Trametinib activates endogenous neurogenesis and recovers neuropathology in a model of Alzheimer’s disease

Enhancing adult neurogenesis in the brain has been suggested as a potential therapeutic strategy for AD. We developed a screening platform, ATRIVIEW®, for molecules that activate neuronal differentiation of adult mouse NSCs. The most potent hit from an FDA-approved drug library was SNR1611 (trametinib), a selective MEK1/2 inhibitor. We found that trametinib increases the levels of P15INK4b and Neurog2, suggesting a mechanism by which MEK1/2 inhibition induces neuronal differentiation. Oral administration of trametinib increased adult neurogenesis in the dentate gyrus and subventricular zone of the 5XFAD AD mouse model. Surprisingly, we also found that trametinib enhanced adult neurogenesis in the cortex. Consequently, trametinib rescued AD pathologies such as neuronal loss and cognitive impairment in 5XFAD mice. Finally, trametinib induced neurogenic differentiation of NSCs derived from AD patient iPSCs, which suggests its potential therapeutic application. Altogether, we suggest that restoration of endogenous adult neurogenesis by trametinib may be a promising therapeutic approach to AD.


Introduction
Alzheimer's disease (AD) is characterized by impaired cognitive functions and memory loss caused by neuronal degeneration in the brain cortex and the hippocampus.Pathogenic mechanisms of AD have not been clearly elucidated, but it is known that deposition of amyloid plaques or hyperphosphorylated tau tangles damages neuronal networks through diverse events such as autophagic-lysosomal dysfunction, mitochondrial dysregulation, synaptic loss, and neuroin ammation 1,2 .Thus, the development of AD treatments has mainly focused on eliminating amyloid plaques and tau tangles.However, since AD patients have already undergone signi cant neuronal damage, such disease-modifying therapies were unsuccessful.Recently, regenerative approaches, including stem cell therapy, have attracted attention as a treatment for neurodegenerative diseases such as Parkinson's disease and AD 3 .Unfortunately, these approaches also have many challenges, such as securing a su cient amount of stem cells to be administered to the patient, the risk of tumor formation caused by a viral infection of stem cells, the technology of differentiating stem cells to the required cell type, or the problem of immune rejection 4 .
Adult neurogenesis has been observed in the subventricular zone (SVZ) of the lateral ventricle and the subgranular zone (SGZ) of the hippocampal dentate gyrus.Several studies have also reported evidence of adult neurogenesis in the neocortex, striatum, amygdala, hypothalamus, substantia nigra, and brain stem 5,6 .Recently, it was reported that neurogenesis in the adult brain, especially in the hippocampus, is maintained until 90 years of age, but it decreases sharply in AD patients 7,8 .Compared with healthy individuals, the number of neural progenitor cells (NPCs) in the SVZ and SGZ in the brain of AD patients increased abnormally 9 .At the same time, however, the number of migrating neuroblasts and newly differentiated mature neurons decreased due to the slowed progression of essential processes in neurogenesis along with disease progression 8 .In addition, profound impairment of adult hippocampal neurogenesis could be observed from the early stage of Alzheimer's disease 8 .Therefore, restoration of endogenous adult neurogenesis may be a promising therapeutic approach for AD.
Here we report that trametinib, a speci c MEK1/2 inhibitor, is identi ed by the phenotype-based screening platform, ATRIVIEW®, to induce neuronal differentiation of adult NSCs from an AD mouse model.
Trametinib treatment to NSCs induces expression of cell cycle inhibitor P15 INK4b 19 and proneuronal factor Neurog2 20 in vitro.By inhibiting MEK/ERK signaling, trametinib promotes adult neurogenesis of NSCs in SVZ, DG, and also in the brain cortex of 5XFAD AD model mice.Trametinib also rescues neuronal cell number and axonal length in the brain and recovers cognitive functions of AD model mice.Moreover, we provide evidence that trametinib can induce neurogenic differentiation of NSCs derived from AD patient iPSCs.

Trametinib treatment
Trametinib (Medchemexpress, Monmouth Junction, NJ) was micronized and suspended in the vehicle containing 5% mannitol, 1.5% hydroxypropyl methylcellulose, and 0.2% sodium lauryl sulfate.5XFAD mice and their age-matched WT mice (male, n = 7 ~ 10 per group) were divided into vehicle-and trametinib-treated groups.Vehicle or 0.1 mg/kg of trametinib was administered to mice of each group for 1 or 2.5 months by oral gavage once a day.All the mice were sacri ced by the perfusion method, and brain samples were processed for biochemical and immunohistochemical analyses.

Novel object recognition test
To test novel object recognition, mice were habituated in an empty open eld arena (42 cm×42 cm).For the training trial, mice were placed in an open eld arena with two identical objects for 5 min each.Next day, the test trial was performed for 5 min with one of the two familiar objects replaced with a new one.
Video tracking was performed, and the time that mice explored the novel object and time that mice explored the familiar object were measured.A Discrimination Index (DI) was de ned as (novel objectfamiliar object)/(novel object + familiar object).

Fear conditioning test
For the training test, mice were placed into a fear chamber.After allowing them to move freely for 200 seconds in a darkroom box (fear chamber), a sound of 460Hz, 75db was delivered for 5 seconds, and stimulation was given with a current of 0.5mA strength for 2 seconds immediately after the sound.These sounds and electrical stimulation were repeated a total of 4 times.For the contexture test, 24 hours after the training test, the mice were placed in the fear chamber and contextual fear was measured for 120 seconds.Freezing behavior was analyzed every 60 sec to report the percentage of freezing.
For transient transfection, the NSCs were cultured for 1 day and then transfected with the required plasmids using the Lipofectamine2000 transfection reagent (Thermo Fisher, cat.no.11668019) according to the manufacturer's instructions.For MEK knock-down experiments, NSCs were transfected with shRNA-MEK1 and shRNA-MEK2 using lipofectamine.After 4 hrs of incubation, cells were further grown in the growth factor depleted medium.After 48 hrs, cells were harvested.Mouse Map2k1 #: GAGGCCTCTCAGCTCATAT (Dharmacon V3SM11241-234594572) and mouse Map2k2 #1: ACATGTCCCCAGAGCGGCT (Dharmacon V3SM11241-235104473) were used.

Drug screening
For ATRIVIEW®, we used NSCs from Tg2576 mouse which were kindly provided by Dr. Manho Kim at Seoul National University Hospital.For neurosphere cultures from Tg2576 mice, cells were grown in DMEM/F12 with B27 supplements in 75 cm 2 asks in suspension.bFGF (20 ng/ml) and human EGF (20 ng/ml) were added to the media to allow the cells to form neurospheres.It is known that in growth factor free medium, Tg2576-derived NSCs start to deposit Aβ 22 .For drug screening, adult NSCs derived from Tg2576 were plated on poly-L-ornithine/laminin coated 96 well plates.After 24 hrs, the cells were treated with 500 nM of FDA-approved drugs (994 compounds) and incubated in growth factor free medium for 48 hrs.Immunocytochemistry was performed using anti-Tuj1 antibody (Cell signaling, 4466; 1:200) which is a neuron-speci c marker 24 and Alexa Fluor 488 (Thermo, A21121; 1:200).The nucleus was stained with 4',6-diamidino-2-phenylindole (DAPI).Fluorescence was measured using a uorometer, Varioskan Flach (Thermo Fisher, VLBL00D0 N12391).

Immunocytochemistry
NSCs and iPSC derived-NSCs were cultured on glass coverslips.After washing three times with PBS for 5 min, cells were xed in 10% formalin for 10 min at room temperature.The cells were then washed with PBS and permeabilized in 0.1% Triton X-100 for 15 min.Cells were placed in blocking solution containing 5% BSA and 10% normal goat serum (NGS) in PBS for 1 hr at room temperature and then incubated with primary antibodies in 1% BSA and 10% NGS in PBS overnight.After washing, cells were incubated with uorescence conjugated secondary antibodies for 1 hr at room temperature.After washing, cells were incubated with 4',6-diamidino-2-phenylindole (DAPI) for 5 min.Coverslips were mounted using mounting medium (Biomeda, M01) and visualized by confocal microscopy using a LSM880 microscope (Carl Zeiss).Antibodies are indicated in Table S2.

Western blotting
Cultured cells and mouse brain tissues were washed twice with ice-cold phosphate-buffered saline (PBS) and extracted by homogenizing with Ripa buffer (50 mM Tris-HCl, 150mM NaCl, 0.25% deoxycholic acid, 1% NP-40, 1mM EDTA, and protease inhibitors, pH 7.4).Lysates were centrifuged at 13,000 rpm for 20 min at 4°C, and the protein concentration in the supernatant was determined using the Bradford assay (Bio-rad, 5000006).Proteins from each sample were subjected to 8% ~ 15% SDS-PAGE, and the resolved proteins were transferred to nitrocellulose membrane.The membranes were blocked with 5% nonfat milk powder in Tris-buffered saline/Tween 20 (TBST) for 1 hr at room temperature and then incubated with primary antibodies overnight at 4°C.After washing, membranes were incubated with secondary antibodies for 1 hr at room temperature.Peroxidase activity was visualized with enhanced chemiluminescence.The detected signals were quanti ed using an iBright™ imaging system (Thermo Fisher, A44115).Antibodies are indicated in Table S2.

Quantitative PCR (qRT-PCR)
Quantitative PCR analysis was performed as previously described 25 .Total RNA was extracted from cells using TRIzol (Invitrogen, 15596018).Reverse transcription was performed using M-MLV reverse transcriptase (Thermo Fisher, 28025013).qRT-PCR was performed using the SYBR™ Green PCR master mix (Thermo Fisher, 4368706) according to the manufacturer's guidelines.Results were expressed relative to the housekeeping gene GAPDH (Glyceraldehyde-3-Phosphate Dehydrogenase).For analyzing the expression of mouse or human genes, the primers are indicated in Table S3 and S4.

Immunohistochemical analysis
Mice were perfused with ice-cold phosphate buffered saline (PBS) and followed by 4% paraformaldehyde (PFA).Brains were dissected and analyzed by immunohistochemistry 26 .For para n sections, brain hemispheres were embedded in para n and prepared into sagittal sections of 5 µm slices.Para n sections were depara nized, and antigen retrieval was performed in citrate buffer (pH 6.0).For cryosections, brain hemispheres were embedded in O.C.T compound and prepared into sagittal sections of 10 µm slices.The cryosections were air dried before use.For immunostaining, the sections were incubated with primary antibodies overnight at 4℃.This step was followed by incubation with secondary antibodies for 1 hr at room temperature.The sections were counterstained with DAPI.Immuno uorescent images were captured using a LSM880 Laser-Scanning confocal microscope (Carl Zeiss).For diaminobenzidine (DAB) staining, immunohistochemistry was performed with the DAB kit (Vector Lab., SK4100).Immunostained cells were counted in the area using the Icy micromanager program (Institut Pasteur) or Zen software (Carl Zeiss).Cortical neurons were counted in the somatosensory cortex of 5XFAD mice.
Whole cell RNA sequencing RNA was isolated from mouse whole brains, and cDNA libraries for RNA sequencing were prepared using the TruSeq Stranded mRNA Prep Kit (Illumina, San diego, CA) according to the manufacturer's guidelines 27 .The libraries were sequenced on the Illumina Nextseq500 platform, and the reads were mapped to the reference Mouse mm10 genome using Tophat v2.0.13.The total number of reads mapped to the transcriptomes were 24,532 genes, and the genes with 0 count in at least one sample were removed before differential expression analysis.There was a total of 15,727 genes after the removal of genes with 0 count.To de ne differentially expressed genes (DEG), we set up a stringent statistic cutoff of fold change (FC) of ≥ 1.5 and a false discovery rate (FDR) < 0.05.A total of 160 DEGs (107 genes were up regulated and 53 genes were downregulated) were identi ed between the vehicle-treated group and trametinib-treated group in the second week.Gene ontology was performed with the biological process using the Panther database.The signi cance threshold for analyses was set to 0.05 using Fisher's exact test-adjusted p-values.

Statistical analysis
All data were analyzed in Prism (GraphPad Software).All graph data are presented as the mean ± s.e.m.Data distribution was assumed to be normal, and this was not formally tested.Mice were randomly assigned to experimental groups, and no animals or data points were excluded from analyses.Investigators were blinded to group allocation during data collection and analysis.No statistical methods were used to predetermine sample sizes; sample sizes were determined on the basis of pilot studies, and randomization procedures were not used.Statistical difference between groups was determined using one-tailed unpaired Student's t-test or one-way ANOVA followed by Dunnett's analysis or Fisher's LSD analysis.P values and degrees of freedom are described in the gure legends.Signi cance was reported as *P < 0.05, **P < 0.005 or ***P < 0.001.

ATRIVIEW®, the CNS regenerative drug screening platform, and properties of its hit trametinib (SNR1611)
Because drugs against AD aiming for Aβ clearance or interference of its production have not been successful in restoring already damaged cognitive functions, we hypothesized that enhancing adult neurogenesis from endogenous NSCs will restore neuronal network integrity in affected brain regions of AD patients.Thus, we developed a phenotypic drug discovery platform, ATRIVIEW®, aiming to identify small molecular compounds inducing neuronal differentiation of adult NSCs derived from AD model mice (AD-NSCs) (Fig. 1a).Using NSCs derived from Tg2576 AD mouse model, we screened 994 small molecules from the FDA-approved drug library to identify small molecules inducing neuronal differentiation.By comparing the level of Tuj1 (class III beta-tubulin, neuronal marker) expression using uorescent immunocytochemical analysis, 48 compounds were found to increase neuronal differentiation (at least 1.7-fold increase to the DMSO-treated control) (Supplementary Fig. 1a).Among them, the most effective drug inducing the neuronal differentiation of NSCs was SNR1611, which turned out to be trametinib (Mekinist®), a speci c MEK1/2 inhibitor and FDA-approved anti-cancer drug.Upon immunocytochemistry, it was observed that trametinib treatment induced morphological changes of NSCs into neuron-like cells (Fig. 1a).Another MEK1 inhibitor, Cobimetinib 28 , and a CDK inhibitor, Dinaciclib 29 , were among the selected drugs.Semagacestat 30 , a g-secretase inhibitor, and Duloxetine 31 , a serotonin-norepinephrine reuptake inhibitor, which have been known for their therapeutic potential for AD were also screened (Supplementary Fig. 1a).In the presence of growth factors in the medium (20 ng/ml EGF and 20 ng/ml bFGF), trametinib inhibited proliferation (assessed by PCNA level) and induced neuronal differentiation (by Tuj1 level) but not astrocytic differentiation (by GFAP level) of adult NSCs from C57B/L6 mice (Supplementary Fig. 1b).Regardless of Aβ 1− 42 oligomer (10 mM) presence, both MEK1/2 inhibition by trametinib (Supplementary Fig. 1c) and MEK1/2 knock-down (Supplementary Fig. 1d) induced neuronal differentiation of NSCs.These ndings indicate that inhibition of MEK/ERK signaling by trametinib is a potential approach to induce neuronal differentiation of adult NSCs.

Inhibition of MEK/ERK signaling by trametinib induces neuronal differentiation via induction of P15 INK4b and Neurog2 expression and protects against cell death
It has been reported that MEK/ERK signaling is activated in the brain of AD patients compared to the normal brain 32,33 and Aβ plaques activate this signaling 34 .To nd out whether MEK/ERK signaling is also activated in NSCs of the AD model mice with Ab plaque phenotypes, we measured the pERK level in the brain of WT and AD model mice, 5XFAD.It was con rmed that the MEK/ERK signaling pathway was activated in the SGZ (Fig. 1b, c) and SVZ (Fig. 1d, e) of 7.5-month-old 5XFAD mice compared with WT.Administration of trametinib (for 2.5 months to 5-month-old 5XFAD mice) reduced the level of pERKs in both areas (Fig. 1b-e).Thus, we further tested whether trametinib induces the neuronal differentiation of adult NSCs isolated from the 5XFAD mouse.The size of neurospheres derived from 5XFAD were smaller than that from wild type (WT) (Fig. 1f, g).In addition, trametinib strongly increased the level of Tuj1 but not the level of GFAP in NSCs isolated from 5XFAD mouse, indicating its induction of neuronal but not astrocytic differentiation.(Fig. 1h, i).Meanwhile, we con rmed that other MEK1/2 inhibitors (AZD8330, PD184352, Refametinib, PD318088, AS703026) also induced neuronal differentiation (Supplementary Fig. 2) in embryonic NSCs, indicating that MEK1/2 inhibition activates neuronal differentiation.Notably, we observed that trametinib was the most effective MEK inhibitors to neuronal differentiation.
Since trametinib was discovered as a drug to induce the CDK inhibitor (such as P15 INK4b )-mediated cell cycle arrest by inhibiting MEK1/2 35 , we tested whether it would induce neuronal differentiation through cell cycle arrest in NSCs as well.We found that trametinib induced P15 INK4b (cell cycle arrest) and Neurog2 (proneuronal factors) expression and increased the protein levels of P15 INK4b and Neurod 1 (neuronal differentiation 1) in adult NSCs from 5XFAD mice (Fig. 1j-m).Whole brain RNA-Seq analysis was also supported that inhibition of MEK/ERK signaling is the mechanism responsible for the neuronal differentiation.In Go term analysis, the second week of trametinib administration appeared to be the critical period for neuron development, pyramidal neuron differentiation and regulation of neuron migration (Supplementary Fig. 3).Among the 107 genes whose expression was increased by 1.5 times or more, it was con rmed that the expression of Ebf1, Nhlh2, Irx5, Ebf3, Irx3, Sox14, Cdh1, Tead4 genes, known as the target gene of Ngn2, was also increased by trametinib administration (Supplementary Table 1).
To investigate whether apoptosis was induced by Aβ accumulation as in the previous results 22 in adult NSCs isolated from 5XFAD mice and whether trametinib protects against this, we examined the expression of the apoptosis markers.The Aβ accumulation was observed in the adult NSCs from 5XFAD mice, and active caspase-3 was also observed simultaneously.Trametinib treatment, however, reduced the level of Aβ accumulation and active caspase-3.In addition, the level of cleaved Poly (ADP-ribose) polymerase (PARP), another apoptosis marker) was also decreased by trametinib treatment (Supplementary Fig. 4a-e).MEK1/2 knock-down by shRNA also reduced the expression of active caspase 3 (Supplementary Fig. 4f, g).
Taken together, these results demonstrate that activation of MEK/ERK signaling is related to the apoptosis of adult NSCs of 5XFAD mice and that trametinib protects against cell death.Furthermore, trametinib enhanced neuronal differentiation through induction of CDK inhibitors and proneuronal factors.

Trametinib induces hippocampal and SVZ neurogenesis in 5XFAD mice
Since trametinib induced the neuronal differentiation of NSCs derived from AD model mice, we examined whether oral administration of trametinib also induces hippocampal neurogenesis in 5XFAD mice.Fivemonth-old 5XFAD mice were administered 0.1mg/kg trametinib for 2.5 months, after which all mice were sacri ced at the age of 7.5 months (Fig. 2a).In the DG of the vehicle-administered 5XFAD mice, Sox2+/GFAP + NSCs 36 increased, but the number of both Neurod1 + neuroblasts and Tuj1 + immature neurons decreased compared with WT control (Fig. 2b, c).In contrast, the number of neuroblasts and immature neurons in the same brain region of the trametinib-administered 5XFAD mice increased, while the number of radial glial cells decreased (Fig. 2b, c).
Next, we asked whether trametinib also induces neuronal differentiation of NSCs in the SVZ of 5XFAD mice 37 .The number of NSCs expressingSox2 and GFAP (type B1 cells) in the SVZ of 5XFAD mice decreased compared with WT control mice (Fig. 2d, e).Also, the number of Dcx + neuroblasts (Type A cells) signi cantly decreased in 5XFAD mice.Trametinib increased the number of Sox2+/GFAP + NSCs, Ki67 + proliferating cells and Dcx + neuroblasts in the SVZ of 5XFAD mice to a level comparable to that observed in WT (Fig. 2d, e).These data indicate that the generation of type A neuroblasts from NSCs is impaired in 5XFAD, and trametinib restores the transition to facilitate adult neurogenesis in the SVZ.
Then we asked further whether trametinib induced the migration of newly formed neuroblasts in the SVZ to the olfactory bulb (OB).While the number of Dcx + neuroblasts was signi cantly reduced in the OB granular layer of 5XFAD mice, trametinib indeed recovered the neuroblast population (Fig. 2d).These data demonstrate that inducing adult neurogenesis in the SVZ by trametinib restores migration of neuroblasts to the OB, which is disrupted in AD conditions.Altogether, trametinib administration evidently restores adult hippocampal and SVZ neurogenesis, which is impaired in 5XFAD mice.

Trametinib induces cortical neurogenesis in the 5XFAD mouse brain
We asked whether trametinib can induce adult cortical neurogenesis in 5XFAD mice which is characterized by neuronal loss in the brain cortical layer V 38 .Previous studies demonstrated that the production of new neurons occurs in the brain cortex after cortical injury, and those new neurons originate from cortical glial cells or NSCs migrating from the SVZ [39][40][41] .Therefore, we investigated which cell population differentiated into cortical neurons.We counted the number of Sox2, GFAP, and Neurod1 triplepositive cells in the sagittal sections of the 5XFAD mouse somatosensory cortex.The Sox2 and GFAP double-positive cell is well known as a neural stem/progenitor cell in the DG and SVZ 42 .Sox2, especially, is known as a reprogramming factor in the adult brain 43 , and Neurod1 is expressed in dividing neuroprogenitor cells 43 .Thus, these triple positive (Neurod1+/Sox2+/GFAP+) cells may act as neurogenic progenitors in the cortex.After 0.1 mg/kg trametinib was administered to ve-month-old 5XFAD mice for 2.5 months, mice were sacri ced at the age of 7.5 months (Fig. 3a).Interestingly, triple-positive cells were signi cantly increased by trametinib administration in the 5XFAD mouse cortex (Fig. 3b, c).We also con rmed cortical neurogenesis using EdU incorporation analysis.We administered trametinib for 1.5 months to 7-month-old 5XFAD mice while giving them 200mg/kg of EdU injection 30 days before sacri ce at 8.5 months of age (Fig. 3d).The number of EdU-positive cells in the cortex increased with trametinib administration (Fig. 3e, f).Furthermore, the number of EdU/NeuN double-positive cells signi cantly increased (Fig. 3g, h), suggesting that trametinib distinctly induced neurogenesis and produced new neurons in the cortex.We also questioned whether trametinib can induce cortical neurogenesis in the late-symptomatic stage of AD and administered trametinib to 9-month-old 5XFAD mice for 1.5 months (Fig. 3i).The level of Dcx in the cortex was increased by trametinib administration (Fig. 3j, k), and EdU/NeuN double-positive cells were also increased (Fig. 3l, m).These results strongly suggest that NPCs exist in the AD brain cortex and that trametinib activates cortical neurogenesis in 5XFAD mice.

Trametinib induces functional rescue of AD pathogenesis by restoring neuron numbers and neuronal structure
To investigate whether adult neurogenesis, especially cortical neurogenesis, induced by trametinib supports the restoration of neurons in 5XFAD mice, we administered daily 0.1 mg/kg of trametinib to 5month-old 5XFAD mice for 2.5 months (Fig. 4a) or 12-month-old 5XFAD mice for 1 month (Fig. 4d).The number and axonal length of neurons in cortical layer V of the somatosensory cortex was decreased in the 7.5-month-old trametinib-treated 5XFAD mice, whereas 0.1 mg/kg of trametinib administration for 2.5 months to 5-month-old 5XFAD mice improved neuronal numbers and axonal length (Fig. 4b, c).The number and axonal length of neurons in cortical layer V of the cortex and subiculum were also signi cantly increased in 13-month-old 5XFAD upon trametinib administration compared with vehicletreated 5XFAD mice (Fig. 4e, f).This was an unexpected effect of trametinib because, at this late stage, neuronal loss in the cortex is believed to be too severe to allow restoration.
Consequently, we further examined whether trametinib administration recovers AD pathologies and improves cognitive functions in 5XFAD mice.In the fear conditioning test and novel object recognition test, vehicle-administered 5XFAD mice showed cognitive impairment, whereas administration of 0.1 mg/kg trametinib for 1.5 months to 9-month-old 5XFAD or for 1.5 months to 7-month-old 5XFAD mice improved cognitive functions as shown by an increase in the percentage of freezing or discrimination index measures (Fig. 4g, h) with no changes in locomotor functions (Supplementary Fig. 5a, b).These data indicate that trametinib recovers damaged neurons and cognitive function in 5XFAD AD model mice.

Trametinib induces neurogenic differentiation of AD patient iPSC derived-NSCs
Next, we asked if the neurogenic effect of trametinib observed in 5XFAD AD model mice has human relevance using NSCs derived from human iPSCs.When we treated AD patient iPSC-derived NSCs with trametinib, the level of pERKs was signi cantly decreased (Fig. 5a-c).The level of SOX2 did not change in the NSCs, but the level of an early post-mitotic neuronal marker Dcx was increased by trametinib (Figure .5D and 5E).To assess the neurogenic effect of trametinib on those NSCs, cells were immunostained for Dcx 2 days after trametinib treatment.The number of Dcx + cells was increased by trametinib (Fig. 5f, g).
The level of DCX mRNA also increased in trametinib-treated NSCs (Fig. 5h).These neurogenic effects of trametinib are also observed in healthy donor iPSC derived-NSCs (Supplementary Fig. 6).From these data, we suggest that trametinib has potential to induce neurogenic differentiation in AD patients.

Discussion
AD cannot be explained by a single causal factor since it develops through complex interplays among neuronal cells throughout the brain 44,45 .Aβ aggregation, autophagic defect, and in ammatory responses form a negative feedback loop toward neuronal death and subsequently, defective neural activity 46 .The development of AD therapeutics has focused on the elimination of toxic materials such as Aβ plaques or tau tangles so far.With the FDA's recent approval of aducanumab which targets the elimination of Aβ, controversy has arisen because the decision was made based on a surrogate endpoint -a signi cant reduction in Aβ plaques -rather than on a clinical endpoint 47 .Therefore, it is still necessary to change the drug development strategy that can modify the disease.
Our study originated from the idea that adult neuronal differentiation from NSCs could be an effective therapeutic approach against various neurodegenerative diseases.Therefore, we established a regenerative drug screening platform, ATRIVIEW®, using adult NSCs from Tg2576 mice which produce an Aβ-induced toxic environment 22 .Of the screened compounds, trametinib was not only the most effective in inducing neuronal differentiation but also protective of the differentiated neuron-like cells in the Aβ 1− 42 oligomer-induced toxic environment.In particular, this neuronal differentiation was accompanied by cell cycle arrest and proneuronal factor expression by trametinib.Moreover, we provide evidence that trametinib manifested its potential as an inducer of neuronal differentiation in AD patient iPSC derived-NSCs as well as adult NSCs derived from the 5XFAD AD model mice.
In 5XFAD mice, we demonstrated that oral administration of trametinib restores impaired neurogenesis in the DG and SVZ.The 5XFAD mice express human amyloid precursor protein (APP) and presenilin1 (PSEN1) with a total of ve AD-linked mutations 38 .These mice have a relatively early and aggressive presentation of amyloid plaques than other AD model mice.Furthermore, neuronal loss is also evident in the 5XFAD mice in the brain cortex and hippocampal subiculum.In the DG of 5XFAD mice, NSCs do not differentiate into neuroblasts or mature neurons e ciently (Fig. 2b, c).In the SVZ of 5XFAD mice, differentiation of NSCs into neuroblasts (type A cells) is severely impaired (Fig. 2d, e).Both this process and transition of type B1 NSCs to type C cells are induced by trametinib administration.In healthy brains, neuroblasts generated in the SVZ migrate to the OB via the RMS (rostral migratory stream), which is not observed in 5XFAD mice.Trametinib successfully replenished neuroblasts in the OB, which migrated from the SVZ (Fig. 2d).
In addition, there are several evidence of cortical neurogenesis in the 5XFAD mouse brain.We found that NPCs exist in the 5XFAD mouse brain cortex from immunohistochemical analyses (Fig. 3).In addition, newborn neurons which were speci ed by EdU/NeuN double staining were also increased in the cortex by trametinib administration.Thus, it is possible that the activation of cortical neurogenesis by trametinib supports the increase in neuron number in the cortex of 5XFAD mice and the rescue of cognitive impairment.
Finally, we asked about the possibility of clinical translation of trametinib using AD iPSC derived-NSCs.
We used familial AD (PSEN1 M146I) iPSC derived-NSCs and con rmed trametinib could induce neuronal differentiation.The PSEN1 M146I mutation is a widespread mutation for AD patients with PSEN1 mutations 48 .Therefore, it is reasonable to expect that trametinib can induce neurogenesis in an AD patient's brain.
It is known that adult neurogenesis is disrupted in AD patients 49 .Several studies have shown that TACs, involved in the early stages of neurogenesis in the hippocampus, are increased in a pathogenic environment, but their maturation is defective and unable to proceed further 8 .Our results suggest that restoring endogenous neurogenesis in the DG/SVZ and the cortex can contribute to the recovery from AD. Notably, our research showed that trametinib restores the supply of newly generated neurons even in the severe stage of the disease, at least in the AD animal model.In addition, trametinib induces neurogenic differentiation of AD patient iPSCs-derived NSCs.These ndings indicate that the regenerative approach using trametinib would be effective in treating AD and other neurodegenerative diseases.μm.Normalized by the WT-vehicle group.f, g Adult NSCs from 5XFAD mice were cultured to form neurospheres in the medium contained with 20 ng/ml EGF and 20 ng/ml bFGF.Gross images were obtained at 4 th day after seeding (f) and neurosphere size were quanti ed (g).Scale bars, 50 μm.h-l Adult NSCs from 5XFAD mice were seeded on the plates and treated with 100 nM of trametinib in the medium without growth factors for 48 hrs.Cell lysates were subjected to immunoblot analyses of TUJ1, GFAP, pERKs and ERKs (h) and quanti ed with band intensities (i).The level of Cdkn2b and Ngn2 mRNA was analyzed using qRT-PCR.Each sample was normalized to the expression of Gapdh (j).Normalized by cortex layer V area and renormalized to 5XFAD-vehicle group.Scale bars, 50 μm.The length of Tau-stained cells was measured in cortex layer V. n=3 mice per group.gNine-month-old 5XFAD mice were administered with the vehicle or trametinib for 1.5 months.Fear conditioning test was performed, and the average of freezing % in 2 minutes was calculated.n=12 mice for WT-vehicle group, n=22 mice for 5XFAD-vehicle group, n=11 for 5XFAD-trametinib group.P values were obtained by one-way ANOVA test.h Seven-month-old 5XFAD mice were administered with the vehicle or trametinib for 1.5 months.Novel object recognition test was performed.Heat map analysis of animal tracking following novel object recognition test (left panel) and the discrimination index in 5 minutes was calculated (right panel).n=5 mice for 5XFAD-vehicle group, n=8 for 5XFAD-trametinib group.Data are representative of three independent experiments and values are expressed in mean ± SEM.P values were obtained by oneway ANOVA test (c, g) and Student's t-test (f, h).*p < 0.05 and ***p < 0.001.Quanti cation data were done blind with respect to the experimental group.

Figure 1 Identi
Figure 1 Cell lysates were subjected to immunoblot analyses of P15 and Neurod1 (k) and quanti ed with band intensities (l).m The proposed mechanism how MEK1/2 inhibition induces the cell cycle arrest and neuronal differentiation.Data are representative of three independent experiments and values are expressed in mean ± SEM.P values were obtained by one-way ANOVA test (c, e) and Student's t-test (i-l).*p < 0.05, **p < 0.005 and ***p < 0.001.