Screening And Clinical Validation Of Cytokine Markers Of Disease Activity And Extra-articular Manifestations In Rheumatoid Arthritis

Objective: Rheumatoid arthritis(RA)is an autoimmune disease characterized by progressive joint destruction,often manifested by symmetrical swelling and pain in multiple small joints.The etiology is still unknown.The assessment of RA is complex.The disease affects the joints,but also the lungs,blood system,kidney and other systems.One of the important reasons for the unsatisfactory therapeutic effect of RA at the present stage is that the activity/remission of the disease cannot be accurately assessed,and the organ involvement of RA cannot be accurately judged,which leads to the difficult control of the treatment timing of RA.Currently,the commonly used assessment of disease activity in clinical practice includes DAS28,CDAI,SDAI,Boolean remission,etc.Due to the inclusion of more patients or doctors’ subjective factors,there is an evaluation bias.Therefore,screening biomarkers that can evaluate the disease activity/remission and organ involvement of RA,and establishing more accurate and objective disease activity determination models are still a major focus of current research on RA.In this study,we first analyzed the demographic characteristics and common clinical examination results of 196 patients with RA admitted to Peking University People’s Hospital,and summarized the clinical and immunological characteristics of patients with different disease activity and organ involvement.Then,we detected the results of previous experiments in the discovery cohort and screened the cytokines that might be related to RA disease activity and extra-articular manifestations in the database,and further verified the screened cytokines related to RA disease activity in the validation cohort.And to evaluate the role of cytokines related to articular manifestations in patients with RA.Finally,an assessment model of RA disease activity was established based on existing common clinical indicators.Methods: Part I: Through retrospective and prospective collection of relatively complete clinical medical records of 196 inpatients and outpatients with RA,demographic characteristics and clinical laboratory test results of RA patients were recorded,and clinical characteristics of subgroups of RA patients combined with extrasarticular manifestations were analyzed.Part II: Discovery of 72 RA patients in the cohort,25 cytokines in serum of RA patients were detected by multifactor and ELISA:S100A8,IL-1beta/IL-1F2,IL-1ra/IL-1F3,TNF-α,IL-6,IL-8/CXCL8,IFN-β,CXCL13/BLC/BCA-1,IFN-γ,Dkk-1,IL-17/IL-17 A,RAGE/AGER,S100A9,SOST/Sclerostin,M-CSF,Leptin/OB,CCL23/MPIF-1,Progranulin/PGRN,IFN-α,IL-21,GM-CSF,CXCL10/IP-10/CRG-2,s CD25,LBP,SRA,The screened cytokines associated with RA disease activity were verified in the validation cohort.Part III: Evaluating the significance of the 25 cytokines tested in indicating extrarticular manifestations/systemic involvement of RA.Part IV: Verification of RA patients in the cohort,the predictive model of RA disease activity was established by combining the screened cytokines related to RA disease activity with existing common clinical indicators.Results: Part 1:RA was divided into 4 groups according to DAS28 score from high to low,and group 1had the highest DAS28 score.By comparing the clinical laboratory characteristics of RA patients in different disease activity groups,it was found that the number of joint swelling and pain,ESR,CRP,RF and CCP were correlated with the disease activity of RA.In addition to disease activity,we mainly analyzed risk factors and predictors for the articular manifestations of RA.The levels of RF,CCP,HRF-Ig G and GPI in RA patients with rheumatoid nodules were significantly higher than those in the group without extra-articular manifestations,and the elevated level of CCP was an independent risk factor.Similarly,independent risk factors for RA combined with pulmonary interstitial disease,kidney involvement,blood system involvement,co-infection,and metabolic disorders were analyzed at the same time,but no new clues were found that were different from the known traditional risk factors.Part II: Through literature review,we screened the factors that may be highly correlated with disease activity and extra-articular manifestations of RA from more than 1000 cytokines related literatures,plus 25 highly correlated factors found in the previous work of our research group.In this part,it was found that serum levels of IL-6,S100A9,IFN-α,s CD25 and SRA in RA patients were positively correlated with DAS28-ESR score of disease activity,while LBP was negatively correlated.This was further validated in the validation cohort of RA patients.The levels of SRA,LBP,IL-6,s CD25 and GM-CSF in RA patients in different disease activity groups(groups 1-4)were significantly decreased with the decrease of RA disease activity,and were positively correlated with DAS28 score.RAGE/AGER level increased significantly with the decrease of RA disease activity,and was negatively correlated with DAS28 score,which may be a protective factor.In ESR and CRP group,the positive rate of SRA,LBP and IL-6 was high(about 95.0%),while in ESR and CRP group,the positive rate of LBP was the highest 85.5%,and the positive rate of IL-6 and SRA was also high(75.0%).The results indicated that CRP was negative in ESR and CRP.In patients with high disease activity,LBP,IL-6 and SRA may be suggestive of disease activity.In RF and CCP group,the positive rate and absolute value of LBP,SRA and IL-6 were higher [98(93.3%),94(89.5%),89(84.9%)].In RA patients with single positive CCP and single positive RF,the positive rates of LBP and IL-6 were still the highest(both 100%).In both RF and CCP negative group,the positive rate of LBP was the highest(100.0%),while the positive rate of IL-6 was significantly decreased.These results suggest that LBP,SRA,and IL-6 are highly suggestive of disease activity,and LBP is highly suggestive of disease activity regardless of whether CCP antibody and RF are positive or negative,but IL-6 is highly suggestive of disease activity only in the antibody positive group.Part III: In terms of extra-articular manifestations of RA patients,IFN-β and SRA levels were significantly increased in patients with rheumatoid nodules,while Dkk-1 and Leptin/OB levels were significantly lower.The levels of IL-1beta/IL-1F2,IL-1ra/IL-1F3,IFN-γ,IL-17/IL-17 A and s CD25 in patients with pulmonary interstitial disease were significantly higher,while the levels of RAGE/AGER and Leptin/OB were significantly lower.Through further multivariate analysis,s CD25 was an independent risk factor for RA pulmonary interstitial lesions in each articular presentation group,and RAGE/AGER appeared to be a protective factor.IL-21 levels were significantly higher in the subgroup of patients with renal involvement.The levels of IL-6,CXCL13/BLC/BCA-1,M-CSF,Progranulin/PGRN and GM-CSF were significantly higher in the subgroup with blood system involvement.The level of S100A9 was significantly higher in the subgroup of co-infection.The level of DKK-1 was significantly higher in the subgroup of patients with metabolic disorders.Part Four:We verified the correlation between the above cytokines and RA disease activity in the validation cohort,and established a disease activity prediction model based on Logistic regression and BP neural network using factors independently associated with RA disease activity combined with clinical indicators and acute time-phase reactants(erythrocyte sedimentation and CRP).Logistic regression predicted that s CD25,M-CSF,ESR,low density lipoprotein and the number of swollen joints were finally included.The prediction accuracy of the models were 78.9% respectively.The maximum AUC of ESR,number of swollen joints and M-CSF was 0.897.BP neural network included the above 5parameters,and the accuracy of the model was 78.4%.AUC was 0.864 in the increased or decreased RA disease activity group,and the model was better.Conclusions: 1.The number of swollen and painful joints,ESR,CRP,RF and CCP are related to the disease activity of RA.2.The disease activity of RA was associated with the increase of various cytokines,including IL-6,S100A9,IFN-α,s CD25,SRA,and LBP.These cytokines have suggestive significance for disease activity in RA patients.3.s CD25 and RAGE/AGER were associated with lung interstitial lesions in RA patients,which may be predictive factors of ILD occurrence in RA.4.In RA patients with negative ESR and CRP and high disease activity,LBP,IL-6 and SRA may have suggestive significance for disease activity;5.A predictive model of disease activity in patients with RA was established based on Logistic regression and BP neural network,and five indexes including s CD25,M-CSF,ESR,low density lipoprotein,and number of swollen joints were included as evaluation indexes.This model may become an objective assessment model of disease activity in RA.