Novel Recurrent Mutations in Eldheim-Chester Disease Patients Identified By Whole Exome Sequencing and Whole Genome Sequencing

Background: Erdheim-Chester disease (ECD) is a type of systemic, non-Langerhans histiocytic disorder characterized by diffuse organ damage with infiltration of CD68-positive, CD163-positive, and CD1a-negative histiocytes. Although most patients have mutations associated with the hyperactivation of the MAPK pathway, including BRAF, MAP2K1, and NRAS, the remaining have no known driver mutations. So far, there is no specific target of treatment for them. To elucidate novel driver mutations and establish new treatment strategies in these patients, we performed whole-exome sequencing (WES) and whole-genome sequencing (WGS) using patient samples with ECD. In addition, we performed WES with multi-organ lesions in a patient with ECD to reveal the mutational profiles of each organ lesion.

Methods: We performed a nationwide survey and collected clinical samples of ECD in Japan. We collected 22 samples of ECD lesions from 15 adult patients. All cases were pathologically proved. Twenty of 22 samples were formalin-fixed and paraffin-embedded tissue (FFPE) and were examined by WES. For WGS, DNA was extracted from 2 raw samples of ECD lesions. In addition, we collected multi-organ lesions in 3 patients. One patient developed myelodysplastic syndrome (MDS) and subsequent acute myeloid leukemia (AML). Therefore, we also performed WES with each bone marrow (BM) sample to reveal the relationship between ECD and these myeloid malignancies. We used 6 samples of peripheral blood, 1 BM sample, 1 skin sample, and 1 oral mucosa sample as normal controls.

Result: We detected known driver mutations in seven of 15 cases (47%). Among them, BRAF V600E was detected in 5 cases (8 samples), MAP2K1 C121S in 1 case (1 sample), and NRAS Q61R in 1 case (2 samples) by WES and WGS. A mean of 69 nonsynonymous mutations per patient was identified in normal-tumor analysis (range, 2-491) and 188 in tumor-only analysis (range, 17-3598) of WES, and 3134 in normal-tumor analysis (range, 2588-3680) of WGS. The median variant allele frequency (VAF) for the 11 samples with known activating kinase mutations identified by WES and WGS was 14.4% (range, 6.3-34.7). We could not detect known driver mutations in the other 8 cases (53%). Therefore, to reveal novel driver mutations, we focused on these 8 cases. Notably, EPHA2 P786L, MYBPC3 D798N, TDRD5 P115L, and TCEAL4 F17L mutations are recurrently found in 2 out of the 8 cases, suggesting new driver mutations are contained in these. The VAFs of EPHA2 P786L were 9.1% and 9.3%, MYBPC3 D798N 5.5% and 11.9%, TDRD5 P115L 10.2% and 18.7%, and TCEAL4 F17L 7.5% and 9.0% in each case. We also analyzed multiple organ samples of a case with BRAF V600E mutation (5 samples of ECD lesions, BM with MDS, BM with AML, and skin as normal control). Interestingly, BRAF V600E mutation was identified in 3 samples (bone lesion, heart, and intestine) but was not in other samples (kidney lesion, dura matter, BM tumor, BM with MDS, and BM with AML), suggesting mutational profiles are different depending on the organ of the lesion.

Conclusion: Our nation-wide analysis revealed that no well-known driver mutations were found in more than half of the ECD cases. We identified EPHA2 P786L, MYBPC3 D798N, TDRD5 P115L, and TCEAL4 F17L mutations as candidates of novel driver mutations in ECD. To elucidate the role of these mutations in pathogenesis of ECD, further functional analyses are warranted. In addition, we revealed heterogeneity of mutational profile of multi lesions in an ECD patient with BRAF V600E mutation. It is noteworthy that mutational profiles might be different depending on the organ of the lesion.