iPSC-derived human microglia-like cells to study neurological diseases

Functional validation of iMGLs

Next, iMGLs were validated as surrogates of microglia using both functional and physiological assays. To this end, cytokine/chemokine secretion by iMGLs was measured following stimulation by Lipopolysaccharide (LPS), IL-1β or IFNγ. IL-1β and IFNγ are two cytokines that are elevated in AD patients and mouse models (Abbas et al., 2002; Blum-Degen et al., 1995; Patel et al., 2005; Wang et al., 2015) (Figure 3C). Basally, iMGLs secrete 10 of the examined cytokines at low but detectable levels (Table S2). However, in response to IFNγ or IL-1β, iMGLs secrete 8 different chemokines including TNFα, CCL2, CCL4, and CXCL10. As expected, iMGLs robustly responded to LPS with induction of all measured cytokines except for CCL3 (Table S2 for values). Collectively, iMGLs differentially release cytokines/chemokines based on their cell-surface receptor stimuli, a finding that closely aligns with the responses observed in acutely isolated primary microglia (Rustenhoven et al., 2016).

iMGLs express the microglial-enriched purinergic receptor P2ry12, which sense extracellular nucleotides from degenerating neurons, and is critical for microglial homeostatic function (De Simone et al., 2010; Moore et al., 2015) (Figure S2A,B). In response to ADP, iMGLs chemotax toward ADP and produce detectable calcium transients (Figure 3D,E), that were both negated by a P2ry12-specific inhibitor, PSB0739. These physiological findings further underscore that iMGLs respond appropriately to stimuli and express functional surface receptors, such as P2ry12, enabling quantitative analyses of microglial physiology.

Microglia, like astrocytes, play a critical role in synaptic pruning (Aguzzi et al., 2013; Paolicelli et al., 2011; Stephan et al., 2012). Because in vitro synaptosome phagocytosis assays are an established surrogate to study pruning, the ability of iMGLs to phagocytose human synaptosomes (hS) was quantitatively assessed. In comparison to MD-Mφ, iMGL phagocytosis of pHrodo-labeled hS was less robust (Figure 3F, G). However, iMGLs preferentially internalized hS when compared to E. coli particles and normalized to MD-Mφ (Figure S4D, E) supporting the notion that iMGLs and microglia are more polarized toward homeostatic functions than MD-Mφ.

As iMGLs and primary microglia express both C1q and CR3 (CD11b/CD18 dimer), iMGLs were used to assess whether synaptic pruning in human microglia primarily involves this pathway as seen in mice (Chung et al., 2013). Using an additive-free CD11b antibody, iMGL phagocytosis of hS was significantly reduced (−40.0%, ***p<0.0001)(Figure 3H, I). In contrast, an inhibitor of Mertk (UNC569), also implicated in synaptic pruning, only marginally decreased iMGL hS phagocytosis (−12.6%, *p<0.05)(Figure 3H, I). Similar to studies in murine KO models, our data indicates that MERTK plays a minor role in human microglia-mediated synaptic pruning (Chung et al., 2013), whereas C1q/CR3 is integral for microglia-mediated synaptic pruning in humans.

Utility of iMGLs to study Alzheimer’s disease

Impaired microglia clearance of beta-amyloid (Aβ) is implicated in the pathophysiology of AD and strategies to enhance clearance of AD pathology are being actively pursued by biopharma. Therefore, we examined whether iMGLs can phagocytose Aβ or tau, two hallmark AD pathologies. Like primary microglia, iMGLs internalize fluorescently labeled fibrillar Aβ (Figure 4B). iMGLs also recognize and internalize pHrodo-labled brain–derived tau oligomers (BDTOs) (Figure 4B). Fluorescence emitted indicates trafficking of pHrodo-conjugated BDTOs to the acidic lysosomal compartment showing that iMGLs can actively ingest extracellular tau that may be released during neuronal cell death (Villegas-Llerena et al., 2015) and support recent findings that microglia may play a role in tau propagation in AD and other tauopathies (Asai et al., 2015). Together, these findings suggest that iMGLs could be utilized to identify compounds in high-throughput drug-screening assays that enhance Aβ degradation or block exosome-mediated tau release.

Microglia genes are implicated in late onset AD, yet how they modify disease risk remains largely unknown. Thus, iMGLs were utilized to begin investigating how these genes might influence microglia function and AD risk. Hierarchical clustering using just these 25 AD-GWAS genes also demonstrates that iMGLs resemble microglia and not peripheral myeloid cells (Figure 4A). In their investigated basal state, iMGLs and microglia express many AD-GWAS-related genes including those without murine orthologs i.e. CD33, MS4A4A, CR1. Thus, iMGLs can be used to study how altered expression of these genes influence microglia phenotype in a way that cannot be recapitulated in transgenic mice. Therefore, iMGLs were used to investigate the influence of fAβ or BDTOs on AD-GWAS gene expression in microglia (Villegas-Llerena et al., 2015). Following fAβ exposure, iMGLs increased expression of 10 genes (Table S3) including ABCA7 (5.79 ± 0.44), CD33 (6.02 ± 0.41), TREM2 (4.86 ± 0.50, and APOE (2.52 ± 0.19), genes implicated in Aβ clearance/degradation. BDTOs increased expression of 4 genes including CD2AP (4.62 ± 0.45), previously implicated in tau-mediated toxicity (Shulman et al., 2014). In addition, 6 genes were differentially elevated in fAβ compared to BDTOs (Table S3). Interestingly, CD33, TYROBP, and PICALM, genes more enriched in other myeloid cells at baseline, were upregulated by fAβ and BDTOs suggesting that proteinopathies may alter microglia phenotype to resemble invading peripheral myeloid cells (Chan et al., 2007; Prinz et al., 2011; Stalder et al., 2005). In addition to AD GWAS genes, iMGLs express other CNS disease-related genes including APP, PSEN1/2, HTT, GRN, TARDBP, LRRK2, C9orf72, SOD1, VCP, and FUS and therefore, can likely be used to study other neurological diseases such as ALS, HD, FTD, and DLB in which microglia may play a prominent role (Bachstetter et al., 2015; Crotti et al., 2014; Lui et al., 2016; O’Rourke et al., 2016)(Figure S6C).

Chemical Reagents

All cell culture flasks, reagents, supplements, cytokines, and general reagents were purchased from Thermo Fisher Scientific, unless otherwise noted.

Cell Culture

RNA-seq library construction

Cells were harvested and washed three times with DPBS and stored in RNAlater, RNA preservation solution. RNA was extracted from all cell types using RNeasy Mini Kit (Qiagen) following manufacturer’s guidelines. RNA integrity (RIN) was measured for all samples using the Bioanalzyer Agilent 2100 series. All sequencing libraries analyzed were generated from RNA samples measuring a RIN score ≥ 9. The Illumina TruSeq mRNA stranded protocol was used to obtain poly-A mRNA from all samples. 200 ng of isolated mRNA was used to construct RNA-seq libraries. Libraries were quantified and normalized using the Library Quantification Kit from Kapa Biosystems and sequenced as paired-end 100 bp reads on the Illumina HiSeq 2500 platform.

Confocal Microscopy and Bright-field Imaging

Immunofluorescent sections were visualized and images captured using an Olympus FX1200 confocal microscope. To avoid non-specific bleed-through, each laser line was excited and detected independently. All images shown represent either a single confocal z-slice or z-stack. Bright-field images of cell cultures were captured on an Evos XL Cell Imaging microscope. Image analysis was performed using Olympus

FACS and Flow Cytometer Analysis

iHPCs were collected using cold (4°C) sterile filtered and degassed FACS buffer (1X DPBS, 2% BSA, and 0.05mM EDTA) spiked with human SCF (5 ng/ml). Cells were then filtered through 70 μm mesh to remove large clumps, washed with spiked FACS buffer (300 × g for 5 min 18°C), then stained (1:200) using spiked FACS buffer on ice for 1 hour in the dark using the following antibodies: anti CD34-FITC clone 561, anti CD41-PE clone HiP8, anti CD43-APC clone CD43-10G7, anti CD45-APC/Cy7 clone HI30 (Tonbo Biosciences), anti CD235a-PE/Cy7 clone HI264, and ZombieViolet™ live/dead stain, all from BioLegend unless noted. After staining, iHPCs were washed once with spiked FACS buffer and suspended using spiked FACS buffer (500–700 μl) and sorted utilizing the BD FACSARIA Fusion (BD Biosciences). Sorted cells were collected in cold basal iHPC differentiation medium spiked with SCF (10 ng/ml). Collected CD43+ iHPCs were then plated for iMGL differentiation as mentioned above. iMGLs were suspended in FACS buffer and incubated with human Fc block (BD Bioscience) for 15 min at 4°C. For detection of microglial surface markers, cells were stained with anti CD11b-FITC clone ICRF44, anti CD45-APC/Cy7 clone HI30 (Tonbo Biosciences), anti CX3CR1-APC clone 2A9-1, anti CD115-PE clone 9-4D, and anti CD117-PerCP-Cy5.5 clone 104D2, ZombieViolet™ live/dead stain, all from BioLegend. Cells were sorted on a FACS Aria II, FACS Aria Fusion (BD Biosciences) and data analyzed with FlowJo software (FlowJo).

Cytospin and May-Grunwald Giemsa Stain

1×10⁵ cells were suspended in FACS buffer (100 μl) and added to Shandon glass slides (Biomedical Polymers) and assembled in a cytology funnel apparatus. Assembled slides containing cells were loaded in a cytospin instrument and centrifuged (500 rpm, 5 min). Slides were allowed to air-dry for two minutes and immediately stained in May-Grunwald stain (100%; Sigma) for 5 min. Next, slides were washed in PBS for 1.5 min and immediately placed in Giemsa stain (4%; Sigma) for 20 min at room temperature. Slides were washed in double-distilled water 6 times and allowed to air-dry for 10 min. Slides were preserved using glass coverslips and permount (Sigma).

RNA Isolation and qPCR Analysis

Cells were stored in RNAlater stabilizing reagent and RNA was isolated using Qiagen RNeasy Mini Kit (Valencia, CA) following manufacturer’s guidelines. qPCR analysis was performed using a ViiA™ 7 Real-Time PCR System and using Taqman qPCR primers. Analysis of AD-GWAS genes utilized a custom Taqman Low Density Array card using the primers described below.

iMGL Co-culture with Rat Neurons

Rat hippocampal or cortical neurons were cultured for 21 days with 50% media change every 3–4 days. iMGLs were cultured with neurons at a 1:5 ratio (1 × 10⁶ iMGL to 5 × 10⁶ neurons) in 50% iMGL and 50% NB medium. After 3 days, iMGLs were collected for RNA isolation.

iMGL Transplantation in MITRG and Rag5xfAD Brains

All animal procedures were performed in accordance with NIH and University of California guidelines approved IAUC protocols (IAUC #2011-3004). MITRG mice were purchased from Jax (The Jackson Laboratory, #017711) and have been previously characterized (Rongvaux et al., 2014). MITRG mice allow for xenotransplantation and is designed to support human myeloid engraftment. iMGLs were harvested at day 38 and suspended in injection buffer: 1X HBSS with M-CSF (10 ng/ml), IL-34 (50 ng/ml), and TGFβ-1 (25 ng/ml). iMGLs were delivered using stereotactic surgery as previously described (Blurton-Jones, et al, 2009) using the following coordinates; AP: −0.6, ML: ± 2.0, DV: −1.65. Brains were collected from mice at day 60 post-transplantation per established protocols (Blurton-Jones, et al, 2009). Rag5xfAD mice were generated in this lab and previously characterized (Marsh et al., 2016). Rag5xfAD mice display robust beta-amyloid pathology and allow for xenotransplantation of human cells. iMGLs were transplanted into the hippocampi using the following coordinates; AP: −2.06, ML: ± 1.75, DV: −1.95. After transplantation mice were killed and brains collected using previously established protocol. Briefly, mice were anesthetized using sodium-barbiturate and perfused through the left-ventricle with cold 1X HBSS for 4 min. Perfused mice were decapitated and brain extracted and dropped-fixed in PFA (4% w/v) for 48 hours at 4°C. Brains were then washed 3 times with PBS and sunk in sucrose (30% w/v) solution for 48 hours before coronal sectioning (40 μm) using a microtome (Leica). Free-floating sections were stored in PBS sodium azide (0.05%) solution at 4°C until IHC was performed.

Immunocytochemistry and Immunohistochemistry

For ICC, cells were washed three times with DPBS (1X) and fixed with cold PFA (4% w/v) for 20 min at room temperature followed by three washes with PBS (1X). Cells were blocked with blocking solution (1X PBS, 5% goat or donkey serum, 0.2% Triton X-100) for 1 h at room temperature. ICC primary antibodies were added at respective dilutions (see below) in blocking solution and placed at 4°C overnight. The next day, cells were washed 3 times with PBS for 5 min and stained with Alexa Fluor^® conjugated secondary antibodies at 1:400 for 1 h at room temperature in the dark. After secondary staining, cells were washed 3 times with PBS and coverslipped with DAPI-counterstain mounting media (Fluoromount, southern Biotech). For BORG IHC, tissue were collected and dropped-fixed in PFA (4% w/v) for 30 min at room temperature and then washed three times with PBS. BORGs were then placed in sucrose solution (30% w/v) overnight before being embedded in O.C.T (Tissue-Tek). Embedded tissue was sectioned at 20 μm using a cryostat and mounted slides were stored at −20°C until staining. For BORG staining, mounted tissue was removed from storage and warmed by placing at room temperature for 30 min. Tissue were rehydrated and washed with room temperature PBS (1X) 3 times for 5 min. Heat-mediated antigen retrieval was performed by using Citrate Buffer (10mM Citrate, 0.05% Tween 20, pH=6.0) at 97°C for 20 min and then allowed to cool to room temperature. After antigen retrieval, slides were washed three times with PBS. Slides were then washed once in PBS-A solution (1X PBS with 0.1% Triton X-100) for 15 min. Tissue was blocked using PBS-B solution (PBS-A, 0.2% BSA, and 1.5% goat or donkey serum) for 1 h at room temperature. After block, primary antibodies were added to PBS-B solution (250–350 μl/slide) at appropriate dilutions (see below) and incubated overnight at room temperature. The next day, slides were washed with PBS-A solution 3 times for 5 min each. Tissue were blocked for 1 h using PBS-B solution at room temperature. After block, slides were incubated with Alexa Fluor^® conjugated secondary antibodies (all at 1:500) and Hoechst stain (1X) in PBS-B (for 250–300 μl/slide) for 2 h at room temperature in the dark. After secondary staining, slides were washed 5 times with PBS for 5 min. Slides were cover slipped using fluoromount (Southern Biotech). For mouse brain IHC, brains were collected, fixed, and processed as mentioned above. Free-floating sections were blocked in blocking solution (1X PBS, 0.2% Triton X-100, and 10% goat serum) for 1 h at room temperature with gentle shaking. For human TMEM119 staining, heat mediated antigen retrieval was performed prior to blocking, as performed previously (Bennett et al., 2016). Free-floating tissue antigen retrieval was performed by placing floating sections in a 1.5 ml micro centrifuge tube containing 1 ml of Citrate Buffer solution and placing in a pre-heated temperature block set at 100°C. Tissue was heated for 10 min at 100°C then removed and allowed to come to room temperature for 20 min before washing with PBS 3 times for 5 min and then proceeding with blocking step. For AD mouse brain staining of amyloid plaques, floating sections were placed in 1X Amylo-Glo ® RTD™ (Biosensis) staining solution for 10 min at room temperature without shaking. After staining, sections were washed in PBS 3 times for 5 minutes each and briefly rinsed in MiliQ DI water before being placed back in to PBS followed by blocking. Primary antibodies were added to staining solution (1X PBS, 0.2% Triton X-100, and 1% goat serum) at appropriate dilutions (see below) and incubated overnight at 4°C with slight shaking. The next day, sections were washed 3 times with PBS and stained with Alexa Fluor^® conjugated secondary antibodies at 1:400 for 1 h at room temperature with slight shaking in the dark. After secondary staining, sections were washed in PBS 3 times for 5 min and mounted on glass slides. After mounting, slides were cover slipped with DAPI-counterstain mounting media (Fluoromount, southern Biotech). Primary antibodies:

• mouse anti-β3Tubulin (1:500; BioLegend, 801201)

• mouse anti-human Cytoplasm (SC121,1:100; Takara Bio Inc., Y40410),

• mouse anti-human Nuclei (ku80, 1:100; Abcam, ab79220)

• chicken anti-GFAP (1:500; Abcam, ab4674)

• rabbit anti-Iba1 (1:500; Wako; 019-19741)

• goat anti-Iba1 (1:100; Abcam ab5076) * recommend use with Alexa Fluor 488 or 555 secondary antibody only.

• mouse anti-ITGB5 (1:500; Abcam, ab177004)

• mouse anti-MMP-9 (1:500; EMD Millipore, AB19016)

• mouse anti-human Mertk (1:500; BioLegend, 367602)

• rabbit anti-P2ry12 (1:125; Sigma; HPA014518)

• rabbit anti-Pros (1:500; Abcam, ab97387)

• rabbit anti-PU.1 (1:500; Cell Signaling Technology, 2266S)

• rabbit anti-human Tmem119 (1:100; Abcam, ab185333)

• goat anti-human Trem2 (1:100; R&D Systems, AF1828)

• rabbit anti-Tgfbr1 (1:500; Abcam, ab31013).

• rabbit anti-Cx3cr1 (1:1000; Bio-Rad/AbD Serotec, AHP1589)

Other primary antibodies used:

• rabbit anti-Amyloid Fibrils (OC) (1:1,000, EMD Millipore, AB2286)

• rabbit anti-Amyloid Oligomeric (A11) (1:1,000, EMD Millipore, AB9234)

• mouse anti-Amyloid 1-16aa (6e10) (1:1,000, BioLegend, 803001)

ADP migration and calcium imaging assays

Trans-well migration assays to ADP was performed as previously described (De Simone et al., 2010; Moore et al., 2015). iMGLs (5.5×10⁴ cells/well) were cultured in serum-free basal media without cytokines pre-exposed to DMSO or PSB0739 (50 μM, Tocris) for 1hr at 37ºC in 5% CO₂ cell culture incubator. Cells were then washed three times with basal medium and plated in trans-well migration chambers (5 μm polycarbonate inserts in 24 wells; Corning) containing Adenosine 5′-phosphate (ADP, 100 μM; Sigma) in the bottom chamber in 37ºC in 5% CO_2. After 4 hours, cells were washed three times with PBS (1x) and fixed in PFA (4%) for 15 minutes at room temperature. Cells were stained with Hoechst stain for 10 mins to visualize nuclei of cells. A blinded observer counted total cells per slide and then scrubbed cells off top surface using a cotton-swab, washed with PBS, and recounted to record migrated cells. Migration was reported as migrated over total cells per well. Fluorescent images of cells were captured using Olympus IX71 inverted microscope.

For calcium imaging, iMGLs were plated on poly-L-lysine-coated coverslips and 1 hour later were incubated with Fura-2-AM (Molecular Probes) calcium dye diluted in Ringer solution containing (in mM): NaCl 140, KCl 4.5, CaCl₂ 2, MgCl₂ 1, HEPES 10, glucose 10, sucrose 5, pH=7.4. After 1-hour incubation, the dye was washed out 3 times using Ringer solution and treated for 1 hour with either P2ry12 inhibitor PSB0739 (50 μM, Tocris) or Vehicle (DMSO) and used for experiments. Baseline Ca²⁺ signal (I₃₄₀/I₃₈₀) were measured for more than 100 s and then ADP (10 μM) was introduced under steady flow after baseline measurement. Ca²⁺ recordings were performed on Zeiss (Axiovert 35)-based imaging setup and data acquisition was conducted with Metafluor software (Molecular Devices). Data analysis was performed using Metafluor, Origin Pro, and Prism 6.0.

Phagocytosis Assays

iMGLs and MD-Mϕ, were incubated with mouse anti CD16/32 Fc-receptor block (2 mg/ml; BD Biosciences) for 15 minutes at 4°C. Cells were then stained with anti CD45-APC clone HI30 (Tonbo Biosciences) at 1:200 in flow cytometer buffer. Samples were then analyzed using Amnis Imagestreamer®^x Mark II Imaging Flow Cytometer (Millipore). E.coli, human synaptosome, fAβ, and BDTO phagocytosis was analyzed using the IDEAS software onboard Internalization Wizard algorithm. Additive free Anti-CD11b antibody (BioLegend, #301312) was used for CD11b blockade.

Fibrillar Aβ Preparation

Fibrillar fluorescent amyloid-beta (fAβ_1-42) was generated as described previously (Koenigsknecht-Talboo and Landreth, 2005). Fluorescently labeled Aβ peptide (Anaspec; Fremont, CA) was first dissolved in NH₄OH (0.1%) to 1 mg/ml, then further diluted to 100 μg/ml using sterile endotoxin-free water, vortexed thoroughly, and incubated at 37°C for 7 days. fAβ was thoroughly mixed prior to cell exposure.

BDTO Preparation

Brain-derived tau oligomers were purified by immunoprecipitation as described previously (Lasagna-Reeves et al., 2012). Tau oligomers were isolated by immunoprecipitation with the T22 antibody using PBS-soluble fractions of homogenates prepared from AD brain. These were then purified by fast protein liquid chromatography (FPLC) using PBS (pH 7.4). Additional analyses include Western blots to detect contamination with monomeric tau or large tau aggregates (tau-5, normally appear on top of the stacking gel) and using a mouse anti-IgG to identify non-specific bands. BDTOs were subsequently conjugated to pHrodo-Red per manufacturer’s protocol.

Human synaptosomes

The synaptosome preparation protocol was adapted from (Gylys et al., 2000). Human tissue samples were obtained at autopsy and minced, slowly frozen in 0.32 M sucrose with 10% DMSO and stored at −80 °C. To obtain a crude synaptosome fraction, tissue was thawed in a 37 °C water bath and homogenized in 10 mm Tris buffer (pH 7.4) with proteinase inhibitors (Roche) and phosphatase inhibitors (Sigma-Aldrich) using a glass/Teflon homogenizer (clearance 0.1–0.15 mm). The homogenate was centrifuged at 1000 g at 4 °C for 10 min, the supernatant was removed and centrifuged again at 10,000 g at 4 °C for 20 min. Resulting pellets were suspended in sucrose/Tris solution and stored at −80 °C. Synaptosomes were conjugated to pHrodo-Red per the manufacturer’s protocol.

Mesoscale Multiplex Cytokine and Chemokine Assay

iMGL culture media was replaced with basal media for 2 hours prior to stimulation with IFNγ (20 ng/ml), IL1β (20 ng/ml), and LPS (100 ng/ml) for 24 hours, after which cells were collected for RNA and conditioned media assessed for cytokine secretion. To simultaneously assess multiple cytokine and chemokine analytes from iMGL conditioned media, conditioned media from each treatment group was processed and analyzed using the V-PLEX human cytokine 30-plex kit (Mesoscale) per the manufacturer’s protocol.

Dot blot

Serial dilutions of proteins (2 μl) were blotted on a pre-wet nitrocellulose paper and allowed to dry. After drying, blots were blocked with 5% BSA in 1x Tris-buffered saline with Tween 20 (TBST) for 1 h at room temperature with slight shaking. Next, blots were incubated with primary antibodies (see below) at room temperature for 1 hour. Blots were then washed 3 times for 5 min each with TBST. Blots were then incubated with HRP conjugated secondary antibody (Santa Cruz) at 1:10,000 for 1 h at room temperature with mild shaking. After 1 h, blots were washed 3 times for 5 min each with TBST. After wash, blots were dried on filter paper and incubated with Pierce ECL Western blotting development substrate (Thermo Fisher Scientific) for 10 min in the dark. Blots were imaged on ChemiDoc XRS+ imaging system (BioRad).

AD-GWAS qPCR Primers

The following validated and available Taqman primers were used: APOE Hs00171168_m1, CR1 Hs00559342_m1, CD33 Hs01076281_m1, ABCA7 Hs01105117_m1, TREM2 Hs00219132_m1, TREML2 Hs01077557_m1, TYROBP (DAP12) Hs00182426_m1, PICALM Hs00200318_m1, CLU Hs00156548_m1, MS4A6A Hs01556747_m1, BIN1 Hs00184913_m1, CD2AP Hs00961451_m1, CASS4 Hs00220503_m1, MEF2C Hs00231149_m1, DSG2 Hs00170071_m1, MS4A4A Hs01106863_m1, ZCWPW1 Hs00215881_m1, INPP5D Hs00183290_m1, and PTK2B Hs00169444_m1.

RNA-seq analysis

RNA-seq reads were mapped to the hg38 reference genome using STAR (Dobin et al., 2013) aligner and mapped to Gencode version 24 gene annotations using RSEM(Li and Dewey, 2011). Genes with expression (< 1 FPKM) across all samples were filtered from all subsequent analysis. Differential gene expression analysis was performed on TMM normalized counts with EdgeR (Robinson et al., 2010). Multiple biological replicates were used for all comparative analysis. A p-value ≤ 0.001 and a 2-fold change in expression were used in determining significant differentially expressed genes for respective comparisons. PCA analysis was performed using the R package “rgl” and plotted using “plot3d”. Clustering was performed using R “hclust2” and visualized using Java Tree View 3.0 (http://bonsai.hgc.jp/~mdehoon/software/cluster/software.htm). Differential gene analysis between groups was performed using the R package “limma” and significant genes (adjusted p < 0.01) were used for Gene ontology and pathway analysis. Gene ontology and pathway analysis was performed using Enrichr data base (http://amp.pharm.mssm.edu/Enrichr/) (Chen et al., 2013; Kuleshov et al., 2016).

Statistical Analysis

Statistical analysis was performed using Graphpad Prism 7 software. Comparisons involving more than two groups utilized one-way ANOVA followed by Tukey’s post hoc test and corrected p-values for multiple comparisons were reported. Comparison’s with more than two groups and comparing to a control or vehicle group utilized one-way ANOVA followed by Dunnett’s post hoc test with corrected p-values for multiple comparisons reported. Two-Way ANOVA were followed by Sidak’s multiple-comparison post-hoc test Comparisons of two groups utilized two-tailed Students t-test. All differences were considered significantly different when p<0.05. Statistical analysis for RNA-sequencing is detailed above and all further statistical analysis details are reported in the figure legends.

Data and Software Availability

Raw and normalized RNA-sequence data can be obtained at NCBI. The GEO Accession Super Series ID number for the data reported in this paper is GEO: GSE89189.