Different immunophenotypes and cytokine profiles in non-systemic Juvenile Idiopathic Arthritis (JIA): potential fingerprints for predicting disease course and flare and driving therapeutic decisions.

Background: Juvenile Idiopathic Arthritis (JIA) is the most common chronic rheumatic disease in childhood and its prognosis has dramatically improved thanks to the novel treatments and the step forward performed in understanding the pathophysiology of the disease. However, contrasting data are available regarding clinical and laboratory predictors of response to treatment and relapse after drug withdrawal. Objectives: The purpose of the present study were to identify: 1. In a multicentric retrospective study cohort of JIA treated with the first course of Adalimumab or Etanercept who discontinued the treatment for persistent inactivity, factors associated with relapse. 2. In a monocentric prospective comparative study peculiar subsets of lymphocytes and cytokines/chemokines profile able to predict which patients might benefit of a more aggressive treatment, the disease status at the moment of blood sample collection and relapse after drug withdrawal Methods: For Aim 1 we performed a multicentric Italian retrospective cohort study based on medical records reviews of patients with oligoarticular and polyarticular JIA if they stopped therapy for persistent inactivity after the first anti-TNF. As main outcomes we considered the relapse after drug withdrawal. For Aim 2 we performed a prospective comparative monocentric study involving children with a diagnosis of JIA who signed the informed consent and underwent blood sample collection. For both aims we collected demographic, clinical, laboratory, treatment data. From the laboratory perspective we performed lymphocyte subpopulation assessments evaluating cytofluorimetry lymphocyte CD3+, CD3+CD4+, CD3+CD8+, CD19+ and CD16+CD56+. We performed the serum cytokine and chemokine assay using Luminex Multiplex assay for the following cytokines/chemokines: CCL2/JE/MCP1, CCL3/MIP1α, CD25/IL2R α, CXCL6/GCP2, CXCL9/ MIG, CXCL10/IP10/CRG2, G-CSF, IFNγ, IL1 α/IL1F1, IL1β/IL1F2, IL1ra/IL1F3, IL2, IL4, IL6, IL8/CXCL8, IL10, IL12 P70, IL15, IL17/IL17A, IL18/IL1F4, IL23, MIF, TRANCE/TNFSF11/ RANKL, S100A9, TNF α, VEGF α. Statistical analysis was performed using SPSS v 28.0 for Windows. Results: For Aim 1 136 patients were enrolled (102 female, median age at onset 3 years (range 1 15), of whom 55.9% had oligoaJIA, 40.4% uveitis and 72.8% ANA positivity. Adalimumab (59.3%) and Etanercept (40.7%) were started at a median age of 6 years (range 1-16), TNFi were discontinued after a median time of 30 months (range 6-90), increasing the interval (76.5%), reducing the dose (18.4%) and abrupt discontinuation (16.9%). 79.4% of patients relapsed after a median time of 5 months (range 0.5-66). Patients with uveitis relapsed earlier (Log Rank χ²16.4 p<0.0001), while patients who lengthened the interval of administration showed a later relapse (Log Rank χ²6.95 p0.008). Uveitis (HR 2.11 CI 1.34-3.31), age at onset (HR 0.909 CI 0.836-0.987), duration of tapering (HR 0.938 CI 0.893-0.985) and to have a persistent OligoJIA (HR 0.597 CI 0.368-0.968) are significant predictors of disease relapse (Mantel-Cox χ²34.23 p<0.001). For Aim 2 we collected the blood samples from 52 children with JIA (37 Female), with a median age at onset of 47.6 months (IQR 22.3-111.8), of whom 41 samples for cytokines from 34 children (7 contributed twice), and 50 for lymphocyte subsets from 38 children (12 contributed twice). Among the 52 children, 28 children had an oligo JIA (53.8%), 18 a polyJIA RF negative (34.6%), 3 a psoriatic form (5.8%), 1 an ERA 1.9%, and 2 polyarticular JIA RF+ (3.8%). Eight children had uveitis (15.4%), 49 were ANA positive (94.2%), 2 RF positive (3.8%). At the last available follow up, after a median duration of disease of 60 months (IQR 34.2-113.2), 11 children showed an active JIA (21.2%, of whom 6 active off-therapy (11.5%), 5 active on therapy (9.6%)) and 41 were on remission (78.8%, of whom 8 off-therapy (15.4%), 33 on therapy (63.5%)). Lymphocytes subpopulation were performed after a median duration of the disease of 32 months (IQR11.2-81.5) and in 7 children at onset of the disease, while in 43 in other moment of the disease (86%, 14 active on therapy (28%), 15 remission on therapy (30%), 3 active off-therapy (6%), 8 pre-withdrawal (16%) and 3 post-withdrawal (6%). At the moment of the samples 26 were on remission from an articular and ocular perspective (52%), while 24 were active for at least one of the two items (48%).We observed significant differences in the distribution of the proportion of CD3+Cd4+ and the moment of the disease at blood collection (p0.028), and the ratio of CD4+/CD8+ and the disease status based on the ongoing treatment (p 0.035). Moreover, we observed significant differences in lymphocytes distribution based on the ongoing treatment in the proportion of CD3+CD4+ (p0.008), in the ratio CD3+CD4+/CD8+ (p0.016) and in the proportion of CD16+CD56+ (p0.05). Similarly, we observed significant differences in the distribution of the proportion of CD3+CD8+ (p 0.012), absolute value of CD3+CD8+ (p0.015) and in the ration CD4+/CD8+ (p0.026) and the ability to achieve and/or maintain the remission on treatment after lymphocyte subpopulation. As well as, among the 19 patients in whom we were able to timely stop the treatment for persistent remission we identified significant differences between children who maintained remission versus who relapsed in the proportion of CD3+CD8+ and in the ratio CD4+/CD8+. At the moment of the cytokine/chemokine determination the median duration of the disease was 39.2 months (IQR 17.8-102.9), At the moment of the samples 32 were on remission from articular and ocular perspective (78%, of whom 30 on treatment (73.1%) and 2 off-treatment (4.9%)), while 9 were active for at least one of the two items (22%, of whom 3 off treatment (7.3%), and 6 on treatment 14.3%)). In 3 children the cytokine samples were performed at the onset of the disease (7.3%), while in 17 pre-withdrawal (41.5%), while in 21 (51.2%) in other moment of the disease. At this time point, 5 children were off systemic treatment (12.2%), 12 were on methotrexate (12.2%), 20 on adalimumab 29.3%), 4 on etanercept (9.8%). We did not observe a significant difference in the distribution of cytokines/chemokines and the disease status stratified as active and inactive at the moment of blood collection. However, we observed a significant difference in the value of IL-12 P70 among patients at onset, pre-withdrawal and other moment of the disease (p0.020). Significant differences in the value of CCL11/Eotaxin and the status at the last available follow-up stratified as remission off-therapy, active off-therapy, active on therapy and remission on therapy (p0.013) and grouped as active versus inactive (p 0.001) at the last follow-up were observed. Moreover, we evaluated that children who relapsed after drug withdrawal had higher levels of TNFα before drug withdrawal compared to the others. Conclusions: In conclusion, we identified a subset of clinical predictors of disease relapse after the first anti-TNF withdrawal. Moreover we identified in this pilot study a subset of cytokine and chemokines that might help us in better characterize the disease status of our patients and predict the achievement of inactivity on treatment and relapse after drug withdrawal.