BRAF-V600E utilizes posttranscriptional mechanisms to amplify LPS-induced TNFα production in dendritic cells in a mouse model of Langerhans cell histiocytosis

Abstract

Langerhans cell histiocytosis (LCH) is an inflammatory disease characterized by abnormal dendritic cells (DCs) with hyperactive ERK signaling, called “LCH cells.” Since DCs rely on ERK signaling to produce inflammatory molecules in response to pathogenic cues, we hypothesized that hyperactive ERK enhances DCs inflammatory responses. We specifically investigated TLR4-induced TNFα production in LCH cells by utilizing the BRAF-V600E^fl/+:CD11c-Cre mouse model of LCH, which hyperactivates ERK in DCs. We measured LPS-induced TNFα production both in vivo and in vitro using splenic CD11c+ cells and bone marrow-derived DCs with or without pharmacologic BRAF^V600E inhibition. We observed a reversible increase in secreted TNFα and a partially reversible increase in TNFα protein per cell, despite a decrease in TLR4 signaling and Tnfa transcripts compared with controls. We examined ERK-driven, post-transcriptional mechanisms that contribute to TNFα production and secretion using biochemical and cellular assays. We identified a reversible increase in TACE activation, the enzyme required for TNFα secretion, and most strikingly, an increase in protein translation, including TNFα. Defining the translatome through polysome-bound RNA sequencing revealed up-regulated translation of the LPS-response program. These data suggest hyperactive ERK signaling utilizes multiple posttranscriptional mechanisms to amplify inflammatory responses in DCs, advancing our understanding of LCH and basic DC biology.

1 |. INTRODUCTION

Histiocytic diseases involve aberrant mononuclear phagocytes (MNPs) that accumulate in tissues. MNPs are a system of cells derived from the myeloid lineage and include macrophages and dendritic cells (DCs).^1,2 The most common pediatric histiocytosis is Langerhans cell histiocytosis (LCH), with an incidence rate of 4–9 per million children per year, most often in the first 3 years of life.^3–5 Severe forms of the disease are considered cases that involve risk organs such as the liver, bone marrow, and spleen. Patients in this group have a progression-free survival rate under 50%. First- and second-line therapies are highly toxic and commonly result in relapse.^6–9 A better understanding of LCH etiology is required to develop more effective and safer therapeutics.

LCH is characterized by the presence of pathogenic CD207⁺CD1a⁺ DCs (“LCH cells”) within inflammatory lesions that can lead to multisystem organ damage. LCH cells resemble Langerhans cells, but genetic and phenotypic studies of human LCH cells from lesion biopsies suggest that these cells are more closely related to conventional DCs and express CD11c, a common conventional DC marker.^10–12 LCH cells universally exhibit hyperactive ERK due to somatic mutations within the Ras-ERK pathway, most commonly BRAF^V600E, an activating missense mutation.^13–15 Despite this clear molecular defect in LCH, it is not fully understood how LCH cells propagate and sustain inflammatory lesions.

Extensive work has gone into investigating the impact of BRAF^V600E on cellular functions in many contexts. BRAF is a central kinase in the conserved Ras-ERK signaling pathway, a kinase cascade that couples extracellular stimuli to intracellular signals. The phosphorylation of ERK1/2 regulates the function of a wide variety of proteins to mediate numerous cellular functions including cell growth, proliferation, survival, differentiation, and apoptosis.^16–19 In addition to these universal cellular programs, ERK signaling also plays a critical role in many inflammatory functions of specialized immune cells.²⁰ DCs are specialized innate immune cells that play a pivotal role in orchestrating immune responses. DCs are equipped with specialized pathogen recognition receptors including TLRs, which enable them to sense pathogenic materials in their environment. In response to TLR stimuli, DCs rely on ERK signaling to appropriately produce and secrete pro- and anti-inflammatory cytokines and chemokines to communicate with lymphocytes.^21–25 Tight regulation of ERK signaling is necessary to prevent inappropriate immune responses.²⁶

Inflammation in LCH patients occurs both locally within lesions, as measured by the presence of infiltrates^27–28 and cytokines,^11,29–30 and also systemically, as measured by circulating cytokines.³¹ Given the inflammatory nature of LCH lesions and lack of knowledge regarding hyperactive ERK in DCs, we sought to investigate the effects of BRAF^V600E signaling on DC-specific inflammatory functions. We hypothesize that BRAF^V600E expression enhances the response of DCs to pathogenic stimuli. Since TLRs depend on ERK signaling to produce and secrete inflammatory cytokines, we measured a canonical DC TLR response, TLR4-mediated TNFα production, in LCH cells. Using an established transgenic mouse model of LCH,¹³ we discovered increased TLR4-induced TNFα secretion and TNFα protein levels per cell, despite a reduction in TLR4-induced NFkB activation and Tnfa transcription compared with wild type (WT). We therefore investigated ERK-mediated posttranscriptional mechanisms impacting TNFα production and secretion and identified an increase in both the activity of the enzyme that cleaves membrane bound TNFα, TACE, as well as mRNA translation of inflammatory programs. These effects likely contribute to the amplified TNFα secretion and protein levels, respectively. Using TLR4-induced TNFα production as a model, we demonstrate that BRAF^V600E expression disrupts multiple posttranscriptional mechanisms in DCs resulting in an intensified response to environmental cues, likely contributing to the reactive nature of LCH lesions in patients.

2 |. MATERIALS AND METHODS

2.1 |. Mice

Mice were housed in our Association for Assessment and Accreditation of Laboratory Animal Care-certified animal facility and all procedures were approved by the Institutional Animal Care and Use Committee (protocol no. 921). Braf-V600E^fl/+ (017837) and CD11c-Cre (008068) mice were purchased from Jackson Laboratories. Bred Braf-V600E^fl/fl mice were crossed with CD11c-Cre mice to yield both Braf-V600E^fl/+:CD11c-Cre (LCH) and Braf-V600E^fl/+:Cre-negative (WT) mice and tail-snips were used for genotyping. Mice used in experiments were between 6 and 10 weeks of age and efforts were made to ensure equal use of male and female mice in all experiments.

2.2 |. In vivo analysis of LPS-induced TNFα production

Mice were injected with 2 mg/kg of LPS (catalog no. L4524; Sigma–Aldrich) or PBS through the intraperitoneal cavities. Mice were euthanized by CO₂ asphyxiation after 2 h, and blood was drawn from cardiac puncture. Serum was isolated for cytokine analysis by ELISA (BD OptEIA; BD Biosciences) as per the manufacturer’s protocol. Spleens were harvested and single cell suspensions were subjected to CD11c+ magnetic bead isolation following the manufacturer’s protocol (catalog no. 130-052-001; Miltenyi). The CD11c+ population was spun down and resuspended in buffer RLT and stored in −80 prior to RNA extraction.

2.3 |. Processing of organs

Whole bone marrow cells were flushed from the hindlegs of LCH and control mice using cold PBS and single-cell suspensions were generated by mechanical disruption through a 70-μm strainer. Whole spleens were digested with 30 mg/ml DNase I (Roche) and 1 mg/ml collagenase (Roche) at 37°C for 30 min. Single-cell suspensions of splenocytes were generated by mechanical disruption through a 70-μm strainer. All cells were isolated in sterile conditions and red blood cell lysis was performed using ACK lysis buffer (Lonza). Total cells per organ were counted on a Countess Automated Cell Counter (Thermo Fisher Scientific).

2.4 |. Cell culture and LPS stimulation

To generate CD11c+MHCII+ bone marrow-derived DCs (BMDCs), bulk bone marrow cells were plated (5 × 10⁵ cells/well in 24-well plates; 1 × 10⁵ cell/well in 96-well plates; 35 × 10⁶ cells/well in 15 cm plates) and cultured in IMDM (Gibco) supplemented with 10% heat-inactivated FBS (Atlanta Biologicals), penicillin–streptomycin–l-glutamine (Cellgro), and 10 ng/ml GMSCF (Peprotech) at 37°C in 6% CO₂. Media was changed on day 3 and 6 and cells were experimented on day 7 or 8. For splenocyte DCs, we used Miltenyi CD11c-magnetic bead isolation kit following the manufacturers protocol (catalog no. 130-052-001; Miltenyi), and isolated cells were plated in 96-well plates for 24 h. For all experiments, cells were pretreated with IF10 media ± 0.5 μM PLX7904 for 60 min. To stimulate cells, half of the media was removed and replaced with IF10 ± LPS media (2× is 200 μg/ml) for various time points.

2.5 |. In vitro analysis of TNFα production

To measure TNFα in the supernatant, cell supernatants were collected after stimulation and centrifuged for 10 min at 20,000 × g at 4°C to remove dead cells and debris. Supernatants were stored at −20°C and used for detection of TNFα by ELISA (BD OptEIA; BD Biosciences) following the manufacturers protocol. To measure transcription of tnfa, cells were washed with cold PBS and buffer RLT with beta-mercaptopurine was added directly to the cells. Samples were stored at −80 prior to RNA extraction. To measure intracellular levels of TNFα protein, all stimulations, washes, and staining steps were performed in the presence of 5 μg/ml Brefeldin A (catalog no. 420601; Biolegend) until cells were fixed. Levels of TNFα were calculated using the MFI of intracellular PE-Cy7 conjugated to TNFα (catalog no. 557644; BD Biosciences) within the CD11c^hiMHCII^hi population.

2.6 |. SunRiSE assay

Following the protocol detailed by Argüello et al.,⁵⁶ we treated BMDCs with harringtonine (2 μg/ml) for increasing time (minutes) prior to incubation with puromycin (10 μg/ml) for 10 min. We had to increase the harringtonine incubation times to capture a significant decrease in puromycin staining in our BMDCs. Levels of puromycin were calculated using the MFI of intracellular puromycin staining (catalog no. MABE343-AF488; Millipore) within the CD11c^hiMHCII^hi population and normalized to the 0-min time point.

2.7 |. Flow cytometry

Flow cytometry was performed on a Miltenyi MacsQuant and FACS data were analyzed using FlowJo software. All cellular staining was performed at 4°C under conditions protected from ambient light. Forward and side scatter was used to gate out debris and doublets. Live cells were identified by excluding cells staining positive for LIVE/DEAD Fixable Aqua Dead Cell Stain from Thermo Fisher Scientific. Fc block (anti-CD16/32, clone 2.4G2) was used prior to antibody staining. Cells were surface stained for CD11c in FITC (catalog no. 117306; Biolegend) or in APC (catalog no. 117310; Biolegend) and MHCII in PerCPCy5.5 (catalog no. 562363; BD Biosciences). BMDCs were identified as the CD11c^hiMHCII^hi population. For intracellular staining, the BD Cytofix/Cytoperm reagent (catalog no. 554714, BD Biosciences) was used per the manufacturer’s instructions.

2.8 |. RNA isolation and qPCR

The Qiagen RNeasy Micro kit was used to isolate RNA per the manufacturer’s instructions. RNA was quantified by NanoDrop and reverse transcription was carried out using the SuperScript III kit (Invitrogen) on 0.5–1 μg of purified RNA. Quantitative PCR was performed on the cDNA using Power SYBR Green Master Mix (ThermoFisher) and Quantitect gene expression assays with probes for mouse Tnfα (Mn00443258_m1), mouse Tlr4 (Mn00446193_m1), mouse Gapdh (Mn99999915_g1), and mouse Beta-Actin (NM_007393) (Qiagen). Samples were run using a Sequence Detection System 7500 PCR machine (Applied Biosystems). Levels of mRNA were analyzed using the ΔΔCT method, using levels of Gapdh or ActinB as housekeeping genes. Relative fold changes were plotted in GraphPad PRISM and a 2-way ANOVA was applied.

2.9 |. Protein harvest and Western blot

Cells were lysed using M-PER mammalian protein extraction buffer (product no. 78501; Thermo Scientific) in the presence of Protease/Phosphatase Inhibitor Cocktail (catalog no. 5872; Cell Signaling Technology) according to the manufacturer’s instructions, and protein concentrations were assessed using a 96-well plate-based Bradford assay (product no. 5000006; Bio-Rad protein assay reagent concentrate). Protein samples (20–30 μg) were resolved by SDS-PAGE on 4–12% Bis–Tris gels (Invitrogen), electro-transferred to nitrocellulose membranes (catalog no. 10484060; Bio- Rad). Target proteins were stained overnight in primary antibody at 4° (1:1000 dilution), followed by 60 min of secondary staining (1:10,000) at room temperature. The following Western blot antibodies were purchased from Cell Signaling Technology, Inc.: phospho-ERK (Thr202/Tyr204) (clone D13.14.4E, product no. 4370), total ERK (clone L34F12, product no. 4696), NF-κB p65 (clone C22B4, product no. 4763), phosphor-p65 (Ser536) (clone 7F1, product no. 3036), total eEF2 (product no. 2332), and phospho-eEF2 (Thr56) (product no. 2331). Gapdh (NB100–56875) was purchased from Novus Biologicals. Secondary IRDye antibodies include goat anti-rabbit (800CW and 680RD) and donkey anti-mouse (800CW and 680RD) and were purchased from Li-cor. Western blots were imaged on the Odyssey system (make 9120; Li-Cor) and used to measure densitometric digital assessments. Values were normalized to the appropriate loading control before being plotted in GraphPad PRISM.

2.10 |. Degradation assay

After a 3-h LPS stimulation, BMDCs were treated with 100 μg/ml cycloheximide (catalog no. C4859; Sigma Aldrich) to halt nascent protein translation in the presence of 0.5 μg/ml Brefeldin A (catalog no. 420601; Biolegend) for various time points. To capture both internal and surface TNFα, total protein was harvested as described above and 40 μg was used to measure the concentration of TNFα by ELISA (BD OptEIA; BD Biosciences).

2.11 |. TACE and matrix metalloprotease activity

The SensoLyte 520 TACE Fluorometric Activity Assay (catalog no. AS-72085; AnaSpec) and Generic MMP Colorimetric Activity Assay Kit (catalog no. AS-72095; AnaSpec) were used to measure TACE or MMP activity, respectively, in live BMDCs following an optimization of the manufacturers protocol. After a 2-h LPS stimulation, BMDCs were washed and resuspended in serum-free IMDM. Cells (100 × 10⁵; 50 μl) were transferred to a 96-well plate (black plate for TACE assay) 4× per sample (allowing for technical triplicates and 1 negative control for the enzyme substrate solution). Serum-free IMDM ± the provided enzyme substrate was added (50 μl) to each well and the plate was gently agitated for 30 s. For TACE, continuous measurements were recorded every 5 min for 60 min (ex/em = 490 nm/520 nm) on the Infinite M200 Pro microplate reader (Tecan). For matrix metalloprotease (MMP), measurements at 412 nm were recorded every 10 min for 50 min on the SPECTRAmax 340 PC microplate spectrophotometer (Molecular Devices) and activity was calculated using a standard curve. For all samples, values were first normalized to a no-cell control per plate, then to the negative control per sample, and finally normalized to the 0 s time point per sample. Technical triplicates were averaged and plotted using GraphPad PRISM. For TACE, data points outside the linear phase were excluded and a simple linear regression analysis was used to compare K values between genotypes with and without BRAF^V600E inhibition.

2.12 |. Polysome profile and RNA isolation

After a 3-h LPS stimulation, BMDCs in 15 cm-plates were treated with 100 μg/ml cycloheximide (catalog no. C4859; Sigma Aldrich) to block the translocation step in elongation. Cells were washed, counted, and resuspended at 24 × 10⁶ cells/ml in the presence of cycloheximide. Two aliquots of 1.2 × 10⁶ cells were taken for total RNA and protein isolation. The remaining 21.6 × 10⁶ cells per sample were subjected to polysome profiling as previously described.⁷⁷ Briefly, cell pellets were lysed in polysome lysis buffer (20 mM Tris, pH7.5, 1.5 mM MgCl2, 140 mM KCl, 1% Triton X-100, 100 μg/ml CHX, 0.5 mM DTT, protease inhibitor cocktail) for 10 min on ice with gentle rocking. 15–20 OD260 of total cell extract was loaded on a sucrose gradient (7–45%) generated by a Gradient Maker (BioComp Instruments, Canada) and ultracentrifuged at 217,290 × g for 3 h 20 min at 4°C in a SW40 rotor. Polysome profiling was analyzed with a BioComp fractionator. For detection of RNA distribution, a total of 13 fractions (830 μl/fraction) from polysome profiling were collected by a fraction collector (Cat# 4422151, FC-203B; Gilson). For RNA isolation of fractions, 500 μl of fractions with >2 polysomes were pooled per sample and dispensed into 50-ml conical tubes with 2-parts Trizol LS (catalog no. 10296028; ThermoFisher Scientific). After a 5-min incubation at room temperature, the Direct-Zol RNA Miniprep Kit was used as per the manufacturers protocol (catalog no. R2050; Zymo Research). Samples were eluted in 50 μl water and the concentration of RNA was counted on a Nanodrop. A portion of RNA (30 ng) was used to make cDNA in parallel with the total RNA samples for qPCR.

2.13 |. rRNA analysis

To calculate the number of 18S and 28S rRNAs, BMDCs were counted, washed in PBS, and resuspended in buffer RLT to isolate RNA according to the manufacturer’s protocol. Isolated RNA was submitted to the Nucleic Acid/PCR Core Facility at CHOP where samples were run using Bioanalyzer RNA 6000 Nano reagents (Catalog no. 067–1512; Agilent). The measured (ng/μl) of both 18S and 28S and the respective molecular weights were used to calculate mol/μl, which was then divided by the number of cells/μl from that sample. The mol/cell was then multiplied by Avogadro’s number to get molecules/cell.

2.14 |. RNA sequencing

RNA per sample (150 ng) (both total and polysome-associated) was sent to the Next Generation Sequencing Core (NGSC) (University of Pennsylvania, Philadelphia, PA), which generated the cDNA libraries with enrichment in poly(A)-tailed mRNA. An illumina provided kit to remove dimers was applied to all samples and RNA sequencing was performed using the Xp workflow on the NovaSeq 6000 system.

2.15 |. Analysis of RNA sequencing

Salmon (v1.5.2)⁷⁸ (https://salmon.readthedocs.io/en/latest/) was used to count mRNA and polysome data against the transcriptome defined in Gencode (vM27). Several Bioconductor (v3.14) packages in R were used for the subsequent steps. Genomic features of transcriptomic data were annotated and summarized using the R package “Tximeta” (v1.12.3)⁷⁹ (10.18129/B9.bioc.tximeta) and Ensembl (accessed in R with “biomaRt” (v2.50.0)⁸⁰ using the data set “mmusculus_gene_ensembl”). RNAseq counts for protein coding genes were normalized within anota2seq using Rlog, and significant changes in translation, mRNA abundance and buffering were identified using default parameters⁵⁷ (DOI: 10.18129/B9.bioc.anota2seq). For each contrast, the results were visualized in scatter plots of polysome-associated mRNA log₂ fold change versus total mRNA log₂ fold change. These data have been deposited at the Gene Expression Omnibus with the accession number GSE193610. Differential expression analysis and normalizations were done using DESeq2 and PCAs were plotted using pcaExplorer. Gene ontology analysis was done on the DAVID website using ensemble IDs of translationally up-regulated genes. Gene names of translationally up-regulated genes were entered into the Aura2 website under the “batch” function for regulatory elements.

3 |. RESULTS

3.1 |. Hyperactive ERK increases TLR4-mediated TNFα secretion in DCs

We utilized a transgenic LCH mouse model (BRAF-V600E^fl/+:CD11c-Cre), which expresses a single allele of mutant BRAF^V600E under the control of CD11c-promoter driven Cre-recombinase¹³ and littermate controls (BRAF-V600E^fl/+) (Figure 1(A)). In this LCH model, the majority of DCs exhibit hyperactive ERK and mice spontaneously develop multisystem LCH-like disease.¹³ To test how this impacts the in vivo TLR4 response, LCH and WT mice were injected with a sublethal dose of LPS or vehicle 2 h before circulating serum TNFα was measured (Figure 1(B)). We observed a 400-fold increase in LPS-induced circulating TNFα in the LCH mice compared with the control mice. Both genotypes had almost undetectable levels of TNFα at baseline, suggesting that the increase in circulating TNFα is due to an increased LPS-response, and not from increased baseline levels.

FIGURE 1.

BRAF^V600E expression increases the LPS-induced TNFa response in vivo and in vitro. (A) Schematic diagram of the LCH mouse model. (B) Circulating TNFα levels measured by ELISA from serum 2 h after i.p. injection of LPS (2 mg/kg) or PBS. (C) BMDCs from WT (V600E^fl/+, Cre-negative) and LCH (V600E^fl/+, CD11c-Cre1) mice were treated with V600E-inhibitor for 1 h before whole cell lysates were immunoblotted for total and phosphorylated levels of ERK1/2 (n = 3). (D–E) BMDCs were stimulated with LPS (100 ng/ml) for the indicated time points with (right) or without (left) V600E-inhibitor pretreatment (PLX7904, 0.5 μM). TNFα in the cell supernatant was measured by ELISA (n = 3) (D) and the area under the curve was plotted (n = 3) (E). (F and G) BMDCs stimulated with LPS (100 ng/ml, 120 min) ± V600E-inhibitor pretreatment (PLX7904, 0.5 μM, 1 h) in the presence of BrefeldinA (5 μg/ml). (D) Intracellular levels of TNFα in CD11c+MHCII+ population were measured by flow-cytometry and a representative histogram of TNFa is shown. (E) The median fluorescent intensities ± sem was plotted. All data are representative of 3 independent experiments. All data were analyzed using a 2-way ANOVA

It is known that LCH mice have increased inflammatory infiltrates in multiple organs, first described by Berres et al.¹³ and replicated in our laboratory in mice as young as 6 weeks old (data not shown). These infiltrates include potent TNFα producers, including T-cells, NK cells, and macrophages. We acknowledge these possible DC-extrinsic contributors to the circulating TNFα and measured the TNFα secretion directly from cultured DCs by employing the widely used in vitro model of BMDCs.³² After 7 days in culture supplemented with GMCSF, the majority of the cell population displayed a DC phenotype, demonstrated by the expression of CD11c, MHC-II, and co-stimulatory molecules (Figures S1(A) and S1(B)). We used the specific BRAF^V600E inhibitor, PLX7904 (V600Ei), which targets BRAF^V600E at a higher affinity than BRAF^WT (IC₅₀ of BRAF^V600E is 0.0042 versus 0.14 for WT)³³ to selectively reduce the levels of phosphorylated-ERK in the BRAF^V600E-BMDCs to those of WT (Figures 1(C) and S1(C)). This system was used to measure the secretion pattern of TNFα over time in LPS-stimulated BMDCs (Figures 1(D) and 1(E)). BRAF^V600E-BMDCs secrete consistently more LPS-induced TNFα seen as early as 2 h of stimulation. Inhibition of BRAF^V600E almost completely reverses the secreted levels of TNFα to that of WT, indicating this phenotype has an acute dependence on the activated BRAF. To determine if this dependance was due to activation of the pathway downstream of BRAF, we pretreated BMDCs with a MEK inhibitor prior to measuring the LPS-induced TNFα secretion, which also demonstrated an acute reversibility (Figure S1(D)).

We repeated this experiment on cultured CD11c+ splenocytes sorted from LCH and WT mice (Figure S2). Inhibiting BRAF^V600E had no effect on the TNFα secretion of WT-DCs but reduced that of the BRAF^V600E-DCs by 1.4-fold. The less impressive reversal of secreted TNFα could be due to a fundamental difference between BMDCs and splenic DCs or to the fact that sorting splenocytes prior to culturing alters their general responsiveness to subsequent stimulation.

3.2 |. Hyperactive ERK increases TLR4-mediated protein levels of TNFα in DCs

To determine whether LPS-stimulated LCH cells have more TNFα protein on a per-cell basis, we analyzed levels of intracellular TNFα by flow cytometry in LPS stimulated CD11+MHCII+ BMDCs in the presence of Brefeldin A, a Golgi apparatus toxin that effectively prevents protein secretion (Figures 1(F) and 1(G)). The results demonstrate a uniform increase in TNFα protein production in the BRAF^V600E-BMDC population compared with the WT. Acute inhibition of BRAF^V600E reveals a partial reversibility for this phenotype. These data suggest that BRAF^V600E expression leads to increased TLR4-mediated TNFα production and secretion on a per-cell basis.

3.3 |. Hyperactive ERK decreases Tlr4 mRNA and subsequent Tnfa transcription in LPS-stimulated DCs

In parallel to the above experiments, we measured Tnfa mRNA from CD11c+ splenocytes from the LPS stimulated mice from Figure 1(B) (Figure 2(A)) and from the BMDCs at the 4-hour time point from Figure 1(D) (Figure 2(B)). We observed an increase in baseline levels of Tnfa mRNA (Figures 2(A) and 2(B)) consistent with previous microarray studies of LCH cells from human³⁴ and mice.³⁵ However, we found a consistent reduction in LPS-stimulated Tnfa transcripts in the BRAF^V600E-expressing cells compared with the WT in both models. There is no acute effect of inhibiting BRAF^V600E on the levels of Tnfa transcripts (Figure 2(B)).

FIGURE 2.

BRAF^V600E-DCs have reduced LPS-mediated TLR4 signaling and Tnfa transcription. (A) The spleens from the mice used in Figure 1(B) were subjected to magnetic bead isolation of CD11c+ cells and RNA was used for qPCR of Tnfa. A representative plot of mean ± sem from 1 of 3 independent experiments (n = 3–5) was analyzed by a 2-way ANOVA. (B) Tnfa qPCR of LPS or PBS-stimulated BMDCs (100 ng/ml, 3 h) (left), or LPS-stimulated ± V600E-inhibitor pretreatment (PLX7904, 0.5 μM, 1 h) (right). Representative plots of mean ± sem from 1 of 3 independent experiments (n = 3–5) was analyzed by a 2-way ANOVA. (C and D) BMDCs were stimulated with LPS (100 ng/ml) for the indicated time points with (right) or without (left) V600E-inhibitor pretreatment (PLX7904, 0.5 μM). Whole cell lysates were immunoblotted for total and phosphorylated levels of p65 and ERK. Representative blots (C) and normalized quantifications (D) are shown from 1 of 3 repeat experiments (n = 3). Symbols are mean ± sem. Area under the curve was calculated and p values from an unpaired t-test are indicated on the plots in (D) (ns p > 0.05; *p < 0.05). (E) Tlr4 qPCR from unstimulated BMDCs ± V600E inhibition (PLX7904, 0.5 μM, 1 h). Data plotted and analyzed as in (B)

To determine if the reduced LPS-mediated Tnfa transcription is due to a reduction in TLR4 signaling, we measured NF-κB signaling by immunoblotting whole cell lysates of BMDCs (Figures 2(C) and 2(D)). The LPS-induced phosphorylation of p65 in BRAF^V600E-BMDCs does not reach that of the WT and is not rescued upon BRAF^V600E inhibition. Consistent with this decrease in LPS-mediated signaling, qPCR analysis of RNA indicates a reduction in Tlr4 expression at baseline (Figure 2(E)). Taken together, these data demonstrate that the BRAF^V600E-mediated increase in LPS-induced TNFα protein and secretion is despite diminished Tlr4 levels and subsequent NF-κB signaling, which persists to the transcriptional level.

3.4 |. Hyperactive ERK increases the activity of TACE in DCs

To identify the mechanisms contributing to the aberrant TNFα production, we investigated ERK-driven, posttranscriptional mechanisms that could lead to this phenotype. To address the reversible increase in TNFα secretion, we looked at both cleavage of membrane bound TNFα and reuptake of soluble TNFα from BMDCs. The secretion of proTNFα from DCs requires the plasma-membrane shedding by a metalloprotease.^36–37 TACE is the major protease responsible for this cleavage,³⁸ and its activity is regulated by ERK.^39–40 To test whether the increase in secreted TNFα is due to an ERK-dependent increase in TACE activity, we used an in vitro fluorescent based TACE activity assay and detected a significant and V600Ei-mediated reversible difference in TACE activity between the 2 genotypes of BMDCs (Figure 3(A)). Our data indicate that TACE activity is significantly increased at baseline in the BRAF^V600E-BMDCs compared to the WT (p value = 0.0032), demonstrating that the presence of BRAF^V600E is sufficient to increase TACE activity in BMDCs. Upon LPS stimulation, TACE activity is increased proportionally in both genotypes, maintaining a similar difference between WT and BRAF^V600E-BMDCs (p value = 0.0039). Although the BRAF^V600E-mediated increase in TACE activity is not intensified upon LPS stimulation, our previous data indicate that the availability of its substate TNFα is. Furthermore, we showed that this increase in TACE activity is acutely reversible upon V600E inhibition. Additionally, we measured Adam17 transcripts, which encode the TACE protein from BMDCs and found an irreversible decrease compared with WT controls (Figure 3(B)). Although the Adam17 mRNA increases with V600E inhibition in the WT samples, the lack of change in the WT slope in Figure 3(A)) between the LPS and LPS + inhibitor samples suggests that there is no functional effect of the inhibitor on TACE activity levels in WT BMDCs. Taken together, this suggests that hyperactive ERK is mediating an increase in TACE-activity in the BRAF^V600E-BMDCs, despite lower levels of its transcription. Given the fully reversible nature of the increased secreted TNFα compared to the partially reversible nature of the intracellular protein levels of TNFα, we suggest that the fully reversible increase in TACE activity may likely contribute to the secretion pattern of TNFα coming from BRAF^V600E cells. Since many MMPs have also been shown to cleave proTNFα in vitro,^41–43 we measured MMP activity using the a generic MMP colorimetric assay but detected no differences between the genotypes suggesting a role for TACE activity in the amplified TNFα secretion from BRAF^V600E-DCs (Figure 3(C)).

FIGURE 3.

BRAF^V600E increases TACE activity in DCs. (A) BMDCs were pretreated with (right) or without (middle or left) V600E-inhibitor (PLX7904, 0.5 μM) for 1 h prior to a 2-h stimulation with or without LPS (100 ng/ml). Cells were then subjected to a TACE-activity fluorescent-based assay over time. The fold-change was calculated for each sample and the mean ± sem of data from 2 independent experiments (n = 3–4) were plotted and analyzed using a simple linear regression (comparison of slopes; ns p > 0.05; **p < 0.01). (B) qPCR analysis of Adam17 from BMDCs stimulated with LPS (100 ng/ml) for 3 h ± V600E-inhibitor pretreatment (PLX7904, 0.5 μM). A representative plot of mean ± sem from 3 indpenedent experiments (n = 3) is shown and analyzed by a 2-way ANOVA. (C) BMDCs stimulated with LPS (100 ng/ml) for 1 h were subjected to a generic MMP-activity assay over time (n = 3) and mean ± sem was plotted and analyzed using a simple linear regression (ns p > 0.05). (D) BMDCs were treated with 5 ng/ml of recombinant murine TNFa. After 4 h, the remaining TNFa in the supernatant was measured by ELISA and normalized to a no-cell control. The mean ± sem of % of TNF uptake was plotted from 3 independent experiments (n = 2–3) and tested with a 2-way ANOVA (interactive p value indicated in the top left corner of the plot; ns p > 0.05)

Additionally, we considered that extracellular, soluble TNFα is taken up by cells through receptor-mediated endocytosis.^44–45 To determine if a decrease in the internalization of secreted TNFα could be a contributing mechanism to the increased levels of TNFα found in the culture supernatants, we measured the percent of re-uptake from our BMDCs (Figure 3(D)). We found no significant difference in the percentage of TNFα uptake between genotypes, suggesting this mechanism does not affect the increased levels of secreted TNFα.

3.5 |. Hyperactive ERK irreversibly increases RNA translation in LPS-stimulated DCs

To address the partially reversible increase in intracellular TNFα protein in LCH cells, we considered 2 ERK-driven, posttranscriptional mechanisms: the rate of degradation of produced TNFα protein⁴⁶ and the translation of Tnfa mRNA.^47–48 To determine if a decrease in TNFα protein degradation could be contributing to the increased pool of TNFα in LCH cells, we treated LPS-stimulated BMDCs with cycloheximide to halt nascent protein translation in the presence of brefeldin A to prevent the secretion of TNFα. To calculate the total remaining TNFα protein within the cell and on the plasma membrane, we performed ELISAs of whole cell lysates. We observed no difference in the rate of TNFα degradation over time, ruling out decreased degradation as a factor (Figure 4(A)).

FIGURE 4.

BRAF^V600E-BMDCs increase translation initiation and elongation. (A–E) BMDCs were stimulated with LPS (100 ng/ml) for 3 h ± V600E-inhibitor pretreatment for 1 h (PLX7904, 0.5 μM). (A) BMDCs were then treated with cycloheximide (CHX) + Brefeldin A (5 μg/ml) for indicated time points. TNF from 40 μg of whole cell lysates was measured by ELISA. Mean ± sem of fold-changes from 3 repeat experiments (n = 3) were plotted and fit by nonlinear regression (comparison of K values; ns p > 0.05). (B and C) BMDCs were subjected to polysome profiling. (B) A representative plot of RNA concentrations per fraction. (C) qPCR analysis of tnfa from isolated total and pooled-polysomal RNA (≥2 polysomes) was used to calculate the ratio of polysomal-bound tnfa mRNA. Mean ± sem of fold changes were plotted from 1 of 2 independent experiments (n = 2–3) and analyzed by a 2-way ANOVA and Sidak’s multiple comparison test (adj p value for V600E; **p < 0.01). (D) Concentration of ribosomal subunits was measured from total RNA using a BioAnalyzer and the per-cell number of 18S and 28S subunits from a representative experiment of 3 repeats (n = 2–3) was plotted and analyzed by a 2-way ANOVA (interaction p value; ns p > 0.05). (E) Whole cell lysates were immunoblotted for total and phosphorylated eEF2. A representative blot is shown (top) and quantified mean ± sem from 1 of 2 repeat experiments (n = 3–5) was plotted (bottom) and a mixed-effects analysis was applied (interaction p values; *p < 0.05; **p < 0.01). (F and G) BMDCs pretreated with or without V600E-inhibitor (PLX7904, 0.5 μM) were treated with harringtonine (2 μg/ml) for the indicated times before the addition of puromycin (10 μg/ml, 10 min). Cells were subjected to flow cytometry and gated on CD11c+MHCII+. (F) Representative puromycin histograms from 1 of 3 repeat experiments (n = 3). (G) Combined puromycin MFIs from 3 independent experiments were plotted and fit by nonlinear regression analysis (comparison of k values; **p < 0.01)

Translational efficiency (TE) is defined as the ratio of ribosomal bound mRNA to total mRNA and has been described as a good proxy for translatability.⁴⁹ To analyze Tnfa TE, we measured the number of polysome-bound Tnfa transcripts by performing polysome profiling, which fractionates transcripts based on the number of bound ribosomes. Higher numbers of bound polysomes on a transcript indicates faster translation since more ribosomes are working on the translation at a time.⁵⁰ Cell lysates of equal RNA content were fractionated in a sucrose gradient, and the UV absorbance of each fraction was recorded. Our polysome profiling data indicate a remarkable, and irreversible increase in polysome-bound transcripts in the BRAF^V600E-BMDCs compared with the WT (Figure 4(B)). To confirm that Tnfα is one of those transcripts, we measured Tnfa transcripts by qPCR from total and pooled-polysomal RNA and calculated about a 30-fold increase in the ratio of polysome-bound Tnfa transcripts in BRAF^V600E-BMDCs compared with WT (Figure 4(C)), confirming that translation efficiency of Tnfa is increased. Two-way ANOVA testing reveals a small, statistically significant interaction of genotype on the effects of inhibitor, but clearly not enough in magnitude to reverse the mutant phenotype to that of WT.

Since ERK signaling is involved in the complex upstream network of initiation and primarily plays a role in regulating ribogenesis,^51,52 we questioned whether the number of rRNAs could be altered in the BRAF^V600E-BMDCs. We calculated the number of 18S and 28S rRNAs per cell but found no difference in the availability of ribosomes between genotypes (Figure 4(D)). We considered the role that mTOR could be having on initiation, since it is recognized as the master regulator of translation initiation.⁵³ We found a slight increase in phosphorylation of mTOR protein p70S6K in BRAF^V600E-BMDCs compared with WT by western blot, suggesting a potential role for mTOR in the increased polysome occupancy (Figure S3). Together, these data demonstrate an increase in the engagement of ribosomes to transcripts in the BRAF^V600E-BMDCs compared with WT, indicating an increase in translation that is insensitive to acute V600E inhibition.

3.6 |. Hyperactive ERK increases translation elongation in LPS-stimulated DCs

The rate of ribosomal movement along bound mRNA contributes to the rate of protein production. One of the main translation elongation factors in eukaryotes is eEF2, which catalyzes the GTP-dependent translocation step during translation elongation to help push ribosomes along mRNA. The phosphorylation of eEF2 prevents it from binding with ribosomes, thus preventing elongation.⁵⁴ The only known kinase of eEF2 is eEF2K, which is modulated by ERK through p90^RSK1.⁵⁵ We demonstrated a decrease in phosphorylated-eEF2 in the BRAF^V600E-BMDCs compared with WT both at baseline and after LPS stimulation (Figure 4(E)), indicating a derepression of eEF2 activity. BRAF^V600E inhibition rescues this phosphorylation after LPS stimulation.

To functionally measure translation elongation, we used SunRiSE, a puromycin-based flow cytometry assay to calculate the rate of ribosomal run-off from mRNA, described by Argüello et al.⁵⁶ By treating cells with harringtonine, an inhibitor of translation initiation, for various time points prior to puromycin addition, we can measure the time it takes currently bound ribosomes to release their transcript. The decrease in puromycin signal over time of harringtonine treatment directly correlates with the speed of ribosomal run-off, and thus rate of elongation. Intracellular flow cytometry of these BMDCs demonstrates that BRAF^V600E expression increases the rate of puromycin decay (Figures 4(F)–4(G)). The 1.4-fold decrease in puromycin half-life suggests faster elongation in the BRAF^V600E-BMDCs, which is completely rescued after acute BRAF^V600E inhibition. Taken together, these data suggest a multistage increase in translation, with only elongation being reversible. This combination is likely responsible for the partial reversal of intracellular TNFα protein levels by BRAF^V600E inhibition (Figures 1(F) and 1(G)).

3.7 |. Hyperactive ERK selectively increases translation of inflammatory programs in LPS-stimulated DCs

We used a high-throughput approach to gain a better understanding of the global impact of the increased translation in LCH cells compared with WT. Using RNA-seq data from both polysome fractions and total RNA, we utilized the Bioconductor R package, Anota2Seq,⁵⁷ to identify the significant differences in the translatome between genotypes after LPS stimulation. Given the impressive increase in polysome bound mRNA depicted in the polysome profiling (Figure 4(B)), we were not surprised that Anota2Seq identified a large number of genes considered to have increased translation in LCH cells and not WT (Figure 5(A), dark blue, pink, and light green; 1582 genes). However, the data suggest a large number of those transcripts (dark blue; 69.2% of genes with increased translation) are offset by a decrease in transcription and are predicted to have no change in protein levels, for example, “buffered” mode. The remaining 487 transcripts with increased translation are predicted to yield an increase in protein levels (pink and light green); however, 210 genes (light green; 13.3% of genes with increased translation) can be attributed to an increase in mRNA transcription, e.g., “transcription up” mode. Thus, only 277 transcripts (pink; 17.5% of genes with increased translation) can be attributed to an increase in translation alone, for example, “translation up mode.”

FIGURE 5.

BRAF^V600E promotes translation of inflammatory mediators by selective and irreversible polysome binding (A) Scatter plot of log2 fold changes for polysome-associated mRNA (Y-axis) versus total mRNA (X-axis). Colors designate differentially regulated transcripts through translation, buffering, or abundance according to the anota2seq analysis of the inhibited contrast and is quantified in the bar graph (bottom). Unchanged mRNAs are shown in grey (n = 3). (B and C) Translationally up-regulated genes from the uninhibited and inhibited contrasts were combined and submitted for GO term (biologic process) enrichment analysis through DAVID. An FDR cutoff of 5% was applied before plotting the fold enrichment. [count, FDR]. For (B), genes from the “translation up” list that are associated with the GO terms are displayed to the right. (D) The translationally up-regulated gene list was uploaded to the AURA2 “batch mode” online mouse database using the “regulatory element enrichment” analysis mode. The resulting list was plotted after a 1% FDR cutoff was applied

We applied the Anota2Seq algorithm to compare WT and BRAF^V600E-BMDCs both with and without the V600E-inhibitor. Given the lack of V600E-inhibitor effects on the polysome profiling of both WT and BRAF^V600E-BMDCs (Figure 4(B)), we were not surprised to find little differences between ANOTA2Seq outputs when comparing the effect of inhibitor within each genotype (Figures S4(B) and S4(C)). The lack of significantly differentially regulated transcripts plotted in Figures S4(B) and S4(C) indicates that V600E inhibition does not acutely effect polysome occupancy of specific transcripts for both WT and BRAF^V600E-BMDCs.

As noted above, when comparing WT with BRAF^V600E-BMDCs, ANOTA2Seq identified many genes with statistically significant increases in “translation up” mode independent of inhibitor status (Figure 5(A)). With V600E-inhibiton, ANOTA2seq indicated a larger number of genes in the “translation up” mode (277 genes) (Figure 5(A)) compared with without V600E-inhibiton (33 genes) (Figure S4(A)). Given the lack of V600E-inhibitor effect on the regulation of transcripts within each genotype, we were originally surprised by this distinction when comparing between genotypes. Upon further investigation, we realized that 79% of the “translation up” gene list from the uninhibited comparison (Figure S4(A)) appears on the “translation up” gene list of the V600E-inhibited comparison (Figure 5(A)). The significant overlap suggests that V600E inhibition had not significantly altered the content of the “translation up” genes. Many of the differences between gene lists could be accounted for by changes in the adjusted p value cut-off rather than biologic alterations to the translatome. For instance, without inhibitor, TNFα is in the “translation up” gene list with a significant unadjusted p value (p = 0.03) but fell just below the threshold when adjusted with Benjamini–Hochberg correction. With inhibitor, TNFα remained significant even after correction. This is consistent with the significant increase in Tnfa transcription efficiency measured by qPCR from both of these samples (Figure 4(C)). Given that a direct comparison of the effect of inhibitor showed essentially no translational differences for each genotype (Figures S4(B) and S4(C)), and the fact that the content of the “translation up” gene lists between genotypes were similar with and without inhibitor, we decided that using the list of genes that met the adjusted p value cutoff for WT versus LCH from either “with” or “without” inhibitor conditions would most accurately depict the differences in the LCH translatome. This combined list of “translation up” genes was used for all subsequent analysis and referred to as the “LCH translatome” in our model.

To understand the impact of the altered LCH translatome on biology, we searched for enrichment of biologic functions defined by Gene Ontology (GO) classifications on the DAVID platform^58–59 within the “translation up” gene list. Using a 5% FDR cutoff, we identified 3 GO terms all associated with immune processes and listed the identified genes for each term (Figure 5(B)). This suggests an increase in translation efficiency of inflammatory mRNAs. In particular, the term “response to LPS” confirms our hypothesis that BRAF^V600E amplifies the TLR4 response in DCs, which is mediated by an increase in translation of TLR4-induced genes.

We compared associated GO terms of the transcriptome to the translatome to determine if traditional transcriptome analysis would have generated similar results. We used DESeq2 to detect differentially expressed genes (DEGs) (fold-change > 1.5, FDR cutoff 15%) from the total mRNA, which generated a list of 346 up-regulated genes. Enrichment analysis of the up-regulated DEGs indicated different GO terms compared with that of the translatome and unrelated to the LPS response (Figure 5(C)). This suggests that BRAF^V600E amplifies the LPS response posttranscriptionally, in a means not discernible by traditional transcriptional analysis, by selectively up-regulating the translation of inflammatory transcripts.

To identify common upstream regulators whose activity could be responsible for this selective up-regulation of translation, we submitted the list of genes in “translation up” mode to Aura2, which exploits various data mining tools to identify known binding elements of transcripts and detects enrichments within a list of genes.⁶⁰ Using the mouse database and an FDR cutoff of 1%, 7 RNA binding proteins were identified (Figure 5(D)). As a control, we also submitted the list of genes in “buffered mode,” and Aura2 did not identify any common binding elements from this list, further suggesting that these 7 RNA binding proteins are specifically interacting with the purely translationally controlled transcripts. Interestingly, the top Aura2 hit from the translation mode analysis, Ezh2, has increased protein levels in many histiocytoses, including LCH. Tian et al.⁶¹ demonstrated a correlation in staining of Ezh2 with increased phosphorylation of ERK1/2 but not to myc or p-stat, indicating a role for ERK1/2 in regulation of Ezh2 protein expression. Although Ezh2 is typically known as a lysine methyltransferase and transcriptional repressor, a recent paper described its role in accelerating ribosome function as well as contributing to IRES-dependent translation initiation in cancer cells.⁶² Additionally, 4 out of the 7 regulatory proteins include Ser/Arg-rich splicing factors (SRSFs), which regulate multiple aspects of the gene expression program. Boutej et al.⁶³ show that LPS-stimulated microglia rely on SRSF3 activity to repress innate immune translation, and that silencing SRSF3 increased the protein translation of multiple inflammatory mediators. Together with our data, these suggest a role for Ezh2 and SRSF3 in LCH-mediated translation of inflammatory programs.

4 |. DISCUSSION

Many laboratories have studied the effects of BRAF^V600E and hyperactive ERK on malignancies arising from varied cell types. Thus, much of our knowledge about the pathogenic effects of BRAF^V600E relates to the biology of proliferation and survival. However, the effects of ERK activity are cell context specific. The inflammatory nature of LCH and the fact that the disease is associated with BRAF^V600E mutation in DCs, an inflammatory cell, led us to hypothesize a role for hyperactive ERK directly in inflammatory functions. The work detailed here demonstrates how hyperactive ERK dysregulates a critical inflammatory function of DCs, the response to pathogenic cues. By using TNFα production in BRAF^V600E-expressing DCs as a readout for the response to LPS stimulation, we discovered an amplified posttranscriptional response, resulting in increased TNFα despite a reduced TLR4 signal. While TLR4 and TNFα signals may not be directly related to in vivo LCH pathogenesis, this provides a tractable model in which to explore mechanisms for ERK-directed posttranscriptional control of inflammatory programs. We identified 2 ERK-dependent mechanisms at play. We discovered an increase in TACE activity, which likely contributes to the increase in secreted TNFα (Figures 3(A) and 3(B) and 1(D) and 1(E)). We also observed an increase in protein translation, which likely contributes to the increase in intracellular TNFα protein we observed (Figures 4 and 1(F) and 1(G)). These data confirm our hypothesis that BRAF^V600E enhances the inflammatory response of DCs to stimuli. We consider that this phenotype may not be specific to DCs, but rather any cell that has the potential to respond to LPS and produce TNFα, such as macrophages. Further studies are required to investigate this phenotype in other cell types and which types of stimuli impact LCH pathogenesis, and to determine if this translates in humans.

ERK activity results in the phosphorylation of TACE at multiple sites, which leads to its activation and relocation to the plasma membrane.⁴⁰ Our BRAF^V600E-BMDCs with hyperactive ERK showed a significant increase in TACE activity, which was reversed upon acute BRAF^V600E inhibition. In addition to TNFα, TACE is also known to cleave over 80 different substrates related to immune processes, development and differentiation, and cell adhesion.^38,64 Although we did not measure these other proteins directly, we hypothesize that their rate of cleavage might also be increased in an ERK-dependent manner in LCH cells. Overactive TACE activity has been observed and reported to contribute to various inflammatory disorders and cancers.^64,65 Not surprisingly, there have been efforts to target TACE as a therapeutic, but safety and toxicity concerns have prevented its clinical progression.⁶⁶

Increased protein translation mediated by BRAF^V600E is also not unique to LCH cells. Although BRAF^V600E-mediated translational increases have been reported in cancers such as melanomas,^67–68 the cell-specific consequences of increased translation have not been fully explored. Unlike melanocytes, which experience increased rates of proliferation in response to the BRAF^V600E-mediated aberrant translation,⁶⁹ it is well established that LCH cells do not hyper-proliferate.^13,27,35,70 Consistent with the dynamic nature and specialized immune functions of DCs, our data demonstrate that the BRAF^V600E-induced translatome is skewed to promote an amplified response to TLR4 stimulation. Our polysome profiling data alone suggest a broad increase in translation efficiency. However, our ANOTA2Seq data suggest that not all of those polysome-bound transcripts would be predicted to yield an increase in protein, as most of the polysome associated transcripts are accompanied by a reduction in transcription predicted to buffer the subsequent protein levels. The Aura2 analysis points to a few RNA-binding proteins that could influence the selection of transcripts that are predicted to yield increases in protein levels. Further studies are required to verify the roles of those proteins in driving the LCH translatome.

Our data demonstrate that the acute reduction of ERK activity in LCH cells does not affect the increased number of ribosomes occupying transcripts (Figures 4(B) and 4(C)). We consider the relatively short duration of the V600E inhibition of the BMDCs and postulate that the mechanism responsible for the increased translation may eventually be sensitive to BRAF^V600E activity; however, it is not captured within our experimental setup. Additionally, an alternate pathway may be contributing to translation initiation. Given mTOR’s well-known role in regulating translation, it is likely that the increased mTOR activity we observed in BRAF^V600E-BMDCs (Figure S3) contributes to the increased polysome binding. A recent study⁷¹ demonstrated that mTOR inhibition by Rapamycin improved overall disease state in an inducible mouse model of LCH. Given our data, and mTOR’s well-established role in translation, one interpretation is that the Rapamycin treatment reduced translation initiation, contributing to the observed improvement in disease state. Additional studies are required to determine the specific mechanisms responsible for the increased translation initiation and their sensitivity to V600E inhibition. On the other hand, ERK inhibition was sufficient to reduce the increased speed at which ribosomes elongate (Figures 4(F) and 4(G)). This is consistent with the fact that ERK regulates eEF2K activity, which controls elongation through eEF2. Without BRAF^V600E inhibition, these cells have increased polysome occupancy and elongation rates leading to increased production of inflammatory proteins. When BRAF^V600E is inhibited, the combination of persistent increased polysome occupancy and the decelerated elongation rate likely results in a partial reduction in protein production, as demonstrated by partially reduced levels of intracellular TNFα (Figures 1(F) and 1(G)).

In a time when next-generation sequencing and transcriptomic data are commonplace in medical research and informing clinical decision making,^72–73 this work reinforces the importance of complimentary proteomic data. In this particular example, uncovering the LCH translatome highlights a BRAF^V600E-phenotype that is not acutely reversible, translation initiation, and opens the door to a number of related questions. The current model of LCH pathogenesis, the “misguided myeloid dendritic cell precursor” model, proposes that the stage of cell differentiation when a somatic activating mutation is acquired, determines the organs inflicted and extent of clinical manifestations.^6,70 It has been suggested that hyperactive ERK leads to the generation of a senescence-associated secretory phenotype (SASP) in multipotent hematopoietic progenitor cells, which may lead to inflammation.⁷¹ However, we do not yet have a molecular explanation as to how hyperactive ERK causes misguided DC development and the development of SASP in the first place. Altered translation as a result of excessive ERK activation in DC precursors could represent the molecular link to these downstream events. Future work should investigate how altered translation might be the first step in the misguided myeloid DC precursor model.

Abbreviations:

Adam17

A disintegrin and metalloproteinase 17

Brsoe

based Ribosome Speed of Elongation

BMDCs

bone marrow derived dendritic cells

BRAF

v-raf murine sarcoma viral oncogene homolog B1

DC

dendritic cells

DEG

differentially regulated genes

EEF2

Eukaryotic Translation Elongation Factor 2

EEF2K

EEF2 kinase

EZh2

Enhancer of zeste homolog 2

LCH

Langerhans cell histiocytosis

MMP

matrix metalloprotease

MNP

mononuclear phagocytic system

mTOR

mammalian

p70S6K

70-kDa ribosomal protein S6 kinase

p90RSK1

p90 ribosomal S6 kinase

Ras

Rat sarcoma (family of proteins)

SASP

senescence associated secretory pathway

SRSF

serine/arginine-rich splicing factor

SunRise

SUnSET

SUnSET

Surface sensing of translation

TACE

TNFα converting enzyme

T-cells

T lymphocytes

TE

translation efficiency

V600E

valine to glutamate substitution at amino acid site 600

WT

wild type