New somatic BRAF splicing mutation in Langerhans cell histiocytosis

Langerhans cell histiocytosis (LCH) is an inflammatory myeloid neoplasia with constitutive activation of the MAPKinase RAS-RAF-MEK-ERK cell signaling pathway. We analyzed 9 LCH cases without BRAFV600 and MAP2K1 mutations by whole exome sequencing. We identified a new somatic BRAF splicing mutation in 2 cases. Both cases were childhood single system (SS) LCH cases, with self-healing outcome of the bone lesions. This mutant consisted in a 9 base pair duplication (c.1511_1517 + 2 duplication), encoding for a predicted mutant protein with insertion of 3 amino acids (p.Arg506_Lys507insLeuLeuArg) in the N-terminal lobe of the kinase domain of BRAF. Transient expression of the c.1511_1517 + 2dup BRAF mutant in HEK293 cells enhanced MAPKinase pathway activation, and was not inhibited by vemurafenib but was inhibited by PLX8394, a second-generation BRAF inhibitor able to inhibit signaling of BRAF monomers and dimers. Future LCH molecular screening panel should include this new mutation to better define its prevalence in LCH and its restriction to autoregressive bone SS LCH.


Background
Langerhans cell histiocytosis (LCH) is the most common histiocytosis, and is characterized by inflammatory lesions containing abundant CD1a + CD207+ histiocytes that lead to the destruction of affected tissues [1]. A BRAF V600E mutation, responsible for activation of the MAPKinase RAS-RAF-MEK-ERK cell signaling pathway in pathologic histiocytes, is present in~55% of LCH cases and was associated with recurrence and high-risk presentation [2]. Responses to BRAF inhibitors in patients with BRAF V600E -mutated LCH confirms that BRAF V600E is a driver mutation in LCH [3]. Although~45% do not have BRAF V600E mutation, ERK was reported to be activated in pathologic histiocytes of all LCH samples [4]. Other molecular alterations have also been reported to activate the MAPKinase pathway in BRAF V600E -non mutated LCH, such as MAP2K1 mutations (10-20% of LCH) [5,6], β3-αC loop deletion in the kinase domain of BRAF (6% of LCH) [7], and case reports highlighted mutation on ARAF [8] and MAP3K1 [9]. Fusion events involving BRAF and activating MAPkinase pathway have also been reported in histiocytoses of the L group [7,10].
To identify the mechanism of pathologic ERK activation in the remaining LCH, we performed whole exome sequencing (WES) on selected LCH frozen biopsy samples wild-type for the most common activating mutations reported in LCH. DNA extracted from peripheral white blood cells (PBMC) were used as the "normal" sample for comparison. Mutation function and response to MAPKinase pathway inhibitors were assessed using in vitro constructs.

Results
From the French LCH registry [11], 9 patients fulfilled the following inclusion criteria: i) fresh frozen biopsy tissue and blood samples available, ii) high percentage of lesionsinfiltrating CD207+ histiocytes (>30%), iii) no mutation identified by BRAF V600E pyrosequencing [2] or among the most common activating mutations of PIK3CA, BRAF, KRAS and NRAS with the i-plex mass spectrometric based genotyping technology (Sequenom-Agena Bioscience) [12], iv) negative screening for exon 2-3 MAP2K1 mutations by Sanger sequencing. Among the 9 included patients, 7 had a bone-limited LCH and 2 had a LCH involving several organs ( Table 1).
Detection of duplication at the end of BRAF exon 12 in LCH samples A somatic duplication of 9 base pairs at the end of exon 12 of BRAF (nucleotides c.1511_1517 + 2) was detected in LCH samples from 2 patients (P5 and P6). Both patients were children with self-healing bone lesions. This duplication was not yet reported in the COSMIC database. For both patients, Sanger sequencing of genomic DNA confirmed the BRAF c.1511_1517 + 2 duplication in LCH lesions (Fig. 1a), but failed to detect it within PBMC. This 9 nucleotides insertion at the position +2 of the splice donor site of intron 12 was predicted to change the splicing, with an insertion of 9 nucleotides in the cDNA sequence [GTTACTCAG] at the end of exon 12 (Fig. 1b). Messenger RNA was extracted from lesion of P5, and length analysis of PCR products of cDNA confirmed a 9 nucleotides insertion (Fig. 1c). Insertion was also confirmed by Sanger sequencing (Additional file 1: Figure. S1).
To investigate the prevalence of the somatic BRAF c.1511_1517 + 2 duplication in LCH, we studied 28 additional LCH samples wild-type for BRAF V600 , by length analysis of PCR products. No additional mutated case was found, suggesting that this mutation represents a small proportion of BRAF V600 wild-type LCH (<10%) [7], but more studies are needed to estimate precisely its prevalence.

Functional analysis and response to MAPkinase inhibitors
The insertion of 3 amino acid (p.Arg506_Lys507insLeu LeuArg) coded by this 9 base pair duplication is localized in the smaller N-terminal lobe of the kinase BRAF domain responsible for ATP binding (Additional file 1: Figure. S2). Small deletions localized nearby this region of BRAF were shown to induce MAPkinase pathway activation in LCH and in pancreatic carcinomas [7,13], suggesting that this new mutation may also have functional impact. Immunohistochemistry of samples of P5 and P6 confirmed that areas rich in CD1a + histiocytes contained numerous histiocytes with phosphoERK in their cytoplasms as well as translocated into the nucleus (Fig. 1d). The strong phosphorylation of ERK in LCH lesions of P5 and P6 was also confirmed by Western blot (Fig. 1e).
We then assessed the functional impact of this genomic alteration on BRAF signaling by analyzing phosphorylation of ERK in HEK293 cells transiently transfected with wild-type BRAF, BRAF V600E or BRAF c.1511_1517 + 2dup mutants. cDNA expression of BRAF c.1511_1517 + 2dup, but not wild-type BRAF, resulted in a significant increase in ERK1/2 phosphorylation (Fig. 1f).
We also evaluated the ability of the BRAF V600E inhibitor vemurafenib and the MEK inhibitor trametinib to suppress ERK activation by specific BRAF alterations. Although vemurafenib induced a substantial inhibition of BRAF V600E -induced activation, this drug did not inhibit the MAPkinase activation in cells transfected with the cDNA containing the BRAF c.1511_1517 + 2dup mutant, which is consistent with specific activity of this agent against mutations that result in active BRAF monomers. Trametinib, which blocks active MEK, decreased activation of ERK in cells transfected with BRAF V600E , but with no impact on cells transfected with the BRAF c.1511_1517 + 2dup mutant (Fig. 1f ). We thus further evaluated the effects of other, or combination of inhibitors of the MAPkinase pathway. As expected TCS ERK 11e, which directly inhibits ERK, induced a total extinction of ERK phosphorylation (Fig. 1g). PLX8394 is a second-generation BRAF inhibitor able to inhibit signaling of BRAF monomers and dimers without paradoxical activation of MAPKpathway signaling in cells with wild-type BRAF that has been observed in firstgeneration agents such as vemurafenib [14,15]. PLX8394 induced an almost complete extinction of pERK signal on Western blot, confirming that most of the pathway activation was due to the mutant BRAF. Again vemurafenib or trametinib alone did not suppress ERK activation, but combination of both drugs induced a completed extinction of the pERK signal (Fig. 1g).
To elucidate this last observation, we performed a dose response experiment with vemurafenib and trametinib on BRAF c.1511_1517 + 2dup transfected cells and BRAF V600E transfected cells, in order to test if an increased dose for trametinib was required to block activation by the BRAF c.1511_1517 + 2dup mutation as compared to other LCH-associated BRAF mutations. In our model, while vemurafenib and trametinib induced an inhibition of BRAF V600E -induced activation with relationship between dose and response, these drugs did not inhibit the MAPkinase activation in BRAF c.1511_1517 + 2dup transfected cells regardless of dose (Fig. 1h). Neither vemurafenib nor trametinib used individually, even at the highest concentration, could inhibit phosphorylated MEK1/2 and phosphorylated ERK1/2. Future study should better define mechanisms of resistance of the BRAF c.1511_1517 + 2dup mutation by targeted therapies such as vemurafenib and trametinib.

Conclusions
We report here a new somatic BRAF splicing mutation in LCH, leading to the insertion of 3 amino acids (p.Arg506_Lys507insLeuLeuArg) in the N-terminal lobe of the kinase domain of BRAF. This mutation constitutively    activates the MAPKinase pathway, and was inhibited by the second-generation BRAF inhibitor PLX8394. Thanks to recent substantial effort of LCH expert teams, the unknown part of the molecular spectrum of LCH continues to shrink and identification of these mutations has many potential applications such as targeted therapy, therapeutic riskstratification based on tumor genotype, and quantitative detection of mutant allele in circulating cell free DNA as possible blood biomarkers.