Driver mutations in myeloid and lymphoid cells point to multipotent progenitor origin of diverse histiocytic neoplasms

Abstract

Histiocytic neoplasms are rare myeloid diseases characterized by MAPK pathway-activating genetic alterations. We investigated their hematopoietic origin, with a focus on non-Langerhans cell histiocytoses. Using droplet digital polymerase chain reaction assays specific for BRAF, MAP2K1, or KRAS alterations detected in histiocytosis lesions, we could trace the same driver mutation to circulating blood cells in 13 of 14 patients. In 9 of 13 patients, the mutations were detected in circulating lymphoid cells, indicating that multipotent progenitors probably acquired these alterations. The 9 patients included 5 adults with single-system disease, including 3 with recurrent cutaneous xanthogranulomas. The presence of long-lived mutated progenitor cells in these 3 patients was supported by the detection of the same KRAS or BRAF mutation in xanthogranulomas that developed up to 25 years apart. As proof of concept, we traced the driver mutation to circulating CD34⁺ progenitors in 1 of the 3 patients. Distinct secondary mutations in either KRAS, BRAF, or ARAF were identified in separate xanthogranulomas from the same patient, indicating a 2-hit mutational process underlying the formation of these recurrent lesions. Finally, histiocytes and B cells harboring the same KRAS mutation were identified in the unifocal Langerhans cell sarcoma lesion of the only patient without circulating mutated cells. Together, these data point toward multipotent hematopoietic progenitors as the cell of origin of both single-system and multisystem histiocytoses.

Graphical abstract

Figure 1.

Molecular analysis of blood- and tissue-derived hematopoietic cell populations. (A) Frequency of mutation-carrying cells among total PBMCs or lineage-committed cells isolated from blood samples of patients 1 to 14. Every horizontal row represents 1 patient; the shade of red indicates the percentage of mutant cells (with 0% in white and ≥15% in dark red). Data were generated with QuantaSoft software (Bio-Rad). Percentages were calculated by multiplying the fractional abundance by 2, as these are all heterozygous mutations. Gating strategies for lineage cell sorting are found in supplemental Figure 1A. ∗Less than 3 droplets with mutation-specific amplicons; therefore, these samples are considered not unequivocally positive. (B) Exemplary droplet digital polymerase chain reaction (ddPCR) plots revealing the detection of the BRAF^N518S mutation in gran, B, and NK cells isolated from the peripheral blood of patient 11. Droplets containing only mutation-specific PCR products are depicted in blue, whereas droplets containing only wild-type BRAF amplicons are depicted in green. Orange droplets contain both PCR products; gray droplets contain no PCR products. Assay validation results and further details can be found in supplemental Figure 3. (C) Frequency of mutant cells among sorted subsets from lesional tissue from patients 10 and 14, as indicated. (D) Photographs (left) and positron emission tomography image (right) showing the isolated tumor in the left upper thigh of patient 14, who was diagnosed with LCS. (E) Photomicrographs of immunostained tissue slides of the tumor in patient 14, which demonstrated a high Ki67 proliferation index and a complex immunophenotype. The tumor was characterized by a central core containing CD1a⁺ cells, surrounded by many CD1a⁻ CD163⁺ histiocytes; CD14 stained cells from both populations. (F) ddPCR plots depicting the detection of the KRAS exon 2 indel in CD1a⁺ histiocytes, CD1a⁻ histiocytes, and CD19⁺ B cells sorted from lesional tissue in patient 14. The exact gating strategy used for cell sorting is depicted in supplemental Figure 1B. AOX, adult-onset xanthogranuloma; B, B cells; DC, dendritic cells; gran, granulocytes; indel, insertion-deletion; mono, monocytes; MS, multisystem; NK, natural killer; NT, not tested; PBMC, peripheral blood mononuclear cell; RDD, Rosai-Dorfman disease; SS, single system; T, T cells.

Figure 2.

Analysis of patients with recurrent cutaneous xanthogranulomas. (A) Photographs depicting the cutaneous xanthogranulomas in patients 10 to 12, which developed during multiple decades at varying anatomic locations. (B) Graphic representation showing the development of multiple cutaneous xanthogranulomas over time in these patients. Note that only xanthogranulomas confirmed by histology are depicted, whereas many additionally resected lesions were not sent for pathologic evaluation and are therefore not depicted. Xanthogranulomas that were analyzed by next-generation sequencing (NGS) are depicted by large (green or red) colored dots; the color of the dots and their outline indicate the detected somatic mutations. These mutations are also specified. Percentages indicate the variant allele frequencies of detected mutations. In patient 10, NGS was sometimes performed on 2 separate xanthogranulomas resected at the same time point; these instances are indicated by a “2” in the large dot. (C) Gating strategy used to isolate CD34⁺ HSPCs from live CD45^bright/dim PBMCs in patient 10 (left and middle panels). DNA extracted from these cells was subsequently analyzed using KRAS p.G12R-specific ddPCR, demonstrating the presence of the driver mutation in 1.2% of flow-sorted progenitors (right panel). Neg, negative.