Background/Purpose: The innate immune system may be implicated in the fibrotic processes of systemic sclerosis (SSc); to date little has been done to unravel the immunological events that take place in earlySSc (EaSSc) and definite SSc patients prior to the onset of fibrosis. To this end, we evaluated the response of CD56+ cells (NK/NKT) cells to different stimuli with the potential to trigger immune responses in the early phases of the disease before the onset of fibrosis.

Methods: CD56+ NK/NKT cells were isolated from 22 EaSSc (LeRoy and Medger criteria, e.g. Raynaud phenomenon (RP) + either SSc-specific nailfold videocapillaropic changes and/or SSc-specific autoantibodies and no other manifestations of SSc) and from 13 definite SSc patients (ACR/EULAR 2013 criteria) without evidence of fibrotic features. Cells were stimulated for 24 hours with Toll-Like receptor (TLR) agonists (Pam3CSK4, Poly(I:C), R848) and human recombinant IL-2 and IFNα. Levels of IFNγ, TNFa, IL-6, IL-10, IL-4, IL-13, IL-17, IL-22, IL-8, GM-CSF, RANTES, MIP-1a were measured in cell-free supernatants using a validated multiplex immunoassay based on Luminex technology. Differences in analytes concentrations between groups were tested via the Mann-Withney’s test. We then used the radial coordinate visualization method (Radviz^TM), a projection-based visualization technique, to represent high dimensional data into the orthogonal space to provide an interpretation of the joint effect of cellular responses on the disease status. The best projection is chosen and validated via the VizRank method.

Results: No single analyte could differentiate NK/NKT responses between groups. Joint-effect analysis via VizRank sorted out meaningful results for NK/NKT-products after IL-2 stimulation. The best Radviz^TM representation considering 4 simultaneous attributes is represented in the Graph. This projection had an accuracy to correctly classify patients on the basis of GM-CSF, RANTES, IL-10 and IL-8 levels equal to 82.9%. When these molecules are jointly considered,  the decision boundary between EaSSc and definite SSc patients is clearly marked; most definite SSc lie close to the GM-CSF and far from IL-8 anchor points (e.g have higher GM-CSF levels and low IL-8 levels); on the other hand, most EaSSc patients lie far from the RANTES and close to the IL-10 anchor points (e.g. have low RANTES and high IL-10 levels).

Conclusion: We showed that NK/NKT cells from EaSSc and definite SSc patients present distinct immunological patterns after stimulation with IL-2 and that these patterns accurately differentiate these groups of patients.