre Tnfr1 cneo/cneo mutant mice were used to explore the role of mesenchymal TNF signalling in the development of heart valve disease. Pathogenic VICs and SFs were further analysed by comparative RNA-sequencing analysis. Results Tg197 mice develop left-sided heart valve disease, characterised by valvular fibrosis with minimal signs of inflammation. Thickened valve areas consist almost entirely of hyperproliferative ColVI -expressing mesenchymal VICs. Development of pathology results in valve stenosis and left ventricular dysfunction, accompanied by arrhythmic episodes and, occasionally, valvular regurgitation. TNF dependency of the pathology was indicated by disease modulation following pharmacological inhibition or mesenchymal-specific genetic ablation or activation of TNF/TNFR1 signalling. Tg197-derived VICs exhibited an activated phenotype ex viv o, reminiscent of the activated pathogenic phenotype of Tg197-derived SFs. Significant functional similarities between SFs and VICs were revealed by RNA-seq analysis, demonstrating common cellular mechanisms underlying TNF-mediated arthritides and cardiac comorbidities. Conclusions Comorbidheart valve disease and chronic polyarthritis are efficiently modelled in the Tg197 arthritis model and share common TNF/TNFR1-mediated, mesenchymal cell-specific aetiopathogenic mechanisms.

Background Spondyloarthropathies are a group of joint-related diseases with burden extending to multiple organs. Co-existing conditions have a major impact on the course and outcome of the disease complicating its diagnosis and management. Patients with spondyloarthropathies have two or more comorbidities including inflammatory bowel disease, psoriasis, cardiovascular disease, uveitis, and osteoporosis. In studying these pathologies and for the evaluation of novel therapeutic approaches, it is critical to carefully choose the animal model that closely reflects the complexity of the respective human pathology. We now use the identification of key cellular and molecular drivers. We have now revisited the TNFΔARE mouse, a well-established model of spondyloarthropathies, to investigate the extent of comorbid pathologies and identify the molecular and cellular pathways underlying pathogenesis in this animal model. Materials and methods The TNFΔARE/+ mouse model expresses de-regulated TNF resulting in the spontaneous development of chronic inflammatory joint and intestinal disease that are characteristic of spondyloarthropathies. The translational value of this mouse model has been established over the years through comprehensive studies on the pathogenic role of TNF as well as through the identification of key cellular and molecular drivers. We have now revisited the TNFΔARE/+ mouse model to uncover previously unidentified co-developing pathologies. To this end, we have carried out a targeted phenotypic screening of TNFΔARE/+ animals and a broad systematic phenotyping of these mice at the German Mouse Clinic. Results Targeted phenotyping revealed that TNFΔARE/+ mice, in addition to arthritis and Crohn’s-like IBD, develop aortic valve disease, salivary gland inflammation and inflammation of the periodontium. Systematic phenotyping revealed deviations in the behaviour, neurology, eye morphology and function, nociception, clinical chemistry and allergic responses of these animals along with pathological findings in kidneys and lungs. Some of these pathologies appear to share common TNF-dependent molecular and cellular mechanisms with arthritis and IBD, supporting a common aetiopathogenic axis. Conclusions This work further highlights the translational value of TNFΔARE/+ model as it reveals an even higher degree of complexity simulating the RA multisystemic condition that is observed in human patients. This model offers an ideal platform for the study of co-developing pathologies as well as the evaluation of novel therapeutics targeting multiple manifestations of spondyloarthropathies-related comorbidities.

Introduction and objectives Rheumatoid arthritis (RA) is a chronic condition characterised by inflammation of the joints as well as destruction of bone and cartilage. Studies on TNF transgenic models of polyarthritis established Tumour Necrosis Factor (TNF) targeting the mesenchymal-origin Synovial Fibroblasts (SFs) as a key event instigating the pathology. RA patients often show higher mortality rates, mainly due to the development of extraarticular disease including cardiovascular, gut and skin manifestations. The Tg197 and TnfΔARE/+ mouse models overexpress TNF and develop spontaneous chronic polyarthritis, fully mimicking human RA pathology. Here, we investigate whether these models develop, similarly to human patients, co-morbid heart pathology. Materials and methods We used the Tg197 and TnfΔARE/+ mouse models to evaluate possible arthritis-related cardiovascular disease. For further ex vivo analysis, we isolated Valve Interstitial Cells (VICs), which are the mesenchymal-origin fibroblasts constituting the aortic valve. We used a GFP reporter mouse labelling specifically mesenchymal-origin cells (ColVICre;mTm/mGFP) to target VICs. Similarities of pathogenic VICs and SFs were analysed by comparing their expression profiling with RNA sequencing analysis. Results Both arthritis models examined develop TNF-dependent aortic valve thickening, as indicated by the amelioration of the pathology following treatment with anti-TNF biologics. Aortic valves exhibited significant fibrosis with minimal signs of inflammatory cell infiltration and thickened areas, consisting almost entirely of VICs, indicating proliferation of this cell type as a hallmark of the observed phenotype. Isolated VICs from mutant mice exhibited a more proliferative and migratory phenotype and expressed high levels of TNF. Interestingly, VIC activation resembles the activated phenotype of pathogenic SFs isolated from the same mice. A significant functional correlation between these two pathogenic cells of mesenchymal origin was also supported by RNA-seq analysis, suggesting common cellular mechanisms operating in RA and RA-related heart pathology. Conclusion TNF-driven arthritis models, apart from their arthritic symptoms, develop co-morbid heart valve disease, similarly to reported comorbidities in RA patients. These two co-morbid diseases are shown here to share common mesenchymal-cell driven aetiopathogenesis.

Background Spondyloarthritis (SpA) describes a group of diseases characterised by a paradoxical simultaneous bone destruction and formation manifested as axial and peripheral pathologies. TNF is a key pathogenic factor with the strongest evidence of its leading role arising from the therapeutic effect of its inhibition in SpA patients. The TgA86 is a mouse model that overexpresses TNF and exhibits signs of spontaneous SpA. Our objective was to characterise the TgA86 pathology, identify its similarity to human disease and ultimately explore its potential to be used as the basis of a preclinical platform for the evaluation of SpA therapeutics and diagnostics. Materials and methods TgA86 transgenic mice express deregulated mouse transmembrane TNF and develop with 100% incidence chronic inflammatory arthritis and axial pathology clinically manifested by a characteristic tail bending. We used clinical observations, histopathological analysis, x-ray and μCT radiographic read-outs to establish a standardised pipeline for the evaluation of the TgA86 pathology. Results Histopathological, x-ray and μCT analysis of the vertebral column revealed the presence of pathological findings closely resembling those observed in human SpA patients. More specifically, disease characteristics included soft tissue swelling, bone erosion, increased bone formation and replacement of the vertebra bar-bell shape with a rectangular shape. Moreover, TgA86 mice exhibited a concomitant development of aortic valve disease, an extraarticular co-morbidity observed in a significant percentage of SpA patients. This data support the similarity of the TgA86 pathology to that of the human disease. We further used a standardised set of the above-mentioned read-outs to validate the response of this model to human therapeutics. TgA86 animals treated with the commercially available human therapeutic Etanercept, exhibited significant amelioration of both peripheral and axial pathologies in a reproducible and dose-dependent manner thus establishing the translational value of this model in the evaluation of human therapeutics. Conclusion Biomedcode has standardised and validated a novel TgA86-based highly reproducible and sensitive efficacy preclinical platform for the evaluation of therapeutics targeting SpA. This platform can be further used to validate and evaluate the therapeutic effect of the blockade of additional pathological pathways associated with SpA including IL-17, Wnt signalling and others.

Objective To identify novel microRNA (miR) associations in synovial fibroblasts (SF), by performing miR expression profiling on cells isolated from the human tumour necrosis factor (TNF) transgenic mouse model (TghuTNF, Tg197) and patients biopsies. Methods miR expression in SF from TghuTNF and wild-type (WT) control mice were determined by miR deep sequencing (miR-seq) and the arthritic profile was established by pairwise comparisons. Quantitative PCR analysis was utilised for profile validation, miR and gene quantitation in patient SF. Dysregulated miR target genes and pathways were predicted via bioinformatic algorithms and validated using gain-of-function coupled with reporter assay experiments. Results miR-seq demonstrated that TghuTNF-SF exhibit a distinct pathogenic profile with 22 significantly upregulated and 30 significantly downregulated miR. Validation assays confirmed the dysregulation of miR-223, miR-146a and miR-155 previously associated with human rheumatoid arthritis (RA) pathology, as well as that of miR-221/222 and miR-323-3p. Notably, the latter were also found significantly upregulated in patient RA SF, suggesting for the first time their association with RA pathology. Bioinformatic analysis suggested Wnt/cadherin signalling as a putative pathway target. miR-323-3p overexpression was shown to enhance Wnt pathway activation and decrease the levels of its predicted target β-transducin repeat containing, an inhibitor of β-catenin. Conclusions Using miR-seq-based profiling in SF from the TghuTNF mouse model and validations in RA patient biopsies, the authors identified miR-221/222 and miR-323-3p as novel dysregulated miR in RA SF. Furthermore, the authors show that miR-323-3p is a positive regulator of WNT/cadherin signalling in RA SF suggesting its potential pathogenic involvement and future use as a therapeutic target in RA.

A major stumbling block for research on and treatment of type 1 diabetes is the inability to directly, but noninvasively, visualize the lymphocytic/inflammatory lesions in the pancreatic islets. One potential approach to surmounting this impediment is to exploit MRI of magnetic nanoparticles (MNP) to visualize changes in the microvasculature that invariably accompany inflammation. MNP-MRI did indeed detect vascular leakage in association with insulitis in murine models of type 1 diabetes, permitting noninvasive visualization of the inflammatory lesions in vivo in real time. We demonstrate, in proof-of-principle experiments, that this strategy allows one to predict, within 3 days of completing treatment with an anti-CD3 monoclonal antibody, which NOD mice with recent-onset diabetes are responding to therapy and may eventually be cured. Importantly, an essentially identical MNP-MRI strategy has previously been used with great success to image lymph node metastases in prostate cancer patients. This success strongly argues for rapid translation of these preclinical observations to prediction and/or stratification of type 1 diabetes and treatment of individuals with the disease; this would provide a crucially needed early predictor of response to therapy.

Background Novel molecules that specifically target human TNFα in rheumatoid arthritis pose problems for preclinical assessment of efficacy. In this study collagen antibody-induced arthritis (CAIA) has been induced in human TNFα transgenic mice to provide a novel model that has been optimised for the evaluation of molecules targeting human TNFα. Methods Tg1278TNFko mice lack murine TNFα and are heterozygous for multiple copies of the human TNFα transgene that is expressed under normal physiological control. To establish CAIA, a collagen II monoclonal antibody cocktail (CAb) at 2, 4 or 8 mg was injected i.p. on Day 0 followed by a lipopolysaccharide (LPS) boost (10 or 100 μg) i.p. on Day 1 or Day 4. Animals were assessed for arthritis symptoms using a clinical score, cytokine levels (human TNFα, IL-1β and IL-6) in sera and joints, and histopathology. The dependence of the model on human TNFα was determined by dosing animals with etanercept. Results Tg1278TNFko animals treated with 2, 4 or 8 mg CAb on Day 0, with 100μg LPS on Day 4, had more severe arthritis and earlier symptoms than wild type animals at all doses of CAb tested. Subsequently it was found that the transgenic model did not require LPS at all for arthritis development but a lower dose of LPS (10 μg) was found necessary for reproducible and robust disease (close to 100% incidence, well-synchronised, with high arthritis scores). Furthermore the LPS challenge could be brought forward to Day 1 so that its’ actions to facilitate disease could be separated temporally from the arthritis phase (beginning about Day 4). Etanercept, administered immediately after the serum spike of cytokines associated with LPS had subsided, was able to dose-dependently inhibit arthritis development and this was associated with a marked protection of the joints histologically on Day 14. Etanercept was also able to reverse the signs of arthritis when given therapeutically allowing animals to be matched for disease burden before dosing begins. Conclusions The features of CAIA in Tg1278TNFko animals make the model well-suited to testing the next generation of therapeutics that will target human TNFα in rheumatoid arthritis.