Lineage switching of the cellular distribution of BRAFV600E in multisystem Langerhans cell histiocytosis

Abstract

Most children with high-risk Langerhans cell histiocytosis (LCH) have BRAFV600E mutation. BRAFV600E alleles are detectable in myeloid mononuclear cells at diagnosis but it is not known if the cellular distribution of mutation evolves over time. Here, the profiles of 16 patients with high-risk disease were analyzed. Two received conventional salvage chemotherapy, 4 patients on inhibitors were tracked at intervals of 3 to 6 years, and 10 patients, also given inhibitors, were analyzed more than 2 years after diagnosis. In contrast to the patients responding to salvage chemotherapy who completely cleared BRAFV600E within 6 months, children who received inhibitors maintained high BRAFV600E alleles in their blood. At diagnosis, mutation was detected predominantly in monocytes and myeloid dendritic cells. With time, mutation switched to the T-cell compartment, which accounted for most of the mutational burden in peripheral blood mononuclear cells, more than 2 years from diagnosis (median, 85.4%; range, 44.5%-100%). The highest level of mutation occurred in naïve CD4+ T cells (median, 51.2%; range, 3.8%-93.5%). This study reveals an unexpected lineage switch of BRAFV600E mutation in high-risk LCH, which may influence monitoring strategies for the potential withdrawal of inhibitor treatment and has new implications for the pathogenesis of neurodegeneration, which occurred in 4 patients.

Figure 1.

Clinical progress and distribution of BRAF^V600E alleles in PBMC and sorted cell fractions. (A) Patients 1 and 2 treated with conventional chemotherapy. (B) Patients 3 to 6 treated with inhibitors. Pie charts depict the distribution of mutated cells as a proportion of all mutated cells detected in blood. Each chart is aligned as closely as possible to the corresponding time point. Blue symbols at the top indicate chemotherapy cycles; red bar indicates MAPK inhibitor. Clinical scores and mutation analyses over time are shown in rows (top to bottom). Filled gray areas indicate negative results below 0.0001. Disease Activity Score (DAS) was measured according to Donadieu et al; BRAF^V600E mutation fraction measured in total PBMC (filled symbols and lines) and plasma cell-free DNA (open symbols and broken line); BRAF^V600E mutation fraction measured in sorted subsets; and mutated cells per μL in sorted subsets. Asterisks indicate rapid disease reactivation when treatment was suspended. Red line indicates classical CD14⁺ monocytes; orange line indicates CD16⁺ nonclassical monocytes; blue line indicates CD1c⁺ myeloid DC; and green line indicates CD3⁺ T cells. VBL/pred, vinblastine and prednisone; VCR/pred/Ara-C, vincristine, prednisone, and cytarabine; 2CdA/Ara-C, cladribine and cytrarabine; 6-MP/MTX: 6-mercaptopurine and methotrexate maintenance.

Figure 2.

Comparison and quantitation of alleles in lymphoid subsets. (A) The percentage contribution of each PBMC subset to the total mutation burden (obtained by summing the parts) among CD3⁺ T cells, CD19⁺ B cells, and myeloid cells (monocytes and DCs). (B) The percentage contribution of each T-cell subset to the mutation burden present in T cells among naïve, central memory (CM), effector memory (EM), and effector memory cells re-expressing CD45RA (TEMRA). Supplemental Figure 1 shows the sorting strategy. In panels A and B, the graphs show data from 1 late time point of all patients with a sufficient material available (n = 13; analysis of variance). (C) Summary of the mutant allele fraction in PBMC in all samples as a function of time on treatment. (D) Summary of the percentage contribution from T cells (green) and all myeloid cells (red) to the total mutation burden of these subsets as a function of time since diagnosis (n = 33). Broken lines depict the 95% confidence intervals for a simple linear regression. Dabra, dabrafenib; NK, natural killer; Tram, trametinib; Vem, vemurafenib.