Background/Purpose:

Currently, juvenile idiopathic arthritis (JIA) is treated with methotrexate (MTX) as a first-line agent. If children fail to respond, a biologic (e.g. anti-TNF) is usually added. However, a consistent fraction of patients fail to respond even to the combination treatment. The identification of predictive markers of responsiveness to the therapy, together with the elucidation of the underlying mechanism, is therefore a dramatic unmet clinical need. Our approach is based on the premise that complex diseases like JIA can be stratified based on the presence of multi-parametric signatures which characterize an immune function.

Methods:

We have developed a multidimensional technology platform for identifying immune signatures, which includes phenotypical and functional data on a variety of blood cell types (T and B cells, APCs, NK and NKT cells, gdT cells). These tests operate in semi-high throughput (sHT) mode and can be combined with other HT “omics”. This platform has been applied to a homogenous cohort of patients from the  Trial of Aggressive Therapy (TREAT), including patients treated with the same iTNF (Etanercept) and MTX, sampled before (T0) and after (Tend) the therapy, and subsequently stratified in patients reaching inactive disease state, and patients not achieving ID.

Results:

When Teff and Treg functions (proliferation and suppression assays) were analyzed as a whole, we could not find significant changes between ID and NO ID. However, in depth characterization of the two cell subsets led to the identification of specific phenotypical differences at both time points analyzed. When a principal component analysis was applied to the entire set of parameters analyzed, the group of responding (ID, inactive disease) patients clustered and was clearly distinct from non-responding (NO ID) patients. Three different statistical models were applied to test whether selected candidate biomarkers could segregate patients according to their clinical outcome. Strikingly, although most parameters, taken singularly, were not effective in differentiating between the two disease states, the combination of a restricted set of CD4⁺ T cell markers was able to segregate ID and NO ID patients with up to 90% accuracy in both time points, thus providing both predictive and confirmatory value to the study. Moreover, our analysis identified a population of cells characterized by an activated phenotype which is significantly more represented in the NO ID patients. This subset selectively expressed pro-inflammatory markers and chemokine receptors targeting to inflammatory sites, and bore signs of recent antigen recognition (based on Ki67 expression), possibly indicating their direct involvement in autoimmunity. 

Conclusion:

We have identified a cluster of immune functions with high predictive and prognostic power for discriminating the responsiveness to treatment with MTX/iTNF in JIA, which may be developed as a theragnostic tool. Importantly, our approach has also identified a subset of experienced T cells, significantly more represented in clinical failures, which may comprise a potentially pathogenic population responsible for the resistance to treatment.

---

« Back to 2012 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/predictive-markers-of-therapeutical-outcome-and-their-role-in-the-ethiopathology-of-juvenile-idiopathic-arthritis/