Oncogenic BRAF V600E induces glial proliferation through ERK and neuronal death through JNK

Background: Activating V600E in BRAF is a common driver mutation in cancers of multiple tissue origins, including melanoma and glioma. BRAF V600E has also been implicated in neurodegeneration. The present study aims to characterize BRAF V600E on cell death and survival in three major cell types of the CNS: neurons, astrocytes, and microglia. Methods : Multiple primary cultures and cell lines of glial cells and neurons were employed. BRAF V600E as well as BRAF WT expression was mediated by lentivirus or retrovirus. Blockage of downstream effectors were achieved by siRNA. Gene expression data from patients with Parkinson’s disease was analyzed. Results : In astrocytes and microglia, BRAF V600E induces cell proliferation, and the proliferative effect in microglia is mediated by activated ERK but not JNK. Conditioned medium from BRAF V600E -expressing microglia induced neuronal cell death. In neuronal cells, BRAF V600E directly induces cell death, through JNK but not ERK. We further show that BRAF-related genes are enriched in pathways in patients with Parkinson’s disease. Conclusions : Our study identi�es distinct consequences mediated by distinct downstream effectors in dividing glial cells and in neurons following the same BRAF mutational activation and a causal link between BRAF-activated microglia and neuronal cell death that does not require physical proximity. It provides insight into a possibly important role of BRAF in neurodegeneration as a result of either dysregulated BRAF in neurons or its impact on glial cells.


Background
The BRAF (v-Raf murine sarcoma viral oncogene homolog B) gene encodes a serine-threonine kinase.BRAF is ubiquitously expressed, with high levels in neuronal tissues [1].As part of the Ras-Raf-MEK-ERK pathway, BRAF is activated by RAS at the upstream, and it in turn phosphorylates and activates mitogen activated protein kinase (MEK) kinase at the downstream.MEK then phosphorylates and activates extracellular signal-regulated kinase (ERK).The Ras-Raf-MEK-ERK signaling cascade transduces extracellular signals to the cytoplasm and nucleus and is involved in the regulation of cell survival, growth, proliferation, and differentiation in developing and adult tissues, such as the central nervous system (CNS).In non-neuronal cells, including glial cells, BRAF stimulates cell proliferation [2].
BRAF signaling is both necessary and su cient to promote astrocyte proliferation during development [3].Mutations of BRAF, predominantly valine-to-glutamic acid mutation at residue 600 (V600E) lead to constitutive kinase activation and unregulated cell growth and proliferation.BRAF V600E is one of the most prevalent oncogenic mutations in human cancers, accounting for more than half melanomas and many other cancers including a variety of pediatric and adult gliomas [4].Accordingly, BRAF inhibitors have been developed as a therapy for cancer [5].
The essential role of BRAF in neuronal development and survival has also been supported by a number of studies [6,7].BRAF is essential for survival of embryonic neurons in culture [1].Homozygous BRAF null mice are embryonic lethal due to growth retardation and neuronal defects [8].Mice expressing nonfunctional BRAF in neural stem cell-derived brain tissue displayed disturbed neuronal circuits in the cerebellum and hippocampus, along with impaired neuronal generation [9].Forebrain-speci c knockout of BRAF caused impaired hippocampal learning and memory [10].Further, BRAF controls synaptic transmission and learning behavior [11].Mutant BRAF V600E in neural progenitors resulted in a hyperexitable phenotype in neocortical pyramidal neurons [12], and it contributes to intrinsic epileptogenicity without the formation of any neuronal tumors, whereas high proliferation of glial lineage cells contributes to the tumorigenity of the mutation [13].Constitutive activation of BRAF by low-level expression of BRAF V600E in the mouse germline is associated with increased number of astrocytes with unaltered neuron number in hippocampal area [14].
More recent evidence implicates that BRAF activity may be involved in neuronal cell death, directly or indirectly.BRAF V600E somatic mutation in the erythro-myeloid progenitors, which give rise to tissueresident myeloid cells caused clonal expansion of microglia and subsequently caused synaptic loss and neuronal death in adulthood [15].The degenerative changes can be prevented by BRAF inhibition [16].
Here we report distinct consequences and downstream effectors following BRAF V600E expression in primary cultures and cell lines of three major cell types of the CNS: neurons, astrocytes, and microglia.We also show indirect effects of BRAF V600E on neurons through astrocytes and microglial cells.

Human neuroblastoma cell line SH-SY5Y cells
SH-SY5Y cells (ATCC, Cat# CRL-2266, USA) were seeded at 80% density, the medium was changed to DMEM/F12 supplemented with 1% FBS and 10 mM retinoic acid (Sigma, USA) for differentiation for 3 days [17].The medium was then replaced with conditioned medium from BV2 cells transduced with different vectors for 24 h.Cells were then collected at different time points for subsequent experiments.

Primary cortical mixed culture
Primary cortical mixed culture was prepared from C57BL/6J mice embryonic day 17.To obtain cortical culture, pregnant mice (the Jackson Laboratory) were anesthetized, embryos were dissected, and cortex was collected in PBS on ice.Tissue was incubated in 0.25% trypsin-EDTA (Gibco) at 37°C for 15 min.Trypsinized tissue was transferred into high-glucose DMEM/F12 medium supplemented with 10% FBS.
After centrifugation (1000×g, 5 min), the pellet was resuspended [17,18].Cells were plated onto poly(Llysine)-coated 24-well plates at 10 6 cells per well and cultured in NB-A (Gibco).Cortical mixed cells were cultured for 5 days, then transduced with lentivirus (LV) vectors or retrovirus (RV) vectors for 24 h, and then cultured in NB-A medium for 96 h.

Primary cortical neuronal culture
Primary cortical neurons were prepared from C57BL/6J mice embryonic day 17.The dissected cortical tissue was digested, triturated and centrifuged as for cortical mixed culture.Cells were plated onto poly (L-lysine)-coated 24-well plates at 10 6 cells per well and cultured in NB-A with 2% B27 (Invitrogen, USA).After 24 h in culture, 5ug/ml cytarabine was added to inhibit the growth of glial cells in the medium, and then changed to the original medium 48 h later.Neurons were cultured for 5 days, and then transduced with BRAF V600E , or BRAF WT , or vector lentivirus plus 8 μg/ml polybrene (Sigma-Aldrich, USA) for 24 h.After transduction, cells were cultured in NB-A for 120 h and then collected for subsequent experiments.

Primary microglial culture
Mixed glial cells were obtained from C57BL/6 mice embryonic day 17.Cells were cultured in high-glucose DMEM/F12 supplemented with 10% FBS in humidi ed air containing 5% CO2 at 37°C.The culture medium was replaced with fresh medium 24 h after the initial preparation and every 3 days thereafter.After 1 week, microglial cells were obtained by mechanical shaking of the mixed glial cell cultures for 1 h.Cells were routinely monitored for purity by ionized calcium binding adaptor molecule 1 (Iba1) staining and the population of Iba1+ cells was >95% [19,20].Microglial cells were seeded at 80% density, then transduced with BRAF V600E , or BRAF WT , or vector lentivirus plus 8 μg/ml polybrene for 24 h.After transduction, the cells were cultured for 120 h in high-glucose DMEM/F12 with or without FBS, and the medium as well as the cells were collected for subsequent experiments.

Primary astrocyte culture
Primary astrocytes were obtained from C57BL/6 mice embryonic day 17.The dissected cortical tissue was digested, triturated and centrifuged as for neuronal cultures.The cell pellet was resuspended in highglucose DMEM/F12 supplemented with 10% FBS.Cells were seeded in poly-L-ornithine-coated Petri plates.Nonadherent cells were removed after 20 min [21].Astrocytes were transduced with BRAF V600E , or BRAF WT , or vector lentivirus plus 8 μg/ml polybrene for 24 h.After transduction, the cells were cultured for 120 h in high-glucose DMEM/F12 with or without FBS, and the medium as well as the cells were collected for subsequent experiments.All experiments were performed in accordance with a protocol approved by Institutional Animal Care and Use Committee at Massachusetts General Hospital.

siRNA transfection
Primary neurons and microglial cells were seeded in 6-well plates and transfected with ERK1 siRNA (Cell Signaling Technology, USA), ERK2 siRNA (Cell Signaling Technology), SAPK/JNK siRNA (Cell Signaling Technology) at 80% cell con uence on day 4 in culture.ERK1 siRNA, ERK2 siRNA, JNK siRNA and Lipofectamine RNAiMAX Transfection Reagent (Thermo Fisher Scienti c) were diluted in OPTI-MEM medium (Thermo Fisher Scienti c), mixed gently, and incubated for 5 min to allow complex formation.The cells were transfected by adding the RNAi-Lipofectamine complex dropwise to medium to achieve a siRNA concentration of 100 nmol/l.The cells were collected 48 h after the transfection for subsequent experiments [22].

Cell cycle determination
Cells were seeded in 6-well plates at 70% con uence, followed by virus transduction and siRNA treatment for 3 days.The cells were collected, xed in 70% ethanol overnight, and stained with Tali Cell Cycle Kit (Thermo Fisher Scienti c).We used the Attune NxT Flow Cytometer (Thermo Fisher Scienti c) to determine the percentage of cells in each cell cycle phase.Flowjo software was used to analyze the data.

NO and LDH and MTS assays
Griess Reagent Kit (Thermo Fisher Scienti c) was used to assess NO according to the manufacturer's instructions.Cells were seeded in 96-well plates.We mixed the following reagents in each well: 20 µl Griess Reagent, 150 µl nitrite-containing sample, and 130 µl deionized water.A photometric reference was prepared by mixing 20 µl Griess Reagent and 280 µl deionized water.Absorbance of the nitrite-containing samples relative to the reference was measured in Synergy 2 Multi-Mode Microplate Reader (Bio-Tek, USA) at 548 nm.
For LDH assay, Pierce LDH Cytotoxicity Assay Kit (Thermo Fisher Scienti c) was used.50 µl cell culture medium was transferred to a new 96-well plate, 50 µl reaction mixture was added to each well and incubated at room temperature for 30 min, then 50 µl stop solution was added.Absorbance at 490 and 680 nm was measured using Synergy 2 Multi-Mode Microplate Reader (Bio-Tek , USA).
For MTS assay, MTS Cell Proliferation Colorimetric Assay Kit (BioVision, USA) was used.Cells were seeded in 96-well plates and treated as indicated above.After the treatment, cell culture medium of each well was removed and 110 µl reagent containing 10 µl MTS and 100 µl medium was added to each well and incubated in humidi ed air containing 5% CO2 at 37°C for 1 h.Absorbance at 490 nm was measured using a microplate reader.

Immunostaining and quantitative analysis
Cells were seeded in 24-well plates.After treatment, the medium was removed and the plates were washed with PBS three times, the slides were xed with 4% paraformaldehyde for 15 min and washed with PBS three times, and permeabilized with 0.5% Triton X-100 at room temperature for 20 min.Plates were washed with PBS.5% goat serum was added and incubated at room temperature for 30 min.Primary antibodies were added and incubated at 4°C overnight.Primary antibodies used were microtubule-associated protein 2 (MAP2) (1:500, Thermo Fisher Scienti c), Iba1 (1:500, Abcam, USA), Ki-67 (1:100, Thermo Fisher Scienti c), and glial brillary acidic protein (GFAP) (1:1000, Sigma-Aldrich,) Plates were washed with PBS and incubated with secondary antibody: goat anti-rabbit IgG-Alexa Fluor 488 (1:500, Molecular Probes, USA) or donkey anti-mouse IgG-Alexa Fluor 546 (1:500, Thermo Fisher Scienti c) for 45 min in the dark at room temperature, washed with PBS, and stained with DAPI (Vector Laboratories, USA).For analysis of uorescence intensity, images were captured under an Olympus BX51 microscope (Olympus Optical Co., Tokyo, Japan) with Olympus CAST stereology software and a DP 70 digital camera system using the same camera gain, exposure time and pixel setting for all samples.Integrated optical density (IOD) of staining in the images was analyzed by Image J.

RNA extraction and real-time quantitative PCR analysis
RNA was isolated using the TRIzol LS Reagent (Thermo Fisher Scienti c).SuperScript III First-Strand Synthesis System (Thermo Fisher Scienti c) was used for cDNA synthesis.All primers were obtained from Massachusetts General Hospital DNA Core and are listed in Table 1.StepOnePlus Real-Time PCR System (Thermo Fisher Scienti c) was used.Relative gene expression levels were analyzed using the 2-ΔΔCt method.GAPDH was used as internal control.
Table 1 The sequences of primers.BCA Protein Assay Kit (Thermo Fisher Scienti c).Equal amounts of protein (30 mg) were loaded onto SDS-PAGE.Following separation, proteins were transferred from the gels to methanol-activated polyvinylidene uoride (PVDF; Sigma-Aldrich) transfer membranes, blocked with 5% BSA for 1 h at 37°C.The membranes were incubated with primary antibodies overnight at 4°C.Subsequently, the membranes were incubated with an appropriate horseradish-peroxidase-labeled secondary antibody for 1 h at room temperature.Proteins were visualized using an ECL system (Thermo Fisher Scienti c).The expression levels of phosphorylated proteins were quanti ed by normalizing to their respective total proteins.Data was analyzed using ImageJ software.

ELISA
The release of Interleukin 1β (IL-1β), Interleukin 6 (IL-6) and tumor necrosis factor α (TNF-α) in mouse primary microglial cell culture medium was measured with commercial ELISA kits-Mouse IL-1β (R&D Systems, USA), mouse IL-6 (R&D Systems), and mouse TNF-α (R&D Systems).Cells were seeded in 6-well plates and cultured with DMEM/F12 supplemented with 10% FBS.After virus transduction, the cells were cultured for 24 h in DMEM/F12 without FBS, and the medium was collected, centrifuged at 3000 rpm for 10 min, and the supernatant was subjected to ELISA.Absorbance at 450 nm was measured using Synergy 2 Multi-Mode Microplate Reader (Bio-Tek).IL-1β, IL-6 and TNF-α concentrations were calculated based on the standards and normalized with total protein concentration of the same sample.

Statistical analysis
All experiments were repeated at least three times with at least three replicates within each condition.
Investigators who performed assays were blind to treatment groups.All values are presented as the mean ± SEM.Group comparisons were performed using ANOVA and Tukey's post hoc test.P < 0.05 was considered statistically signi cant.Postmortem SN (substantia nigra) gene expression pro le from patients with Parkinson's disease and function enrichment analysis of BRAF-related genes GDS2821 gene expression data from 16 Parkinson's disease patients was obtained from the National Center for Biotechnology Information's (NCBI) Gene Expression Omnibus (GEO: https://www.ncbi.nlm.nih.gov/geo).Raw data were normalized and processed by R software (version 3.3.3)with the limma package [23] and the normalized gene expression levels were presented as log2-transformed values.The Spearman coe cients of expression of genes and BRAF were calculated, whilst the expression genes with P < 0.05 were de ned as BRAF-related genes.The Database for Annotation, Visualization, and Integrated Discovery (DAVID, http://david.ncifcrf.gov),an online functional annotation tool were applied for gene ontology (GO) enrichment analyses including biological process (BP), cellular component (CC) and molecular function (MF), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis [24].GO terms or KEGG pathways with P < 0.01 were considered statistically signi cant.

Results
BRAF V600E expression leads to neuronal cell death and glial cell proliferation in primary mouse cortical mixed culture Primary cortical mixed culture composing of neurons and glial cells (astrocytes and microglia was transduced with LV expressing human BRAF V600E , or human BRAF WT , or control vector.BRAF overexpression was detected by Western blotting, which showed signi cantly increased BRAF in transduced cells than vector or non-transduced control cells (Figure 1A).Immunostaining for MAP2, a marker of neurons and Iba1, a marker for microglia 120 h after the transduction showed that there was a decrease in the number of neurons and an increase in the number of microglia in the mixed culture transduced with BRAF V600E than vector or BRAF WT (Figure 1A).Similarly, immunostaining for MAP2 and GFAP, a marker for astrocytes showed that there was a decrease in the number of neurons and an increase in the number of astrocytes in the culture transduced with BRAF V600E than vector or BRAF WT .
BRAF WT did not change the number of either MAP2+ or iba1+ or GFAP+ cells (Figure 1C).These results suggest BRAF V600E expression in the cortical mixed culture may lead to different outcomes for neurons and glial.
We next transduced the mixed cultures with BRAF V600E or control vector by RV.Unlike LV, which is capable of infecting both neurons and glial cells, RV infects only dividing cells, not postmitotic neurons.After 120 h, MAP2 and Iba1 staining showed an increase in the number of microglia and a decrease in the number of neurons in RV BRAF V600E -transduced culture than RV vector-transduced culture (Figure 1B).BRAF V600E transduction resulted in an increase in the number of astrocytes and a decrease in the number of neurons than RV vector-transduced cells, which were demonstrated by MAP2 and GFAP staining (Figure 1D).These ndings suggest that the mutant BRAF may cause neuron loss through proliferation of microglia and/or astrocytes.

BRAF V600E expression in astrocytes promotes proliferation
To investigate BRAF V600E effect on individual cell types, we next transduced primary astrocytes prepared from C57BL/6 mouse embryos with lentiviral BRAF V600E , BRAF WT and control vector, respectively.Western blotting showed that, 120 h after the transduction, expression levels of mitogen-activated protein (MAP)/phosphorylated MEK1/2 (p-MEK1/2), phosphorylated ERK1/2 (p-ERK1/2) in BRAF V600Etransduced astrocytes were increased compared with the vector-transduced or non-transduced control groups, suggesting activation of the ERK pathway by BRAF V600E (Figure 2A).Flow cytometry showed that there was a lower percentage of cells in G0/G1 phase and a higher percentage of cells in S phase in BRAF V600E -transduced astrocytes as compared with vector or non-transduced control groups (Figure 2B).
Despite an increase in BRAF expression by nearly 3-fold, BRAF WT did not change cell cycle distribution in cultured primary astrocytes, which can be explained by approximately 500-fold greater kinase activity of BRAF V600E than BRAF WT [25].MTS assay revealed that astrocyte viability was increased signi cantly 96 and 120 h after BRAF V600E transduction compared with vector or non-transduced control group (Figure 2C).Immunostaining for GFAP and Ki67, an indicator of proliferation, showed that total number of GFAP+ cells and percentage of Ki67+ cells in BRAF V600E -transduced group were signi cantly higher than vector or non-transduced control group 96h after transduction, while transduction with BRAF WT showed no signi cant difference compared with vector or the control group (Figure 2D).IL-6, IL-10, and cyclooxygenase 2 (COX2) were increased at mRNA level whereas TNF-α and IL-1β were decreased after BRAF V600E transduction.In addition, nuclear factor erythroid 2 (NF-E2)-related factor 2 (Nrf2) and its downstream genes heme oxygenase 1 (HO1), quinone oxidoreductase 1 (NQO1), glutamate cysteine ligase catalytic subunit (GCLC), and glutamate cysteine ligase modi er subunit (GCLM) were increased signi cantly in BRAF V600E -transduced astrocytes.In addition, there was a dramatic increase in glialderived neurotrophic factor (GDNF) mRNA level in BRAF V600E -transduced cells (Figure 2E).

BRAF V600E expression in microglia induces cell proliferation and activation through ERK
To determine BRAF V600E effects on microglial cells, we transduced BRAF into BV2 cells for 96 h.Western blotting showed that expression of BRAF, p-MEK1/2, p-ERK1/2, and phosphorylated nuclear factor kappa-B (NF-kB) p65 (p-NF-kB p65) in the BRAF V600E group was dramatically increased compared with vector and non-transduced control groups (Figure S1A).Flow cytometry showed lower percentages of cells in G0/G1 and G2/M phases and higher percentage of cells in S phase in BRAF V600E -transduced group compared with vector or non-transduced control groups (Figure S1B).MTS also revealed that transduction with BRAF V600E induced BV2 cell proliferation (Figure S1D).Immunostaining for Iba1 showed that cell number in BRAF V600E -transduced group was signi cantly increased compared with vector and non-transduced control groups.Morphologically, BV2 cells showed a typical branching shape at the resting state in non-transduced control group.Transduction with vector and BRAF WT caused no obvious morphological changes compared with the control group.BV2 cells transduced with BRAF V600E displayed reduced cell area, loss of rami cations, and development of an amoeboid shape (Figure S1C).qPCR analysis showed increases in TNF-α, IL-1β, IL-6, and COX2 in BRAF V600E -treated BV2 cells.mRNA levels of Nrf2 and its downstream genes HO1, NQO1, GCLC, and GCLM were decreased signi cantly after transduction with BRAF V600E compared with vector and non-transduced control groups (Figure S1F).
ELISA showed that in ammatory factors (IL-1β, IL-6, TNF-α) secreted into the culture medium by BRAF V600E -treated BV2 cells was higher than in the control groups (Figure S1G).There was no signi cant difference in NO levels among the groups (Figure S1E).These results support proliferation and activation of BV2 microglia by BRAF V600E .
We next employed mouse primary microglia to further characterize BRAF V600E effects in microglial cells.
Cells were transduced with BRAF vectors for 120 h.BRAF V600E transduction signi cantly increased p-MEK1/2 and p-ERK1/2.In addition, p-NF-κB p65 as well as p-JNK were increased (Figure 3A).To determine the roles of activated ERK and JNK in BRAF V600E effects, we transduced microglial cells with siRNA targeting JNK (si-JNK), siRNA targeting ERK (si-ERK), and control-siRNA.Knockdown e ciency of si-JNK and si-ERK was demonstrated by Western blotting (Figure 3B).After treatment with si-ERK, BRAF V600E -transduced microglia showed signi cantly downregulated ERK as compared with control-siRNA treated cells expressing BRAF V600E .NF-κB phosphorylation was similarly downregulated.si-JNK did not change p-NF-κB in BRAF V600E -transduced microglia (Figure 3C).Iba1 staining showed that the cell number and percentage of Ki67+ cells in BRAF V600E -treated group were signi cantly increased compared with vector and non-transduced control groups.BRAF WT -treated group did not show signi cant difference from vector and non-transduced control groups (Figure 3D).si-ERK treatment blocked BRAF V600E -induced proliferation of microglia.si-JNK did not change Iba1+ cell number and percentage of Ki67+ cells in BRAF V600E -transduced cells (Figure 3E).Non-transduced control microglia displayed the typical branching shape at the resting state.Transduction with vector, BRAF WT and BRAF V600E plus si-ERK caused no obvious morphological changes compared with the control group.BRAF V600E and BRAF V600E plus si-JNK transduction induced enlargement of the microglial cell body, loss of rami cations, and development of an amoeboid shape (Figure 3D and 3E).
Flow cytometry showed that, compared with vector group, percentage of S phase microglia in BRAF V600Etreated group was increased, and this effect of BRAF V600E was diminished when cells are treated with si-ERK but not si-JNK (Figure 3F).MTS assay revealed higher cell viability after transduction with BRAF V600E for 96 and 120 h compared with vector and non-transduced control groups, and si-ERK reversed this effect (Figure 3G).In ammatory markers TNF-α, IL-1α, IL-1β, IL-6, human inducible NO synthase (iNOS), and COX2 were increased at the mRNA level after transduction with BRAF V600E for 120 h.BRAF WT -transduced microglia displayed similar trends but to a lesser extent, supporting BRAF activity dependence of the changes.si-ERK signi cantly inhibited BRAF V600E -induced in ammatory gene expression, while si-JNK did not reverse the effect (Figure 3I).ELISA and nitrite release showed that in ammatory factors (IL-1β, IL-6, TNF-α) and NO secreted by microglia after treatment with BRAF V600E were higher than vector and non-transduced control groups.si-ERK, but not si-JNK reversed secretion of in ammatory factors (IL-1β, IL-6, TNF-α) and NO induced by BRAF V600E (Figure 3H and Figure 3J).

Conditioned medium from BRAF V600E -expressing microglial cells but not astrocytes induces neuronal cell death
We treated differentiated SH-SY5Y cells with conditioned medium from BV2 cells transduced with BRAF V600E , BRAF WT , vector, or from non-transduced control cells for 24 h.MAP2 staining showed a signi cantly loss of neuronal cells in differentiated SH-SY5Y cells cultured with BV2-BRAF V600E conditioned medium as compared to cells cultured with BV2-vector and BV2-BRAF WT conditioned medium.Consistently, these cells showed increased lactate dehydrogenase (LDH) release.Differentiated SH-SY5Y cells cultured with BV2-control, BV2-vector and BV2-BRAF WT conditioned medium exhibited no obvious changes in cell numbers and LDH release as compared to the control group (Figure S2).
We next treated primary cortical neurons with conditioned medium from astrocytes transduced with BRAF V600E , BRAF WT , vector, or from non-transduced control cells for up to 72 h.Astrocytes transduced with BRAF V600E were proliferating with upregulated GDNF, certain cytokines and antioxidants as shown in Figure 2.However, no signi cant difference in number of MAP2+ was demonstrated in all astrocyte conditioned medium-treated groups compared with the control group.LDH release showed no signi cant change among groups at 24, 28, and 72 h (Figure 4A).
Primary cortical neurons were in parallel treated with conditioned medium from primary microglia transduced with BRAF V600E , BRAF WT , vector, or from non-transduced control cells for 72 h.A signi cant decrease in the number of MAP2+ cells were demonstrated in cells treated with microglia-BRAF V600E conditioned medium (Figure 4B).BRAF V600E conditioned medium also induced LDH release in primary neurons (Figure 4C).Treatment with microglia-control, microglia-vector and microglia-BRAF WT conditioned medium did not result in changes compared to the control group.Similar neurotoxicity was demonstrated in microglia-BRAF V600E + si-JNK conditioned medium group, but not in microglia-BRAF V600E + si-ERK conditioned medium group, suggesting that the indirect, toxic effects of BRAF V600E on neurons resulted from ERK-mediated changes in microglia (Figure 4A and 4B).

BRAF V600E expression in neurons directly promotes cell death through the JNK pathway
To characterize direct neuronal effects of BRAF V600E , primary cortical neurons were transduced with BRAF V600E BRAF WT , and vector for 24 h.The cells were then cultured for 96 h in neural culture basal medium.Western blotting con rmed BRAF overexpression and higher levels of p-MEK1/2, p-ERK1/2, p-JNK, caspase-3 and c-Jun in the BRAF V600E -treated group as compared with vector and non-transduced control groups (Figure 5A).Marked increases in c-Jun, B cell leukemia/lymphoma-2 (Bcl-2), Bcl-2 associated X protein (Bax), p53, Fas-ligand (FasL), and TNF-α at the mRNA level were observed after transduction with BRAF V600E for 120 h (Figure 5B).BRAF V600E induced signi cant neurotoxicity as demonstrated by loss of MAP2+ cells and release of LDH in BRAF V600E -transduced neurons compared with vector-and non-transduced neurons (Figure 5C).
si-JNK and si-ERK were used to block signals downstream to BRAF.Knockdown e ciency of si-JNK and si-ERK was veri ed by Western blotting (Figure S3).Reduced p-JNK, caspase-3, and c-Jun was demonstrated in neurons treated with si-JNK + BRAF V600E as compared with cells treated with BRAF V600E (Figure 5D).At mRNA level, si-JNK reversed BRAF V600E -induced changes in c-Jun, Bax, Bcl-2, p53, Fasl and TNF-α in cultured cortical neurons (Figure 5E).Treatment with si-ERK blocked p-ERK and total ERK expression induced by BRAF V600E (Figure 5G).There was no signi cant difference in protein expression of p-JNK, caspase-3 and c-Jun or mRNA expression of c-Jun, Bax, Bcl-2, p53, Fasl and TNF-α between BRAF V600E and BRAF V600E + si-JNK groups (Figure 5G and 5H).These results suggest that activation of the JNK pathway by BRAF V600E is independent of BRAF V600E -induced ERK activation.MAP2 staining as well as LDH assay showed signi cantly reduced number of MAP2+ cells and higher LDH release in cells transduced with BRAF V600E and the changes were partially reversed after co-treatment with si-JNK (Figure 5F).si-JNK also inhibited JNK in vector-transduced neurons but without signi cant in uences on c-Jun, Bax, Bcl-2, p53, Fasl and TNF-α and neuronal survival (Figure 5D and 5F), which suggests that a threshold JNK activity may be required in activating downstream genes and inducing neuronal cell death in this culture system [26].Co-treatment with si-ERK did not change MAP2+ cell number and LDH release in BRAF V600E -transduced cells (Figure 5I).Despite reported role of ERK in neuronal survival [27] and a trend for reduced number of MAP2+ cells and increased LDH release at 120 h, there was no statistically signi cant difference between si-ERK and control-siRNA treated vector-transduced neurons at the time points investigated.

Enrichment of BRAF-related genes in pathways associated with Parkinson's disease
To explore associations of BRAF signaling with neurodegeneration, we performed GO and KEGG pathway enrichment analyses of the GDS2821 dataset, which consists of gene expression pro les in the SN from postmortem brains of patients with Parkinson's disease (n=16).The BRAF-related genes were signi cantly enriched in chemical synaptic transmission, protein polyubiquitination, proteasomemediated ubiquitin-dependent protein catabolic process, etc. at the BP levels; mitochondrial inner membrane, nucleus, cytoplasm, etc. at the CC levels and protein binding, ubiquitin-protein transferase activity, ligase activity, etc. at the MF levels (Figure 6A and Table S1).KEGG pathway analysis indicated that metabolic pathways, oxidative phosphorylation, and synaptic vesicle cycle, were mostly associated with the BRAF-related genes (Figure 6B and Table S2).

Discussion
Our study using multiple in vitro systems provides evidence that oncogenic BRAF V600E induces 1) proliferation and activation of microglial cells through activated ERK but not activated JNK; 2) neuronal cell death directly through activated JNK but not activated ERK, and indirectly through ERK-activated microglia; 3) proliferation of astrocytes.BRAF WT showed no effects or minor effects in the same direct as BRAF V600E , suggesting either speci city of the mutant BRAF actions or more likely, the dose effects of BRAF activation.
Normal BRAF signaling and the Ras-Raf-MEK-ERK pathway control survival, proliferation, senescence, and differentiation of cells as results of their responses to environmental cues.BRAF V600E and the constitutive activation of the pathway induce proliferation and cancer transformation either alone or in synergy with other oncogenic insults in a wide range of dividing cell [28].In melanocytes, for example, BRAF V600E induces MEK and ERK and promotes cell growth and tumorigenicity [29].In addition to malignant melanomas, BRAF V600E is also common in primary brain tumors of glial origin including pleomorphic xanthoastrocytoma, ganglioglioma, epithelioid glioblastoma, and gliomas [30].Our ndings that BRAF V600E induced proliferation of astrocytes in primary cortical mixed culture and primary astrocytes are consistent with previously reported increased cell proliferation in primary mouse astrocytes overexpressing BRAF V600E by retrovirus transduction [31].The overgrowing astrocytes appear to be active in producing cytokines and chemokines, neurotrophic GDNF, and activating anti-oxidant genes, at least at mRNA levels.Although whether physically proximal BRAF V600E -expressing astrocytes may in uence neurons such as in the case of the mixed culture is unknown, neither trophic nor toxic effects were observed in neurons treated with conditioned medium from astrocytes expressing BRAF V600E .The lack of effects might be explained by the relatively small extent of ERK activation and differential changes in downstream effectors.
BRAF V600E similarly activated ERK in BV2 and primary microglial cells.Proliferation and activation of microglia following BRAF V600E expression were clearly demonstrated at both morphological and functional levels.Primary microglial cells expressing BRAF V600E exhibited classical activation state with induction of iNOS and NF-κB as well as various cytokines and chemokines such as TNF-α, IL-1 β, IL-6, superoxide, ROS and NO.Our demonstration of the critical role of ERK in mediating microglial proliferation and activation con rms ERK as the downstream effector of the BRAF V600E action [15,32].
The proliferative and activating responses of microglial cells to BRAF V600E expression conveyed their impact on neuronal survival.Neurotoxicity of conditioned medium from microglial cells overexpressing BRAF V600E is likely a result of proin ammatory cytokines and chemokines secreted by the activated microglial cells.Studies from cancer especially melanoma eld have demonstrated BRAF V600E and altered immunity in tumor initiation and progression [33].Particularly microphages may play a critical role in melanoma resistance to BRAF inhibitors [34].As the resident macrophage within CNS, microglia have long been associated with neurological pathologies.status also predicts risk of neurodegenerative complications in these conditions.ERK activation in activated microglia has also been implicated in common age-related neurodegenerative disorders such as Parkinson's disease and Alzheimer's disease [36,37].Our study further demonstrated in vitro a clear causal link between BRAF V600E -induced and ERK-mediated activated microglia and neuronal cell death that does not require physical proximity as in vivo.
The BRAF V600E expression in neurons likewise directly caused cell death.This direct effect of the mutant BRAF on neurons, however, was not mediated by ERK but JNK.The JNK/c-Jun pathway in BRAF V600E melanomas is involved in adaptive responses to targeted therapy [38,39].The pathway may be paradoxically upregulated following BRAF inhibitor treatment, which may promote cancer cells becoming quiescent and therefore resistant to drug-induced apoptosis.Both EKR and JNK are activated by BRAF V600E expression in both primary microglia and neurons in the present study.Although inhibition of endogenous ERK and JNK in vector treated cells did not appear to affect cell survival and proliferation in either neurons or glial cells in our systems, which may be explained by incomplete block of either using siRNA approach, both ERK and JNK are necessary for neuronal development and maintenance, and their abnormal activation can lead to dysfunction and cell death.JNK activation in neurons is generally considered proapoptotic in various conditions [26,40].A recent study reported that the JNK pathway was activated in neurotoxic amyloid peptide Aβ 42 -treated neurons or neurons from a transgenic mouse model of Alzheimer's disease, which resulted in aberrant cell cycle re-entry and caused neuronal apoptosis [41].
ERK signaling, on the other hand, can be either protective or associated with promotion of neuronal cell death depending on the contexts [42].Growing evidence supports that activation of ERK can mediate cell death in several neuronal systems [43,44].BRAF itself have been shown to play an essential role in neuron differentiation and axon growth [45].BRAF activation was found to be neuroprotective through inhibition of activating transcription factor (ATF)-3 of the ATF/cAMP response element binding protein family [46].The proliferative effect in microglia mediated by ERK and the toxic effect in neurons mediated by JNK following BRAF V600E expression in our systems further highlight speci c cellular contexts in dictating effectors and thus consequences of BRAF activation.
Much effort has been made in developing BRAF inhibitors for targeted cancer therapy.Consistent overall with our results that BRAF activation is neurotoxic, dabrafenib, a BRAF inhibitor and melanoma drug was found to inhibit c-Jun and to activate ERK, paradoxically, and to be neuroprotective in SH-SY5Y cellular and mouse models of Parkinson's disease, though it is unknown whether the protection by dabrafenib was through direct inhibition of BRAF in neurons or indirect inhibition of activated microglia in the in vivo models [16] .The aforementioned mouse study showed that BRAF inhibitor PLX4720 was neuroprotective through apparent inhibition of activated microglia [15].Microglial activation contributes to neurodegeneration in Parkinson's disease [47]

Supplementary Files
This is a list of supplementary les associated with this preprint.Click to download.

Figure 5 BRAFV600E
Figure 5 FigS1.tif FigS2.tifFigS3.tifSupplementarymaterials.docx [35] et al.reported that a somatic BRAF V600E mutation in yolk-sac erythro-myeloid progenitors, which give rise to microglia during development, resulted in clonal expansion of tissue-resident macrophages, including microglia with activated ERK[15].The neurodegenerative disorder developed late in this mouse model, including synaptic loss and neuronal death were driven by ERK-activated microglia and were preventable by BRAF inhibition.The model closely resembles late-onset neurodegenerative disease observed in patients with histiocytoses, which are clonal myeloid diseases characterized by in ammatory lesions[35].Activating somatic mutations in the RAS-MEK-ERK pathway, predominantly BRAF V600E are strongly associated with histiocytosis, and BRAF V600E