BRAF-V600E expression in precursor versus differentiated dendritic cells defines clinically distinct LCH risk groups

Abstract

Langerhans cell histiocytosis (LCH) is a clonal disorder with elusive etiology, characterized by the accumulation of CD207(+) dendritic cells (DCs) in inflammatory lesions. Recurrent BRAF-V600E mutations have been reported in LCH. In this study, lesions from 100 patients were genotyped, and 64% carried the BRAF-V600E mutation within infiltrating CD207(+) DCs. BRAF-V600E expression in tissue DCs did not define specific clinical risk groups but was associated with increased risk of recurrence. Strikingly, we found that patients with active, high-risk LCH also carried BRAF-V600E in circulating CD11c(+) and CD14(+) fractions and in bone marrow (BM) CD34(+) hematopoietic cell progenitors, whereas the mutation was restricted to lesional CD207(+) DC in low-risk LCH patients. Importantly, BRAF-V600E expression in DCs was sufficient to drive LCH-like disease in mice. Consistent with our findings in humans, expression of BRAF-V600E in BM DC progenitors recapitulated many features of the human high-risk LCH, whereas BRAF-V600E expression in differentiated DCs more closely resembled low-risk LCH. We therefore propose classification of LCH as a myeloid neoplasia and hypothesize that high-risk LCH arises from somatic mutation of a hematopoietic progenitor, whereas low-risk disease arises from somatic mutation of tissue-restricted precursor DCs.

Figure 1.

Clinical status, BRAF genotype, and clinical outcomes. (A) Estimates for refractory or recurrent disease in the study population by disease risk (red: high-risk, blue: low-risk). (B) Estimates for survival based on disease risk (red: high-risk, blue: low-risk). (C) Estimates for refractory or recurrent disease by lesion BRAF genotype (red: wild-type, blue: BRAF-V600E). (D) Estimates for survival based on lesion BRAF genotype (blue: wild-type, red: BRAF-V600E).

Figure 2.

Identification of cells with BRAF-V600E mutation in LCH lesions, circulating cells, and BM aspirates. (A) gDNA was isolated from biopsies of LCH lesions, and the percentage of cells with the BRAF-V600E allele was estimated by qPCR (line = median, 8.0%). (B) gDNA was isolated from PBMCs of patients with active LCH before initial therapy or at a time of recurrence before salvage chemotherapy. The percentage of circulating cells with the BRAF-V600E allele in patients from different clinical risk categories was estimated by qPCR. (C) Estimate of recurrence by circulating BRAF status among those who were clinically classified with low-risk LCH (blue = no circulating cells with BRAF-V600E detected, red = circulating cells with BRAF-V600E detected). (D) gDNA was isolated from gradient-separated leukocytes from synchronous peripheral blood and BM aspirate collected from patients with active LCH. The percentage of cells with the BRAF-V600E allele was determined by qPCR. (Every bar represents a single blood sample; technical duplicates were used in all experiments.) (E) The percentage of circulating cells with the BRAF-V600E allele was determined by qPCR of serial PBMC samples in one patient with disease refractory to multiple salvage therapies (red), and finally cured with clofarabine (green). (Every data point represents a single blood sample; technical duplicates were used in all experiments.)

Figure 3.

BRAF-V600E localizes to circulating myeloid DCs and monocytes in peripheral blood. (A) Sorting strategy for peripheral blood mononuclear cells into monocyte (CD14⁺), myeloid DC (lineage⁻ HLADR⁺ CD11c⁺ BDCA2⁺), and plasmacytoid DC (lineage⁻ HLADR⁺ BDCA2⁺) fractions. (B) Percentage of BRAF-V600E cells in circulating monocyte and DC fractions from the indicated LCH patients. (Every bar represents a single sorted blood sample; technical duplicates were used in all experiments.)

Figure 4.

BRAF-V600E localizes to CD34⁺ hematopoietic progenitor cells in BM. (A) Sorting strategy for BM aspirate cells into HSPCs (CD34⁺) and monocytes (CD14⁺). (B) CD34⁺ and CD14⁺ cells were sorted from BM aspirate fractions, gDNA was purified and amplified, and then the BRAF-V600E allele was detected by qPCR. (C) Percentage of CD34⁺ cells with BRAF-V600E in BM aspirate and the percentage of BRAF-V600E cells obtained from CD34⁺ HSPC-derived colonies generated in CFU Assay in two LCH patients. (D) PBMCs were sorted by FACS into CD3⁺ and CD19⁺ fractions. Percentage of peripheral B and T cells with BRAF-V600E in different LCH patients. (B–D: every bar represents a single sorted blood sample; technical duplicates were used in all experiments.)

Figure 5.

BRAFV600E^langerin mice spontaneously developed progressive accumulation of inflammatory infiltrates in peripheral tissue. (A and B) BRAFV600E^langerin mice were generated as described in Materials and methods and sacrificed at 12 wk (A) or 20 wk (B). Representative lung and liver sections were stained by H&E (in total at least 4–5 mice per group were analyzed; arrows indicate inflammatory infiltrates; bars, 100 µm). (C and D) Lung and liver tissue of BRAFV600E^langerin mice and control littermates were harvested at 12 wk of age and single cells suspensions were analyzed by multicolor flow cytometry analysis after digestion with collagenase. Absolute numbers of total MHCII⁺ CD11c⁺ DCs, MHCII⁺ CD11c⁺ CD103⁺ DCs, and MHCII⁺ CD11c⁺ CD11b⁺ DCs are presented (C, all cells are pregated on singlets, viable CD45⁺ cells, n = 3 per group, representative of two independent experiments). Absolute number of tissue infiltrating total hematopoietic CD45⁺ cells, F4/80⁺ macrophages, total CD3⁺ NK1.1⁻ T cells, total CD3⁺ NK1.1⁻ CD4⁺ CD8⁻ T cells, total CD3⁺ NK1.1⁻ CD4⁺ Foxp3⁺ T reg cells, total CD19⁺ B220⁺ B cells, and total CD3⁻ NK1.1⁺ NK cells are depicted in D (all cells are pregated on singlets, viable CD45⁺ cells, n = 3 per group, representative of two independent experiments). (E) Representative picture of IF staining for langerin in liver lesion of 12-wk-old BRAFV600E^langerin mouse. Bars, 100 µm. Data are shown as mean ± SEM. *, P < 0.05.

Figure 6.

BRAFV600E^CD11c mice develop an aggravated phenotype with massive infiltration of MHCII⁺ CD11c⁺ langerin-expressing cells in peripheral tissues. (A) Representative pictures of the spleen and liver of BRAFV600E^CD11c mice (right) and control littermates (left) at 12 wk of age (n = 16 per group). (B) Representative H&E staining of liver and lung tissue sections of BRAFV600E^CD11c mice and corresponding controls at the indicated ages (arrow marks multinucleated giant cells; bars, 100 µm; n = 5–11 animals per group). (C) H&E staining of back skin tissue sections of 12-wk-old BRAFV600E^CD11c mice and controls (arrows indicate multinucleated giant cells; bars, 100 µm; n = 4 per group). (D) Absolute numbers and percentages of MHC II⁺ CD11c⁺ DC among total CD45⁺ lung and liver cells in 12-wk-old BRAFV600E^CD11c mice and control littermates were determined by multicolor flow cytometry (all cells pregated on singlets, viable cells, n = 3–5 per group, representative of three independent experiments). (E) Representative image of frozen liver tissue section isolated from the liver of 16-wk-old BRAFV600E^CD11c mice stained with anti-langerin mAb (left) and anti-CD11c mAb (right) or with both langerin and CD11c mAb (merged) and analyzed by confocal microscopy (bars, 100 µm). (F) Langerin mRNA expression in whole liver tissue of BRAFV600E^CD11c mice and corresponding controls was analyzed by qPCR (normalized to Gapdh mRNA expression, n = 6–11 per group). All data are shown as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 7.

Expression of BRAF-V600E in early BM-resident DC progenitors in BRAFV600E^CD11c mice results in multi-systemic high-risk disease and is associated with high local cytokine expression and the recruitment of additional inflammatory cells. (A) Relative mRNA expression of BRAF-V600E in sorted BM lineage negative sca1⁺ c-kit⁺ (LSK), lin⁻ sca1⁻ c-kit⁺ CD34⁺ CD16/32⁺ granulocyte myeloid progenitors (GMPs), lin⁻ sca1⁻ CD135⁺ c-kit^high CD115⁺ monocyte and DC progenitors (MDPs), and lin⁻ sca1⁻ CD135⁺ c-kit^low CD115⁺ common DC progenitors (CDPs) and lung/liver CD45⁺ MHCII⁺ CD11c⁺ DCs in BRAFV600E^CD11c mice as assessed by mutation-specific qPCR (data normalized to total BRAF expression, no BRAF-V600E expression detected in corresponding cells of control mice, n = 3, pooled data of three independent isolations). (B and C) Relative numbers of viable singlet CD11b⁺ CD115⁺ Flt3⁺ pre-DCs and CD11c⁺ MHCII⁺ DCs among circulating viable, singlet CD45⁺ cells (B) and hemoglobin levels (C) in BRAFV600E^CD11c mice, BRAFV600E^langerin mice, and control mice (n = 4–10 per group, pooled data of four independent experiments, each data point represents one mouse). (D) Immunohistochemical analysis of liver and lung tissue sections stained with indicated mAb in BRAFV600E^CD11c mice (top) and control littermates (bottom; bars, 50 µm). (E) Absolute numbers of tissue-infiltrating total hematopoietic CD45⁺ cells, F4/80⁺ macrophages, total CD3⁺ NK1.1⁻ T cells, total CD3⁺ NK1.1⁻ CD4⁺ CD8⁻ T cells, total CD3⁺ NK1.1⁻ CD4⁺Foxp3⁺ T reg cells, total CD19⁺ B220⁺ B cells, total CD3⁻ NK1.1⁺ NK cells, and total SiglecF⁺ CD11c⁻ eosinophils in lung and liver of BRAFV600E^CD11c mice and control littermates as assessed by multicolor flow cytometry (n = 4–5 per group, representative of two independent experiments). (F) Local expression of indicated chemokines and cytokines in peripheral liver tissue was assessed by qPCR (data normalized to Gapdh expression, n = 4–5 per group). All data are shown as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.