Plasma Biomarker Profiling In Langerhans Cell Histiocytosis: Risk-Stratifying The Inflammatory Storm

Abstract

IntroductionDespite indistinguishable histology and the common feature of Birbeck granules in lesion biopsies, clinical presentation of patients with Langerhans Cell Histiocytosis (LCH) is highly variable, from single lesion cured by curretage, to multi-system disease requiring aggressive chemotherapy or stem cell transplant. Risk stratification for Langerhans Cell Histiocytosis has historically assigned clinical risk groups based on anatomic location and extent of LCH lesions, which is the basis for dose and duration of chemotherpy on recent Histiocyte Society trials. In this study, we test the hypothesis that distinct subgroups of patients with LCH may be identified by relative levels of circulating biomarkers. MethodsPre-therapy plasma was collected on 97 patients with LCH (82 Pediatric: 17 High-Risk, 23 Multisystem/Multifocal “Non-risk”, 42 Single Lesion “Non-risk”; 15 Adult: 5 High-Risk, 5 Multisystem/Multifocal “Non-risk”, 5 Single Lesion “Non-risk”) and 49 control subjects (32 Pediatric, 17 Adult). Quantitative levels of plasma proteins (158 analytes) was determined by multiplex analysis with Millipore MagPix kits and the Luminex plate reader. Data were analyzed with both unsupervised and supervised methodologies. ResultsConsensus clustering with non-negative matrix factorization (NMF) clusters identified three groups which were analyzed along with clinical categories. Significant clinical variables included age (adult samples clustered in NMF group 1) and LCH risk category (High-Risk LCH samples clustered in NMF group 3). Samples from patients with the BRAF-V600Emutation or relapse within 1 year did not cluster into any NMF group with signifiance. Additionally, supervised analysis identified specific molecules that were significantly differentially expressed between different clinical categories after multiple testing correction (FDR<0.10): Pediatric LCH vs Adult LCH (72 molecules significant, largest differences in MMP-3, MMP-2 and osteopontin); Pediatric Control vs Pediatric LCH (66 molecules significant, largest differences in SDF-1a, IL-20, MIP-1d, FGF-2 and sIL-4R); Pediatric Low-Risk vs Pediatric High-Risk (47 molecules significant, largest differences in sTNF-R11, sTNF-RI, I-309, sIL2Ra and osteopontin). While previous studies have analyzed expression differences of cytokines in LCH lesions and plasma, in this study the most striking differences are between control vs LCH samples are chemokine molecules. The largest differences between Low-Risk and High-Risk LCH patients include inflammatory cytokines and receptors. ConclusionsDespite mounting evidence supporting pathogenesis of LCH as a myeloid neoplasia arising from immature dendritic cell precursors, these results are consistent with exuberant chemokine and cytokine expression in patients with active LCH, supporting a potential role for inflammation in pathogenesis. This study demonstrates the feasibility of identifying novel LCH sub-groups according to plasma protein profiles with unsupervised analysis, and significant differences can be detected in protein levels between clinical risk groups. Future studies will validate the clinical utility of plasma biomarkers in diagnosis, risk-stratification and determining response to therapy. Finally, feasibility of collecting plasma compared to viable lesions makes plasma studies ideal for prospective collection and analysis in cooperative group studies. Disclosures:No relevant conflicts of interest to declare.