Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner

Article

Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner Graphical Abstract

Authors Moumita Datta, Ori Staszewski, Elena Raschi, ..., Patrick Matthias, Melanie Meyer-Luehmann, Marco Prinz

Correspondence [email protected] (M.D.), [email protected] (M.P.)

In Brief Little is known about the epigenetic signals that control microglia function in vivo. Datta et al. show that histone deacetylases Hdac1 and Hdac2 are essential for microglial survival and expansion during development but not during steady state. In Alzheimer’s disease mouse model, deletion of microglial Hdac1 and Hdac2 reduces amyloid pathology and improves cognitive function.

Highlights d

Combined loss of Hdac1 and Hdac2 reduces microglia survival during development

d

Hdac1 and Hdac2 are functionally redundant in adult microglia during steady state

d

In Alzheimer’s disease model, microglial Hdac1 and Hdac2 loss reduces amyloid burden

d

Hdac1 and Hdac2 loss enhances microglial amyloid phagocytosis and improves cognition

Datta et al., 2018, Immunity 48, 1–16 March 20, 2018 ª 2018 Elsevier Inc. https://doi.org/10.1016/j.immuni.2018.02.016

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

Immunity

Article Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner Moumita Datta,1,7,8,* Ori Staszewski,1,8 Elena Raschi,2 Maximilian Frosch,1 Nora Hagemeyer,1 Tuan Leng Tay,1 Thomas Blank,1 Mario Kreutzfeldt,3 Doron Merkler,3 Stephanie Ziegler-Waldkirch,4 Patrick Matthias,5 Melanie Meyer-Luehmann,4 and Marco Prinz1,6,9,* 1Institute

of Neuropathology, Faculty of Medicine, University of Freiburg, Freiburg 79106, Germany

2Department of Experimental, Diagnostic and Speciality Medicine, School of Medicine, Alma Mater Studiorum University of Bologna, Bologna

40126, Italy 3Department of Pathology and Immunology, Geneva Faculty of Medicine, Geneva 1211, Switzerland 4Department of Neurology, Faculty of Medicine, University of Freiburg, Freiburg 79106, Germany 5Friedrich Miescher Institute for Biomedical Research, Basel 4002, Switzerland 6BIOSS Centre for Biological Signalling Studies, University of Freiburg, Freiburg 79104, Germany 7Present address: Institute for Immunology, Faculty of Medicine, Ulm University, Ulm 89081, Germany 8These authors contributed equally 9Lead Contact *Correspondence: [email protected] (M.D.), [email protected] (M.P.) https://doi.org/10.1016/j.immuni.2018.02.016

SUMMARY

Microglia as tissue macrophages contribute to the defense and maintenance of central nervous system (CNS) homeostasis. Little is known about the epigenetic signals controlling microglia function in vivo. We employed constitutive and inducible mutagenesis in microglia to delete two class I histone deacetylases, Hdac1 and Hdac2. Prenatal ablation of Hdac1 and Hdac2 impaired microglial development. Mechanistically, the promoters of pro-apoptotic and cell cycle genes were hyperacetylated in absence of Hdac1 and Hdac2, leading to increased apoptosis and reduced survival. In contrast, Hdac1 and Hdac2 were not required for adult microglia survival during homeostasis. In a mouse model of Alzheimer’s disease, deletion of Hdac1 and Hdac2 in microglia, but not in neuroectodermal cells, resulted in a decrease in amyloid load and improved cognitive impairment by enhancing microglial amyloid phagocytosis. Collectively, we report a role for epigenetic factors that differentially affect microglia development, homeostasis, and disease that could potentially be utilized therapeutically.

INTRODUCTION Microglia are the resident tissue macrophages of brain, crucially involved in scavenging of dying cells, pathogens, and molecules through pathogen-associated molecular pattern (PAMP) receptors (Casano and Peri, 2015; Ransohoff, 2016). Dysregulation of steady-state microglia is currently considered the pathogenic

basis for a group of neurodegenerative and neuroinflammatory conditions, called ‘‘microgliopathies’’ (Prinz and Priller, 2014). These include roles for several microglia molecules such as colony-stimulating factor 1 receptor (CSF1R) in hereditary diffuse leukoencephalopathy with spheroids (HDLS) (Rademakers et al., 2011), the ubiquitin-specific peptidase (USP) 18 (Goldmann et al., 2015) and CD33 in Alzheimer’s disease (AD) (Hollingworth et al., 2011), and Trem2 in frontotemporal dementia (Guerreiro et al., 2013). Together with the macrophages at CNS interfaces such as the meninges and the perivascular spaces, microglia are derived from erythromyeloid precursors (EMP) from early prenatal sources (Goldmann et al., 2016). They exhibit virtually no turnover with circulating monocytes (Mildner et al., 2007) and a remarkable longevity based on a low random self-renewal rate coupled to slow homeostatic apoptosis (Ajami et al., 2007; Tay et al., 2017). Microglia identity and homeostasis are tightly regulated by alterations of the microenvironment within (Lavin et al., 2014) and outside (Erny et al., 2015) of the brain as well as by several transcription factors such as PU.1 and IRF8 or cytokines such as interleukin (IL) 34, macrophage colony-stimulating factor (MCSF), and their receptor CSF1R (reviewed in Kierdorf and Prinz, 2017). Macrophage activation requires precise regulation of gene expression that is governed by epigenetic mechanisms, such as DNA methylation and chromatin and histone modifications (Amit et al., 2016). However, very little is known about epigenetic factors that regulate microglia identity during their development from EMP, during homeostasis, and in microglia function. Histone deacetylases (HDACs) are histone-modifying enzymes that, along with histone acetyl transferases (HAT), reversibly regulate acetylation status of histones. Among the different classes of HDACs, class I HDACs (HDAC1, 2, 3, 8) can assemble into multi-component co-repressor complexes and often act as transcriptional repressors (Das Gupta et al., 2016). Immunity 48, 1–16, March 20, 2018 ª 2018 Elsevier Inc. 1

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

A

C

B

D

E

F

G

H

I

(legend on next page)

2 Immunity 48, 1–16, March 20, 2018

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

In this study, we investigated the role of HDAC1 and HDAC2 in the maintenance of microglia identity and function. Specifically, we used constitutive and inducible CX3C chemokine receptor 1 (Cx3cr1) promoter-driven Cre recombinase combined with conditional alleles for both Hdac1 and Hdac2 to generate mice whose microglia lacked these enzymes from development, in normal adulthood, or during neurodegeneration. Prenatal ablation of Hdac1 and Hdac2 caused spontaneous microglia impairment, including blockage of proliferation and enhanced apoptosis while postnatal deletion was largely compatible with microglial viability and function. In contrast, Hdac1 and Hdac2 depletion was able to prevent microglia exhaustion and dysfunction in a mouse model of AD leading to reduced amyloid load and improved cognitive functions. RESULTS Expression of Hdac1 and Hdac2 and Their Targets Are Developmentally Regulated in Microglia Microglia development from early yolk sac precursor to adult microglia follows a detailed stepwise process (Kierdorf et al., 2013). To determine whether Hdac1 and Hdac2 and their target genes were differentially expressed during microgliogenesis, we examined their messenger (m)RNA levels using microarray as described recently (Figure 1; Hagemeyer et al., 2016). Specifically, we analyzed CD31+c-kit+ EMPs from embryonic day E8.0 and their progeny, namely CD45+CX3CR1loF4/80lo A1 cells (from E9.0), CD45+CX3CR1hiF4/80hi A2 cells (from E9.0), embryonic E14.5 CD45loCX3CR1hiF4/80hi microglia (eMG), and adult CD45loCX3CR1hi microglia as defined previously (Kierdorf et al., 2013). First we determined whether Hdacs were significantly enriched as upstream regulators in the promoters of genes altered from EMP to adult stage. Hdac1 displayed the most significant enrichment (Figure 1A, left). Analysis of the activation Z-score for all Hdacs also revealed maximal inhibition of gene expression by Hdac1 and Hdac2 from EMP to adult stage (Figure 1A, right). We then investigated the expression of Hdac1 and Hdac2 during microglial development. From unbiased microarray data (Figure 1B) as well as from quantitative real-time (qRT)-PCR (Figure 1C), we found that both genes were highly induced during EMP stage, after which their expression was gradually decreased, reaching lowest level in the adult stage. Next, we analyzed the regulation of Hdac1 and Hdac2 target genes in microglia. As depicted in

the heatmap (Figure 1D), the magnitude of alteration (both up- and downregulation) of Hdac1 target genes was highest at EMP stage and gradually diminished as microglia progressed along developmental path. This was also evident from the reduction in network size for Hdac1 target genes from EMP (Figure 1E) to eMG (Figure 1F). Similarly, magnitude of alteration for Hdac2 target genes also showed gradual reduction from EMP to adult microglia as evident from the heatmap (Figure 1G) and the size of gene network (Figures 1H and 1I). Thus, Hdac1- and Hdac2-mediated gene regulations both in terms of number of genes and their relative expression level were maximal in microglia prior to birth and were reduced in adult stage. Prenatal Hdac1 and Hdac2 Ablation Induces Microglia Malformation To study the functional role of Hdac1 and Hdac2 on pre- and neonatal microglia, where these epigenetic modifiers were highly induced, we probed the effect of Hdac1 and Hdac2 deficiency by generating Cx3cr1Cre Hdac1fl/flHdac2fl/fl mice. In these mice, Cre recombinase is expressed starting from E9.0 in primitive A2 yolk sac macrophages that give rise to microglia (Hagemeyer et al., 2016). Cx3cr1Cre Hdac1fl/flHdac2fl/fl (Cre+) and the corresponding control Hdac1fl/flHdac2fl/fl (Cre
) mice were analyzed at defined time points of pre- and postnatal development (Figure 2A) and Hdac1 and Hdac2 absence in microglia was validated by qRT-PCR (Figure 2B). To check whether the deletion was specific for microglia, we also measured Hdac1 and Hdac2 expression in neurons, astrocytes, and oligodendrocytes using both immunofluorescence and qRT-PCR (Figures S1A and S1B). Only microglia from Cre+ mice showed significantly lowered Hdac1 and Hdac2 expression on protein and mRNA level and not the other cell types. We next analyzed the impact of Hdac1 and Hdac2 deficiency on microglia by quantifying Iba-1+ cells and observed reduction of cortical microglia number in Cre+ animals at E16.5, postnatal day (P) 0 and at 2 weeks of age compared to the Cre
littermates (Figure 2C). Differences in microglia cell numbers were not present 4 weeks later, suggesting homeostatic proliferation of microglia that occurs when microglia have to fill spatial niches as described previously (Elmore et al., 2014). We observed higher microglia proliferation upon pulse labeling with 5-ethynyl-20 -deoxyuridine (EdU) in Cre+ mice compared to controls (Figures S2A–S2E). IMARIS-based three-dimensional (3D) morphometric measurements of microglia revealed significantly shorter processes and reduced numbers of

Figure 1. Hdac1 and Hdac2 Are Developmentally Regulated during Microglia Development (A) Representation of Hdacs as upstream regulators of genes altered in different stages of microglia development: EMP, A1, A2 population, eMG, and adult microglia (P42) as assessed from unbiased microarray analysis (n = 3, one experiment) in wild-type mice. The p value (left) and the activation Z score (right) for each Hdac as upstream regulator is shown as color-coded heatmap. (B) Relative mRNA expression of Hdac1 and Hdac2 in microglia during development as obtained from the microarray depicted in (A). Bars represent mean ± SEM. Expression is normalized with respect to adult microglia. (C) qRT-PCR for Hdac1 and Hdac2 expression in the aforesaid stages of microglia development (n = 3, except EMP n = 2, one experiment). Bars represent mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001. (D) Heatmap depicting expression of Hdac1 target genes (top 20 up- and downregulated) at the different time points of microglia development. Color coding: Log2 (fold change). (E and F) Network analysis showing relative interactions between Hdac1 and its target genes altered at the EMP (E) and eMG stages (F). (G) Heatmap depicting expression of Hdac2 target genes at the different time points of microglia development. Color coding: Log2 (fold change). (H and I) Network analysis showing relative interactions between Hdac2 and its target genes altered at the EMP (H) and A2 stages (I).

Immunity 48, 1–16, March 20, 2018 3

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

A

B

C

D

E

Figure 2. Combined Loss of Hdac1 and Hdac2 in Microglia prior to Birth Largely Impairs Microglial Development (A) Scheme of the experimental timeline for the analysis of Cx3Cr1Cre Hdac1fl/flHdac2fl/fl mice. (B) Relative mRNA expression of Hdac1 and Hdac2 in microglia isolated from Hdac1fl/flHdac2fl/fl (Cre
, black circles) and Cx3Cr1Cre Hdac1fl/flHdac2fl/fl (Cre+, red circles) mice at P0, 2, 6, and 16 weeks. Data obtained from qRT-PCR (n = 3–4 mice for each genotype at each time point, two experiments) and expression is normalized to Cre
animals of the corresponding time point. Bars represent mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001. (C) Quantification of Iba1+ microglia in the cortex at the indicated time points (n = 3–6 mice for each genotype each time point, two experiments). Bars represent mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. (D) Upper panels: Representative confocal images of Iba1+ microglia (red) in the cortex of Cre
and Cre+ mice at the specified time points. 40 ,6-diamidino2-phenylindole (DAPI) is in blue. Scale bar: 15 mm. Lower panels: 3D reconstruction of cortical microglia at different time points. Scale bar: 10 mm. (legend continued on next page)

4 Immunity 48, 1–16, March 20, 2018

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

segments, branching, and terminal points (Figures 2D and 2E). The differences in microglial cell numbers were normalized at 6 weeks of age, whereas the morphological differences persisted until adulthood. Cx3cr1Cre Hdac1fl/flHdac2fl/fl mice exhibited no abnormalities of the CNS gross anatomy on a histological level. The number of NeuN+ neurons, GFAP+ astrocytes, and the integrity of the blood-brain barrier were not altered, indicating an effect primarily in the microglia compartment of CNS (Figure S1C). As the Cx3cr1Cre-mediated gene deletion targets all CX3CR1+ myeloid cell types including those in circulation and bone marrow, we utilized the tamoxifen (TAM)-inducible Cx3cr1CreERT2 line that targets only long-living tissue macrophages such as microglia and perivascular and meningeal macrophages in the CNS (Goldmann et al., 2013, 2016; Yona et al., 2013). Resulting Cx3cr1CreERT2 Hdac1fl/flHdac2fl/fl mice were challenged with TAM at E13 and microglia were subsequently analyzed at defined pre- and postnatal time points (Figure S3A). This targeting strategy led to successful deletion of Hdac1 and Hdac2 genes in developing microglia (Figure S3B) and the resulting cells phenocopied the constitutively deleted microglia with reduced numbers and malformed morphology albeit with a slightly reduced extent that might be due to less-efficient gene deletion using this approach (Figures S3C–S3E). Taken together, these data indicate that microglial abnormality in the absence of Hdac1 and Hdac2 during development is due to cell-intrinsic effects rather than non-cell-autonomous mechanisms. Enhanced H3K9ac and H3K27ac in Microglia Lacking Hdac1 and Hdac2 during Development Results in Impaired Proliferation and Apoptosis Induction To obtain mechanistic insights into the microglial phenotype in the absence of Hdac1 and Hdac2, we performed an unsupervised clustering analysis from RNA-seq data of Hdac1- and Hdac2-deficient microglia compared to wild-type cells in both constitutive (Figures 3A–3D) and TAM-inducible (Figures 3F– 3H) mice. Hierarchical clustering of all significantly deregulated genes (p < 0.05) in constitutive Cx3cr1Cre Hdac1fl/flHdac2fl/fl mice (Cre+) compared to Cre
controls divided the P0 group irrespective of Cre status from the other two time points (Figure 3A). This correlated with previously described data indicating that immature microglia at birth have distinct transcriptional signatures compared to young or adult stage (Amit et al., 2016). The extent of gene expression alterations in Cre+ microglia was maximal at birth (P0) depicting 4,338 differentially regulated genes (p < 0.05) and gradually decreased at 2 weeks (1,820 altered genes, p < 0.05) and at 6 weeks (692 altered genes, p < 0.05). This supported the notion that Hdac1 and Hdac2 were indispensable for early microglia development. To gain insights into which cellular processes and pathways were altered, we classified altered genes based on gene ontology (GO) (Figure 3B). Significantly altered biological functions included cell morphology, cell growth and proliferation, cell death and survival, cellular compromise, and cell cycle. This was in agreement

with previous studies that reported cell cycle alterations, growth arrest, and cell death in the absence of Hdac1 and Hdac2 (Lagger et al., 2002; Montgomery et al., 2007). When compared between different time points, the magnitude of alteration of most of these functions showed a gradual decrease from P0 to 6 weeks. Alteration of the cell cycle cluster was significant only at the 2 week time point, most likely due to the observed rapid microglia replenishment in the Cre+ animals after birth (Figures S2B–S2E). Among the significantly altered pathways, we found increased apoptosis signaling (Figure 3C) that could be further confirmed by enhanced TUNEL labeling in Cre+ embryonic microglia (Figure 3E). Molecular functions such as ‘‘retraction of cellular protrusion’’ and ‘‘collapse of growth cone’’ were also significantly altered in Cre+ microglia, supporting the observed changes in microglial morphology (Figure 3D). Microglia from TAM-inducible Cx3cr1CreERT2 Hdac1fl/fl Hdac2fl/fl (Cre+) mice showed similar changes in gene expression pattern when compared to Cre
controls (Figure 3F). Genetic changes, however, were less pronounced in these mice compared to the constitutive Cre line. Maximum number of altered genes was observed at E16.5 (1,697 differentially expressed genes in Cre+ compared to Cre
, p < 0.05) (Figure 3F). Yet, 876 differentially expressed genes were found at P0 (p < 0.05) and no changes were apparent at 2 and 6 weeks. Accordingly, GO functional clusters showed changes primarily at the E16.5 and P0 (Figure 3G) with concomitant robust alteration in the apoptosis signaling pathway (Figure 3H). Hence, elevated numbers of apoptotic TUNEL+ microglia were found in cortical brain sections of Cre+ animals compared to Cre
controls (Figure 3I). To investigate whether the absence of Hdac1 and Hdac2 altered histone acetylation in microglia, we performed genome-wide profiling (ChIP-seq) of H3K9 and H3K27 acetylation (ac) in microglia from 6-week-old Cx3cr1Cre Hdac1fl/fl Hdac2fl/fl mice and their corresponding littermate controls to infer regions of active chromatin (Figure 3J). Although global levels of H3K9ac and H3K27ac showed only subtle changes in microglia lacking Hdac1 and Hdac2, deeper inspection revealed a greater abundance of H3K9ac and H3K27ac at the proximal promoters of several genes regulating cell cycle and cell activation, e.g., Sema6d, Cdkn2c, Ifnar2, and Kcna3 (Figures 3J and S4A). These genes were also upregulated on the mRNA level (Figure 3C). We next classified the genes having altered H3K9ac and H3K27ac level according to their molecular functions and compared them with that of the RNA-seq data (Figure S4B). Similar functions such as cellular development, cell growth and proliferation, and cellular movement were found to be altered in both ChIP-seq and RNA-seq experiments in Cx3cr1Cre Hdac1fl/flHdac2fl/fl mice, suggesting a positive correlation of the data from these two methods. Collectively, these data indicate that absence of Hdac1 and Hdac2 during early prenatal microgliogenesis changes the acetylation marks of defined promoter regions that fundamentally orchestrate microglia survival and expansion, especially by modulating apoptosis, activation, morphology, and cell cycle.

(E) IMARIS based morphometric analysis of microglia from Cre
and Cre+ mice at E16.5, P0, 2, 6, and 16 weeks. Bars represent mean ± SEM of 3–5 mice per genotype per time point, two experiments. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001. See also Figures S1–S3.

Immunity 48, 1–16, March 20, 2018 5

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

A

B

C

F

D

E

G

H

I

J

Figure 3. Absence of Hdac1 and Hdac2 during Development Alters Microglial Histone Acetylation and Transcriptomic Profile (A) Hierarchical clustering of genes that are significantly altered (p < 0.05) between microglia from Cx3Cr1Cre Hdac1fl/flHdac2fl/fl (Cre+) and control Hdac1fl/fl Hdac2fl/fl (Cre
) mice at different time points as obtained from RNA-seq analysis (one experiment). Heatmap displays z-transformed expression values. (B) Line graph showing the alteration of some of the significant cellular functions in microglia from Cre+ mice over different time points. The statistical significance,
Log10 (p value), corrected by Benjamini-Hochberg method for each function is plotted against the time point. (C and D) Heatmaps showing expression (Log2 fold change) of genes involved in some of the significantly altered pathways (C) and biological functions (D) in microglia from Cre+ mice over different time points. (E) Upper panel: Representative confocal images showing co-localization (white arrow) of Iba1+ (green) microglia with TUNEL+ (red) nuclei in the cortex of Cre
and Cre+ mice at E16.5. Scale bar: 15 mm. Lower panel: Quantification of TUNEL+ microglia in the cortex of Cre
(black circles) and Cre+ (red circles) mice. Bars represent mean ± SEM of 3–4 mice per genotype from one experiment. Statistical analysis: unpaired t test, *p < 0.05. (F) Hierarchical clustering of genes that are significantly altered (p < 0.05) between microglia from Cx3Cr1CreERT2 Hdac1fl/flHdac2fl/fl and control mice at different time points as obtained from RNA-seq analysis (one experiment). Heatmap displays z-transformed expression values. (G) Line graph showing the alteration of some of the significant cellular functions in microglia from Cx3Cr1CreERT2 Hdac1fl/flHdac2fl/fl mice over different time points. The statistical significance
Log10 (p value), corrected by Benjamini-Hochberg method for each function, is plotted against the time point. (H) Heatmap showing expression (Log2 fold change) of genes involved in apoptosis signaling pathway in microglia from Cx3Cr1CreERT2 Hdac1fl/flHdac2fl/fl mice over different time points. (I) Upper panel: Representative confocal images showing co-localization (white arrow) of Iba1+ (green) microglia with TUNEL+ (red) nuclei in the cortex of Hdac1fl/fl Hdac2fl/fl (Cre
) and Cx3Cr1CreERT2 Hdac1fl/flHdac2fl/fl (Cre+) mice at E16.5. Scale bar: 15 mm. Lower panel: Quantification of TUNEL+ microglia in the cortex of Cre
(black circles) and Cre+ (red circles) mice. Bars represent mean ± SEM of 3–4 animals per genotype from one experiment. Statistical analysis: unpaired t test, *p < 0.05. (J) ChIP-seq data showing the coverage of H3K9 acetylation (Ac) in the promoter region for the selected genes in Hdac1fl/flHdac2fl/fl (Cre
) and Cx3cr1Cre Hdac1fl/flHdac2fl/fl (Cre+) animals at 6 weeks of age. The first exon (Ex1) for each gene is indicated below. RPKM values are shown in the top left for each panel. Average of two replicates from one experiment is shown. See also Figure S4.

Hdac1 and Hdac2 Are Redundant for Steady-State Maintenance of Adult Microglia To examine the effect of Hdac1 and Hdac2 deficiency in microglia in steady state, we used Cx3cr1CreERT2 Hdac1fl/flHdac2fl/fl (Cre+) mice and performed several analyses up to 24 weeks 6 Immunity 48, 1–16, March 20, 2018

following TAM treatment in adult (6 week) mice (Figure 4A). Upon treatment, microglia from Cre+ mice showed significant downregulation of Hdac1 and Hdac2 compared to control Cre
(Figure 4B). Mutant mice, however, did not develop any overt neurological phenotype for at least 6 months after

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

A

B

C

D

E

Figure 4. Redundant Function of Hdac1 and Hdac2 in Adult Microglia during Steady State (A) Scheme of the experimental timeline for the analysis of Cx3Cr1CreERT2 Hdac1fl/flHdac2fl/fl (Cre+) and Hdac1fl/flHdac2fl/fl (Cre
) control animals. (B) Relative mRNA expression of Hdac1 and Hdac2 in microglia isolated from Cre
(black circles) and Cre+ (red circles) mice at the indicated time points. Data obtained from qRT-PCR (n = 3 mice for each genotype each time point, one experiment) and expression is normalized to Cre
mice. Bars represent mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001. (C) Histological quantification of cortical Iba1+ microglia (n = 3 mice each genotype each time point, one experiment). Bars represent mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test; ns, not significant. (D) Upper panels: Representative confocal images of Iba1+ microglia (red) in the cortex of Cre
and Cre+ mice brain at the specified time points. DAPI is in blue. Scale bar: 15 mm. Lower panels: Representative 3D reconstruction of cortical microglia. Scale bar: 10 mm. (E) Morphometric analysis of Iba1+ cortical microglia from Cre
and Cre+ mice. Bars represent mean ± SEM of 3 mice per genotype, per time point from one experiment. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test; ns, not significant. See also Figure S5.

Immunity 48, 1–16, March 20, 2018 7

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

B

A

C

D

F

E

G

I

H

J

K

L

M

(legend on next page)

8 Immunity 48, 1–16, March 20, 2018

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

induction. We next performed thorough histopathological analysis of microglia cell number and morphology. In contrast to our findings during embryogenesis, we observed no change either in microglia number or morphology in Cre+ mice (Figures 4C–4E). Flow cytometric analysis of adult microglia from Cre+ mice showed no differences in the expression of surface markers CD11b and CD45 (Figures S5A–S5C) that are known to be upregulated upon activation (Prinz and Priller, 2017). In addition, mRNA expression of transcription factors, chemokines, surface receptors, and activation markers did not show major changes in Cre+ microglia at the investigated time points (Figures S5D– S5G). Collectively, these data suggest that the absence of Hdac1 and Hdac2 in adulthood is not critical for steady-state microglia and support the notion that expression of Hdac1and Hdac2-associated genes decline after birth. Genetic Ablation of Hdac1 and Hdac2 from Adult Microglia Reduces Amyloid Plaque and Rescues Spatial Learning and Memory Deficits Unlike in the healthy adult brain, microglia are highly proliferative and actively migrate during neurodegeneration where they are key effector cells, together with other innate myeloid cells such as monocytes, that fundamentally modulate disease burden (Colonna and Butovsky, 2017; Prinz and Priller, 2017). To test whether Hdac1 and Hdac2 were detrimental or beneficial for microglia effector functions during neurodegeneration, we crossed Cx3cr1CreERT2 Hdac1fl/flHdac2fl/fl mice with 5x familial AD (5xFAD) transgenic line that recapitulates major features of AD (Oakley et al., 2006). In the 5xFAD line, amyloid plaques start to appear at around 8 weeks of age, whereas most of the cortex and hippocampus are filled with plaques by 24–32 weeks. We therefore injected 5xFAD Cx3cr1CreERT2

Hdac1fl/flHdac2fl/fl mice (FAD DKO) with TAM at 6 weeks prior to the onset of AD pathology and disease pathogenesis was subsequently monitored (Figure 5A). TAM induction led to significant deletion of Hdac1 and Hdac2 mRNA in microglia in these mice (Figure 5B). Immunofluorescence staining using 6E10 antibody showed robust deposition of Abeta (Ab) plaques in the cortex (Figures 5C and 5D) and hippocampus (Figures 5C and 5E) of FAD DKO mice that was indistinguishable from FAD at 24– 28 weeks but strongly reduced at the age of 44–48 weeks (Figures 5C–5E and S6A). Quantification of plaque area and intensity confirmed a reduction of both parameters in the hippocampus of FAD DKO animals at 44–48 weeks (Figures S6B–S6E, upper panel). Under these conditions, overall microglia abundance as measured by Iba1 immunoreactive area or intensity (Figures S6B–S6E, lower panel) was unchanged in the experimental groups. Both soluble (Figure 5F) and insoluble (Figure 5H) Ab42 were significantly decreased in the cortex and hippocampus but not in the cerebellum of FAD DKO mice at 44–48 weeks when compared to FAD. For Ab40, we could not detect any change in the soluble fraction (Figure 5G) but found a reduction in insoluble Ab40 in the cortex and hippocampus of FAD DKO at 44–48 weeks (Figure 5I). To determine whether the absence of microglial Hdac1 and Hdac2 changed amyloid precursor protein (APP) production, we measured APP in the cortex and hippocampus by immunohistochemistry (Figure S6F) and western blot (Figure S6G) and found no detectable changes. To investigate whether reduced plaque burden in FAD DKO mice led to cognitive improvement, we performed Morris water maze (MWM) tests for spatial learning and memory function. As indicated by the learning curve, FAD mice showed severe learning disability compared to wild-type (WT) while this phenotype in FAD DKO mice was partially rescued (Figure 5J). During

Figure 5. Deletion of Hdac1 and Hdac2 in Adult Microglia Reduces Amyloid Burden and Improves Spatial Learning and Memory in Mouse Model of AD (A) Experimental timeline for the analysis of 5xFAD Cx3Cr1CreERT2 Hdac1fl/flHdac2fl/fl mice. (B) Relative mRNA expression of Hdac1 and Hdac2 in microglia isolated from non-FAD Hdac1fl/flHdac2fl/fl (wild-type WT, black circles, n = 9), 5xFAD Hdac1fl/flHdac2fl/fl (FAD, blue circles, n = 9), and 5xFAD Cx3Cr1CreERT2 Hdac1fl/flHdac2fl/fl (FAD DKO, red circles, n = 7) mice at 24–28 weeks of age. Data obtained from qRT-PCR from three experiments and expression are normalized to WT. Bars represent mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001. (C) Representative confocal images showing 6E10 immunoreactive amyloid plaques (green) and Iba1+ microglia (red) in the cortex and hippocampus of FAD and FAD DKO animals at 24–28 weeks (upper panel) and 44–48 weeks (lower panel). DAPI is in blue. Scale bar: 40 mm. White arrow, plaque; white arrowhead, 6E10 immunoreactive neuron. Inset: zoomed image of plaque, scale bar: 20 mm. (D and E) Histological quantification of amyloid plaques in the cortex (D) and hippocampus (E) of FAD (blue circles) and FAD DKO (red circles) mice at the specified age. Bars represent mean ± SEM of 4–8 mice per genotype per time point from three experiments. Statistical analysis: two-way ANOVA, multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. (F and G) ELISA-based quantification of soluble Ab42 (F) and 40 (G) species in the cortex, hippocampus, and cerebellum of FAD and FAD DKO mice at the specified age. Bars represent mean ± SEM of 4–8 mice per genotype per time point, two experiments. Statistical analysis: two-way ANOVA, multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001; ns; not significant. (H and I) Quantification of insoluble (guanidine hydrochloride extractable) Ab42 (H) and 40 (I) species in the cortex and hippocampus of FAD and FAD DKO mice at the specified age. Bars represent mean ± SEM of 4–8 mice per genotype per time point, two experiments. Statistical analysis: two-way ANOVA, multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. (J) Learning curve for WT (black circles, n = 14), FAD (blue circles, n = 14), and FAD DKO (red circles, n = 13) mice in Morris water maze test at 44–48 weeks of age. Mean escape latency (in seconds) of total 4 trials for each mouse was plotted over 7 days of training time. Each point represents mean ± SEM from two experiments. Statistical analysis: two-way ANOVA, *p < 0.05. (K) Graph representing percentage of time spent in the target quadrant by WT, FAD, and FAD DKO mice on the day of probe trial. Bar: mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05. (L) Percentage alteration in T-maze showing spatial working memory defects in 44– to 48-week-old FAD (n = 7) compared to either WT (n = 7) or FAD DKO (n = 4) mice. Bar: mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001. (M) Repetitive re-entry into same arm with impaired spatial working memory in FAD mice compared to either WT or FAD DKO in T-maze test. Bar: mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001. See also Figure S6.

Immunity 48, 1–16, March 20, 2018 9

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

A

B

C

E

D

G

F

I

H

J

K

(legend on next page)

10 Immunity 48, 1–16, March 20, 2018

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

probe trials, FAD DKO mice also spent significantly more time in the target quadrant compared to FAD, indicating a better memory function in FAD DKO (Figure 5K). Furthermore, when tested in a T-maze paradigm, FAD DKO mice showed significantly improved working and reference memory compared to FAD as evidenced by increased percent of arm alteration (Figure 5L) and minimal number of re-entry into the same arm of the maze (Figure 5M). However, at this stage, we did not observe any significant increase in synapse number in FAD DKO mice compared to FAD (Figures S6H–S6J). Finally, the absence of Hdac1 and Hdac2 in microglia did not change mouse behavior per se as documented in unaltered MWM test for Hdac1fl/flHdac2fl/fl and Cx3cr1CreERT2 Hdac1fl/flHdac2fl/fl mice on a non-FAD background (Figures S6K–S6O). Collectively, our data show a beneficial role of Hdac1 and Hdac2 depletion in microglia in an AD mouse model with reduction of amyloid plaque burden and improvement in learning and memory. Lack of Hdac1 and Hdac2 Increases Microglial Phagocytosis of Amyloid Plaques It has been proposed that plaque-associated microglia are critical in restricting senile plaque formation in AD mouse models by improved migration and increased phagocytosis compared to non-pathology-linked microglia (Keren-Shaul et al., 2017; Meyer-Luehmann and Prinz, 2015). We thus investigated whether Hdac1 and Hdac2 deficiency in microglia altered cellular phagocytosis. We quantified Iba1+ microglia that expressed the lysosomal-associated membrane protein 2 (Lamp-2), designating activated phagocytic microglia (Goldmann et al., 2015), in the cortex and hippocampus of FAD and FAD DKO mice at 24–28 weeks (Figure 6A). Significant increase in Lamp2 intensity in both brain regions of FAD DKO mice was observed (Figures 6B and 6C). This indicated increased phagocytic event in microglia in the absence of Hdac1 and Hdac2. To check whether mutant microglia were more efficient in engulfing plaques, we performed co-localization analysis for Iba1+ microglia and 6E10+ plaques in FAD and FAD DKO mice. Our analysis revealed significantly increased co-localization of 6E10+ plaques

with microglia in hippocampus of FAD DKO (Figures 6D and 6E). We also performed 3D surface reconstruction for plaque-associated microglia (Figures 6F and 6G) as recently described (Paolicelli et al., 2017). We found significantly enhanced Ab engulfment by FAD DKO microglia in the hippocampus compared to FAD (Figure 6H). We further examined the number of plaque-associated microglia and found a small but significant increase in the average number of microglia associated with plaques in FAD DKO mice (Figure 6I). To explore on the molecular level, we performed unbiased transcriptomic analysis of ex vivo isolated microglia from 24- to 28-week-old mice. Data revealed significantly increased expression (p < 0.05) of genes related to phagocytosis in FAD DKO microglia (Figure 6J). These differently expressed genes included apolipoprotein E (Apoe), AXL receptor tyrosine kinase (Axl), and fms related tyrosine kinase 1 (Flt1). All these genes are known to promote phagocytic clearance of Ab (Grommes et al., 2008; Jiang et al., 2008). In contrast, the pro-inflammatory molecules chemokine (C-X-C motif) ligand 10 (Cxcl10) and chemokine (C-C motif) ligand 4 (Ccl4) were downregulated in FAD DKO microglia. Likewise, cytochrome b-245 beta polypeptide (Cybb), a potent generator of reactive oxygen species (ROS), was reduced in FAD DKO mice, indicating an overall decrease of proinflammatory milieu. To investigate how Hdac1 and Hdac2 deletion altered histone acetylation in microglia, we performed ChIP-seq analysis for H3K9 and H3K27 acetylation. We found significant enrichment of genes involved in phagocytosis or related pathways in FAD DKO animals compared to FAD (Figures S7A and S7B). In accordance with the transcriptomic data, we also observed increased H3K27ac in the promoter of Apoe and decreased H3K9ac in Ccl4 in FAD DKO mice (Figure 6K). Additionally, two other genes, namely chemokine (C-C motif) receptor 3 (Ccr3) and Il33 known to be involved in microglial phagocytosis (Xia et al., 1998; Yasuoka et al., 2011), also showed enhanced promoter H3K9 and H3K27ac, respectively, in FAD DKO (Figure 6K). Taken together, our data indicate that in FAD DKO mice, microglia exhibit increased phagocytosis of amyloid plaques and a reduced inflammatory signature.

Figure 6. Absence of Hdac1 and Hdac2 Enhances Microglial Amyloid Phagocytosis (A) Representative confocal images showing co-localization of Iba1+ (red) microglia with Lamp2 (green) in the cortex and hippocampus of FAD and FAD DKO mice at 24–28 weeks. DAPI is in blue. Scale bar: 20 mm. (B and C) Quantification of the Lamp2+ area and Lamp2 fluorescence intensity from the images rendered from the cortex (B) and hippocampus (C) of FAD (blue circles) and FAD DKO (red circles) mice at 24–28 week. Bars represent mean ± SEM of 6–9 mice per genotype from two experiments. AU: arbitrary unit. Statistical analysis: unpaired t test, *p < 0.05; ns, not significant. (D) Representative confocal images of hippocampus from FAD and FAD DKO mice at 24–28 weeks showing co-localization of Iba1+ microglia with 6E10 immunoreactive amyloid plaques. Scale bar: 50 mm. Inset: zoomed image of microglia co-localized with plaque, scale bar: 50 mm. (E) Quantification of volume and intensity of 6E10 channel co-localized with Iba1 in the images mentioned in (D) (n = 4). Bars represent mean ± SEM from two experiments. Statistical analysis: unpaired t test, *p < 0.05, ***p < 0.001. (F and G) IMARIS-based 3D reconstruction of Iba1+ microglia and 6E10+ amyloid plaques from hippocampus of FAD and FAD DKO mice at 24–28 weeks. Orange box in confocal stacks (F) is zoomed in (G) showing reconstructed Iba1 (red) and amyloid plaque (green, left), only Iba1 (middle), and engulfed Ab plaque (right) from FAD and FAD DKO mice. Scale bar: 20 mm. (H) Quantification of volume of Iba1 channel co-localized with 6E10 in the reconstructed images in (G). Bars represent mean ± SEM of 4–5 mice per genotype from one experiment. Statistical analysis: unpaired t test, **p < 0.01; ns, not significant. (I) Quantification of Iba1+ microglia associated with 6E10+ plaques in the cortex of FAD (n = 4) and FAD DKO (n = 4) mice at 24–28 weeks. Bars represent mean ± SEM of one experiment. Statistical analysis: unpaired t test, *p < 0.05. (J) Heatmap showing expression (Log2FC) of genes involved in chemokine signaling and phagocytic clearance of debris in microglia from FAD and FAD DKO mice at 24–28 weeks obtained from microarray analysis (n = 3, one experiment). The fold change is normalized to wild-type expression values. (K) ChIP-seq data showing the coverage of H3K9 and H3K27Ac in the promoter region for the selected genes in FAD and FAD DKO mice at 24–28 weeks. The first exon (Ex1) for each gene is indicated below. RPKM values are shown in the top left for each panel. The averages of two replicates from one experiment are shown. See also Figure S7.

Immunity 48, 1–16, March 20, 2018 11

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

A

B

C

D

E

G

F

H

I

J

K

L

(legend on next page)

12 Immunity 48, 1–16, March 20, 2018

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

Hdac1 and Hdac2 Deficiency in Neuroectodermal Cells Does Not Affect Amyloid Pathology It is previously demonstrated that modulation of epigenetic marks in neurons has vigorous disease-modulating effects in mouse models of AD (Guan et al., 2009). To investigate whether the absence of Hdac1 and Hdac2 in non-microglial cells in the CNS had any effect on the course of neurodegeneration, we generated NestinCre Hdac1fl/flHdac2fl/wt animals on a 5xFAD background (Figure 7A). To avoid embryonic lethality seen in the complete absence of both Hdac1 and Hdac2 in neuroectodermal cells (Hagelkruys et al., 2014), one Hdac2 wild-type allele was retained. Resulting mice exhibited reduced levels of Hdac1 and Hdac2 in neuroectodermal cells (Figure 7B) but not in microglia (Figure 7C). Analysis of the amyloid plaque load in 5xFAD NestinCre Hdac1fl/flHdac2fl/wt animals at 24–28 and 44–48 weeks of age revealed no differences either in the cortex (Figures 7D and 7E) or the hippocampus (Figures 7D and 7F). Similarly, the amount of soluble and insoluble Ab species in the cortex, hippocampus, and cerebellum revealed no significant differences between the two genotypes (Figures 7G–7L). Our results imply that microglia, rather than neurons or other neuroectodermal cells, modulate b-amyloid deposition and consecutive cognitive decline in Hdac1- and Hdac2-dependent fashion. DISCUSSION Here, we described a time- and context-dependent requirement of Hdac1 and Hdac2 for defined in vivo functions of microglia, particularly their proper development and immune capabilities during neurodegeneration. Prenatal ablation of Hdac1 and Hdac2 caused spontaneous microglia impairment, including blockage of proliferation and enhanced apoptosis due to enriched H3K9ac and H3K27ac deposition on the respective gene promoters. Hdac1 and Hdac2 deficiency during neurodegeneration, however, enhanced microglial amyloid phagocytosis, thereby reducing the plaque load and preventing cognitive decline. The in vivo data obtained here were based on our recently developed constitutive and inducible Cx3cr1 promoter-driven Cre mouse line to induce gene deletion in longliving CNS cells such as microglia at time points of interest that

can include both developmental and disease-related aspects of microglia (Goldmann et al., 2013, 2016; Yona et al., 2013; Hagemeyer et al., 2016). We were thereby able to show that the same genetic perturbation, i.e., ablation of Hdac1 and Hdac2, has a differential impact on microglia in the adult and developing system, since microglia are in different activation states. A similar context-dependent requirement has been shown recently in microglia for microRNAs in vivo (Varol et al., 2017). In the mouse CNS, Hdac1 is particularly expressed in neuronal progenitors and glia while Hdac2 is more strongly expressed in post-mitotic neuroblasts and neurons (MacDonald and Roskams, 2008). Previous studies using cell type-specific conditional gene deletion systems demonstrated that single deletion of either Hdac1 or Hdac2 has only subtle effects due to their highly overlapping and redundant functions (Montgomery et al., 2007, 2009; Ye et al., 2009). Consequently, conditional deletion of either Hdac1 or Hdac2 from a subset of neuronal precursors and astrocytes using Gfap Cre line results in normal brain development while double deficiency leads to severe brain abnormalities and premature postnatal death of mice (Montgomery et al., 2009). Deletion of both Hdac1 and Hdac2 from all neuroectodermal cells using Nestin Cre mice causes embryonic lethality due to severe brain malformation, a phenotype that could be rescued by a single allele of Hdac2 but not Hdac1 (Hagelkruys et al., 2014). Furthermore, lack of either Hdac1 or Hdac2 specifically in oligodendrocytes shows no apparent phenotype while the double-deficient situation results in a complete loss of mature oligodendrocytes leading to tremor and death during the first postnatal weeks underlining the vital functions of Hdac1 and Hdac2 for accurate oligodendrocyte development (Ye et al., 2009). Lack of Hdac1 and Hdac2 in microglia caused severe changes of microglia cell expansion during prenatal development by regulating cell proliferation and apoptosis but did not impair the perinatal survival of mice. This finding reflects the extreme plasticity of this cell type that shares its highly adaptive nature with other hematopoietic cells. Indeed, T cell-restricted absence of Hdac1 and Hdac2 resulted in living animals that developed lymphomas at later stages (Dovey et al., 2013; Heideman et al., 2013). In adult microglia, these epigenetic enzymes are

Figure 7. Loss of Hdac1 and Hdac2 in Neuroectodermal Cells Does Not Affect AD Pathogenesis (A) Experimental timeline for the analysis of 5xFAD NestinCre Hdac1fl/flHdac2fl/wt mice. (B) Relative mRNA expression of Hdac1 and Hdac2 from whole brains of wild-type (WT, black circles, n = 3), 5xFAD Hdac1fl/flHdac2fl/wt (FAD Cre
, blue circles, n = 3), and 5xFAD NestinCre Hdac1fl/flHdac2fl/wt (FAD Cre+, green circles, n = 3) mice at 24–28 weeks. Data obtained from qRT-PCR of one experiment and expression are normalized with respect to WT. Bars represent mean ± SEM. Statistical analysis: one-way ANOVA, Bonferroni’s multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001. (C) Left: western blot for Hdac1 and Hdac2 in microglia from FAD Cre
(n = 4) and FAD Cre+ mice. GAPDH is loading control. One representative lane for each genotype is shown. Right: Densitometric analysis thereof. Bars represent mean ± SEM, one experiment. Statistical analysis: unpaired t test; ns, not significant. (D) Representative confocal images showing 6E10+ plaques (green) and Iba1+ microglia (red) in the cortex and hippocampus of FAD Cre
and FAD Cre+ mice at 24–28 weeks (top) and 44–48 weeks (bottom). DAPI is in blue. Scale bar: 40 mm. White arrow, plaque; white arrowhead, 6E10+ neuron. Inset: zoomed image of plaque, scale bar: 20 mm. (E and F) Histological quantification of amyloid plaques in the cortex (E) and hippocampus (F) of FAD Cre
and FAD Cre+ mice at the specified age. Bars represent mean ± SEM of 3–6 mice per genotype per time point from two experiments. Statistical analysis: two-way ANOVA, multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. (G and H) Quantification of soluble Ab42 (G) and 40 (H) species in the cortex, hippocampus, and cerebellum of FAD Cre
and FAD Cre+. Bars represent mean ± SEM of 3–5 mice per genotype per time point from two experiments. Statistical analysis: two-way ANOVA, multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. (I and J) Quantification of insoluble Ab42 in the cortex (I) and hippocampus (J) of FAD Cre
and FAD Cre+. Bars represent mean ± SEM of 4–5 mice per genotype per time point from two experiments. Statistical analysis: Two-way ANOVA, multiple comparison test, *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. (K and L) Similar quantification for insoluble Ab40 in the cortex (K) and hippocampus (L).

Immunity 48, 1–16, March 20, 2018 13

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

present but largely dispensable for the maintenance of adult microglia homeostasis. Specifically, Hdac1- and Hdac2-deficient microglia did not show any signs of spontaneous activation or apoptosis. Histone acetylation has been demonstrated to be involved in memory storage. It restructures chromatin in distinct brain regions that have been implicated in learning and memory, most €ff and Tsai, 2013). Additionprominently in the hippocampus (Gra ally, it is shown that the non-selective HDAC inhibitor sodium butyrate reinstates learning ability and promotes the retrieval of long-term memory in mice even after massive cell loss (Fischer et al., 2007). Furthermore, neuron-specific overexpression of Hdac2, but not Hdac1, decreases dendritic spine density, synapse number, and memory formation (Guan et al., 2009). However, the majority of these studies used global knockouts for individual Hdacs or artificial overexpression systems, or treated whole animals without any cell specificity. For example, complete Hdac2
/
mice show enhanced memory function and altered synapse formation and plasticity in the mouse hippocampus (Guan et al., 2009) but the fact that microglia also express Hdacs was not taken into account. Further, neuron-associated microglia are known to regulate synapse formation during homeostasis and disease and absence of microglia results in impaired memory formation as well (Parkhurst et al., 2013). Using novel cell type-restricted genetic models, we have found Hdac1 and Hdac2 as regulators of microglia activation, proliferation, and phagocytosis that define microglia features during the specific conditions of prenatal development and AD-associated neurodegeneration. In contrast, we found that deficiency of Hdac1 and Hdac2 in neuroectodermal cells was not able to modulate amyloid plaque formation. Not only do these results reveal distinct functions of Hdac1 and Hdac2 in the mouse brain, they also suggest a new strategy for therapeutic intervention of CNS diseases associated with memory impairment. In sum, our data highlight the dynamic change of the microglial epigenetic compartment during distinct microglia conditions such as development and pathologies that could be exploited therapeutically. STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: d d d

d

KEY RESOURCES TABLE CONTACT FOR REAGENT AND RESOURCE SHARING EXPERIMENTAL MODEL AND SUBJECT DETAILS B Mice B Time mating and embryo analysis B Primary cell culture METHOD DETAILS B Ex vivo isolation of microglia and flow cytometry B Immunohistochemistry and immunofluorescence B 3D reconstruction of microglia B Microglial proliferation assay B ELISA B 3D reconstruction of amyloid within microglia B Microarray study B RNA sequencing and qRT PCR

14 Immunity 48, 1–16, March 20, 2018

B

ChIP-sequencing Western blot B Morris water maze (MWM) test B T-maze test QUANTIFICATION AND STATISTICAL ANALYSIS B CNS cell quantification B Quantification of amyloid plaques and co-localization with microglia B Quantification of synapses B Statistical analysis DATA AND SOFTWARE AVAILABILITY B

d

d

SUPPLEMENTAL INFORMATION Supplemental Information includes seven figures and one table and can be found with this article online at https://doi.org/10.1016/j.immuni.2018.02.016. ACKNOWLEDGMENTS The authors thank K. Seidel, M. Oberle, and M. Ditter for excellent technical support; M. Follo and team at lighthouse fluorescence technologies core Facility, University Medical Center, Freiburg for cell sorting; CEMT, University of Freiburg for excellent animal care; and KFB, Center of Excellence for Fluorescent Bioanalytics, Regensburg for Microarray and RNA-Seq analysis. The authors also thank Prof. Dr. Hassan Jumaa from Institute for Immunology, Ulm University for his support. M.P. is supported by the BMBF-funded competence network of multiple sclerosis (KKNMS) (Fo¨rderkennzeichen 01GI1303B), the Sobek-Foundation, the DFG (SFB 992, SFB1160, SFB/TRR167, ReinhartKoselleck-Grant), and the Ministry of Science, Research and Arts, BadenWuerttemberg (Sonderlinie ‘‘Neuroinflammation’’). AUTHOR CONTRIBUTIONS M.D., O.S., and M.P. conceived the study. M.D., O.S., E.R., M.F., N.H., T.L.T., M.K., S.Z.-W., M.K., and D.M. performed experiments and/or analysis. T.B. helped in behavioral analysis. P.M. generated the Hdac1fl/fl and Hdac2fl/fl mice. M.M.-L. helped to design some experiments and provided scientific input. M.D. and M.P. wrote the manuscript. All authors contributed to the editing of the paper. DECLARATION OF INTERESTS The authors declare no competing interest. Received: July 18, 2017 Revised: December 21, 2017 Accepted: February 22, 2018 Published: March 13, 2018 REFERENCES Ajami, B., Bennett, J.L., Krieger, C., Tetzlaff, W., and Rossi, F.M.V. (2007). Local self-renewal can sustain CNS microglia maintenance and function throughout adult life. Nat. Neurosci. 10, 1538–1543. Amit, I., Winter, D.R., and Jung, S. (2016). The role of the local environment and epigenetics in shaping macrophage identity and their effect on tissue homeostasis. Nat. Immunol. 17, 18–25. Casano, A.M., and Peri, F. (2015). Microglia: multitasking specialists of the brain. Dev. Cell 32, 469–477. Colonna, M., and Butovsky, O. (2017). Microglia function in the central nervous system during health and neurodegeneration. Annu. Rev. Immunol. 35, 441–468. Das Gupta, K., Shakespear, M.R., Iyer, A., Fairlie, D.P., and Sweet, M.J. (2016). Histone deacetylases in monocyte/macrophage development, activation and metabolism: refining HDAC targets for inflammatory and infectious diseases. Clin. Transl. Immunology 5, e62.

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut, P., Chaisson, M., and Gingeras, T.R. (2013). STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21.

Jiang, Q., Lee, C.Y., Mandrekar, S., Wilkinson, B., Cramer, P., Zelcer, N., Mann, K., Lamb, B., Willson, T.M., Collins, J.L., et al. (2008). ApoE promotes the proteolytic degradation of Abeta. Neuron 58, 681–693.

Dovey, O.M., Foster, C.T., Conte, N., Edwards, S.A., Edwards, J.M., Singh, R., Vassiliou, G., Bradley, A., and Cowley, S.M. (2013). Histone deacetylase 1 and 2 are essential for normal T-cell development and genomic stability in mice. Blood 121, 1335–1344.

Keren-Shaul, H., Spinrad, A., Weiner, A., Matcovitch-Natan, O., DvirSzternfeld, R., Ulland, T.K., David, E., Baruch, K., Lara-Astaiso, D., Toth, B., et al. (2017). A unique microglia type associated with restricting development of Alzheimer’s disease. Cell 169, 1276–1290.e17.

Elmore, M.R.P., Najafi, A.R., Koike, M.A., Dagher, N.N., Spangenberg, E.E., Rice, R.A., Kitazawa, M., Matusow, B., Nguyen, H., West, B.L., et al. (2014). Colony-stimulating factor 1 receptor signaling is necessary for microglia viability, unmasking a microglia progenitor cell in the adult brain. Neuron 82, 380–397.

Kierdorf, K., and Prinz, M. (2017). Microglia in steady state. J. Clin. Invest. 127, 3201–3209.

 de Angelis, A.L., Jaitin, D., Wieghofer, P., Staszewski, O., Erny, D., Hrabe David, E., Keren-Shaul, H., Mahlakoiv, T., Jakobshagen, K., Buch, T., et al. (2015). Host microbiota constantly control maturation and function of microglia in the CNS. Nat. Neurosci. 18, 965–977. Fischer, A., Sananbenesi, F., Wang, X., Dobbin, M., and Tsai, L.H. (2007). Recovery of learning and memory is associated with chromatin remodelling. Nature 447, 178–182. €ller, P.F., Wolf, Y., Varol, D., Yona, S., Goldmann, T., Wieghofer, P., Mu Brendecke, S.M., Kierdorf, K., Staszewski, O., Datta, M., et al. (2013). A new type of microglia gene targeting shows TAK1 to be pivotal in CNS autoimmune inflammation. Nat. Neurosci. 16, 1618–1626. Goldmann, T., Zeller, N., Raasch, J., Kierdorf, K., Frenzel, K., Ketscher, L., Basters, A., Staszewski, O., Brendecke, S.M., Spiess, A., et al. (2015). USP18 lack in microglia causes destructive interferonopathy of the mouse brain. EMBO J. 34, 1612–1629. Goldmann, T., Jorda˜o, M.J.C., Wieghofer, P., Prutek, F., Hagemeyer, N., Frenzel, K., Staszewski, O., Kierdorf, K., Amann, L., Krueger, M., et al. (2016). Origin, fate and dynamics of macrophages at central nervous system interfaces. Nat. Immunol. 17, 797–805. €ff, J., and Tsai, L.H. (2013). Histone acetylation: molecular mnemonics on Gra the chromatin. Nat. Rev. Neurosci. 14, 97–111. Grommes, C., Lee, C.Y.D., Wilkinson, B.L., Jiang, Q., Koenigsknecht-Talboo, J.L., Varnum, B., and Landreth, G.E. (2008). Regulation of microglial phagocytosis and inflammatory gene expression by Gas6 acting on the Axl/Mer family of tyrosine kinases. J. Neuroimmune Pharmacol. 3, 130–140. Guan, J.-S., Haggarty, S.J., Giacometti, E., Dannenberg, J.-H., Joseph, N., Gao, J., Nieland, T.J.F., Zhou, Y., Wang, X., Mazitschek, R., et al. (2009). HDAC2 negatively regulates memory formation and synaptic plasticity. Nature 459, 55–60. Guerreiro, R., Wojtas, A., Bras, J., Carrasquillo, M., Rogaeva, E., Majounie, E., Cruchaga, C., Sassi, C., Kauwe, J.S., Younkin, S., et al.; Alzheimer Genetic Analysis Group (2013). TREM2 variants in Alzheimer’s disease. N. Engl. J. Med. 368, 117–127. Hagelkruys, A., Lagger, S., Krahmer, J., Leopoldi, A., Artaker, M., Pusch, O., Zezula, J., Weissmann, S., Xie, Y., Scho¨fer, C., et al. (2014). A single allele of Hdac2 but not Hdac1 is sufficient for normal mouse brain development in the absence of its paralog. Development 141, 604–616. Hagemeyer, N., Kierdorf, K., Frenzel, K., Xue, J., Ringelhan, M., Abdullah, Z., Godin, I., Wieghofer, P., Costa Jorda˜o, M.J., Ulas, T., et al. (2016). Transcriptome-based profiling of yolk sac-derived macrophages reveals a role for Irf8 in macrophage maturation. EMBO J. 35, 1730–1744. Heideman, M.R., Wilting, R.H., Yanover, E., Velds, A., de Jong, J., Kerkhoven, R.M., Jacobs, H., Wessels, L.F., and Dannenberg, J.-H. (2013). Dosagedependent tumor suppression by histone deacetylases 1 and 2 through regulation of c-Myc collaborating genes and p53 function. Blood 121, 2038–2050. Hollingworth, P., Harold, D., Sims, R., Gerrish, A., Lambert, J.-C., Carrasquillo, M.M., Abraham, R., Hamshere, M.L., Pahwa, J.S., Moskvina, V., et al.; Alzheimer’s Disease Neuroimaging Initiative; CHARGE consortium; EADI1 consortium (2011). Common variants at ABCA7, MS4A6A/MS4A4E, EPHA1, CD33 and CD2AP are associated with Alzheimer’s disease. Nat. Genet. 43, 429–435.

Kierdorf, K., Erny, D., Goldmann, T., Sander, V., Schulz, C., Perdiguero, E.G., Wieghofer, P., Heinrich, A., Riemke, P., Ho¨lscher, C., et al. (2013). Microglia emerge from erythromyeloid precursors via Pu.1- and Irf8-dependent pathways. Nat. Neurosci. 16, 273–280. Lagger, G., O’Carroll, D., Rembold, M., Khier, H., Tischler, J., Weitzer, G., Schuettengruber, B., Hauser, C., Brunmeir, R., Jenuwein, T., and Seiser, C. (2002). Essential function of histone deacetylase 1 in proliferation control and CDK inhibitor repression. EMBO J. 21, 2672–2681. Lavin, Y., Winter, D., Blecher-Gonen, R., David, E., Keren-Shaul, H., Merad, M., Jung, S., and Amit, I. (2014). Tissue-resident macrophage enhancer landscapes are shaped by the local microenvironment. Cell 159, 1312–1326. Law, C.W., Chen, Y., Shi, W., and Smyth, G.K. (2014). voom: Precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol. 15, R29. Liao, Y., Smyth, G.K., and Shi, W. (2014). featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930. MacDonald, J.L., and Roskams, A.J. (2008). Histone deacetylases 1 and 2 are expressed at distinct stages of neuro-glial development. Dev. Dyn. 237, 2256–2267. Meyer-Luehmann, M., and Prinz, M. (2015). Myeloid cells in Alzheimer’s disease: culprits, victims or innocent bystanders? Trends Neurosci. 38, 659–668. Mildner, A., Schmidt, H., Nitsche, M., Merkler, D., Hanisch, U.-K., Mack, M., €ck, W., Priller, J., and Prinz, M. (2007). Microglia in the Heikenwalder, M., Bru adult brain arise from Ly-6ChiCCR2+ monocytes only under defined host conditions. Nat. Neurosci. 10, 1544–1553. Montgomery, R.L., Davis, C.A., Potthoff, M.J., Haberland, M., Fielitz, J., Qi, X., Hill, J.A., Richardson, J.A., and Olson, E.N. (2007). Histone deacetylases 1 and 2 redundantly regulate cardiac morphogenesis, growth, and contractility. Genes Dev. 21, 1790–1802. Montgomery, R.L., Hsieh, J., Barbosa, A.C., Richardson, J.A., and Olson, E.N. (2009). Histone deacetylases 1 and 2 control the progression of neural precursors to neurons during brain development. Proc. Natl. Acad. Sci. USA 106, 7876–7881. Oakley, H., Cole, S.L., Logan, S., Maus, E., Shao, P., Craft, J., GuillozetBongaarts, A., Ohno, M., Disterhoft, J., Van Eldik, L., et al. (2006). Intraneuronal b-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: potential factors in amyloid plaque formation. J. Neurosci. 26, 10129–10140. Paolicelli, R.C., Jawaid, A., Henstridge, C.M., Valeri, A., Merlini, M., Robinson, J.L., Lee, E.B., Rose, J., Appel, S., Lee, V.M.Y., et al. (2017). TDP-43 depletion in microglia promotes amyloid clearance but also induces synapse loss. Neuron 95, 297–308.e6. Parkhurst, C.N., Yang, G., Ninan, I., Savas, J.N., Yates, J.R., 3rd, Lafaille, J.J., Hempstead, B.L., Littman, D.R., and Gan, W.-B. (2013). Microglia promote learning-dependent synapse formation through brain-derived neurotrophic factor. Cell 155, 1596–1609. Prinz, M., and Priller, J. (2014). Microglia and brain macrophages in the molecular age: from origin to neuropsychiatric disease. Nat. Rev. Neurosci. 15, 300–312. Prinz, M., and Priller, J. (2017). The role of peripheral immune cells in the CNS in steady state and disease. Nat. Neurosci. 20, 136–144. Rademakers, R., Baker, M., Nicholson, A.M., Rutherford, N.J., Finch, N., Soto-Ortolaza, A., Lash, J., Wider, C., Wojtas, A., DeJesus-Hernandez, M., et al. (2011). Mutations in the colony stimulating factor 1 receptor (CSF1R)

Immunity 48, 1–16, March 20, 2018 15

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

gene cause hereditary diffuse leukoencephalopathy with spheroids. Nat. Genet. 44, 200–205. €ndar, F., Diehl, S., Gru €ning, B.A., and Manke, T. (2014). Ramı´rez, F., Du deepTools: a flexible platform for exploring deep-sequencing data. Nucleic Acids Res. 42, W187–W191. Ransohoff, R.M. (2016). How neuroinflammation contributes to neurodegeneration. Science 353, 777–783. Ritchie, M.E., Phipson, B., Wu, D., Hu, Y., Law, C.W., Shi, W., and Smyth, G.K. (2015). limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47–e47. Robinson, J.T., Thorvaldsdo´ttir, H., Winckler, W., Guttman, M., Lander, E.S., Getz, G., and Mesirov, J.P. (2011). Integrative genomics viewer. Nat. Biotechnol. 29, 24–26. Sellner, S., Paricio-Montesinos, R., Spieß, A., Masuch, A., Erny, D., Harsan, L.A., Elverfeldt, D.V., Schwabenland, M., Biber, K., Staszewski, O., et al. (2016). Microglial CX3CR1 promotes adult neurogenesis by inhibiting Sirt 1/p65 signaling independent of CX3CL1. Acta Neuropathol. Commun. 4, 102. Shen, L., Shao, N.-Y., Liu, X., Maze, I., Feng, J., and Nestler, E.J. (2013). diffReps: detecting differential chromatin modification sites from ChIP-seq data with biological replicates. PLoS ONE 8, e65598. Spowart-Manning, L., and van der Staay, F.J. (2004). The T-maze continuous alternation task for assessing the effects of putative cognition enhancers in the mouse. Behav. Brain Res. 151, 37–46. Tay, T.L., Mai, D., Dautzenberg, J., Ferna´ndez-Klett, F., Lin, G., Sagar, Datta, M., Drougard, A., Stempfl, T., Ardura-Fabregat, A., et al. (2017). A new fate mapping system reveals context-dependent random or clonal expansion of microglia. Nat. Neurosci. 20, 793–803.

16 Immunity 48, 1–16, March 20, 2018

Varol, D., Mildner, A., Blank, T., Shemer, A., Barashi, N., Yona, S., David, E., Boura-Halfon, S., Segal-Hayoun, Y., Chappell-Maor, L., et al. (2017). Dicer deficiency differentially impacts microglia of the developing and adult brain. Immunity 46, 1030–1044.e8. Wilting, R.H., Yanover, E., Heideman, M.R., Jacobs, H., Horner, J., van der Torre, J., DePinho, R.A., and Dannenberg, J.-H. (2010). Overlapping functions of Hdac1 and Hdac2 in cell cycle regulation and haematopoiesis. EMBO J. 29, 2586–2597. Xia, M.Q., Qin, S.X., Wu, L.J., Mackay, C.R., and Hyman, B.T. (1998). Immunohistochemical study of the b-chemokine receptors CCR3 and CCR5 and their ligands in normal and Alzheimer’s disease brains. Am. J. Pathol. 153, 31–37. Yasuoka, S., Kawanokuchi, J., Parajuli, B., Jin, S., Doi, Y., Noda, M., Sonobe, Y., Takeuchi, H., Mizuno, T., and Suzumura, A. (2011). Production and functions of IL-33 in the central nervous system. Brain Res. 1385, 8–17. Ye, F., Chen, Y., Hoang, T., Montgomery, R.L., Zhao, X.H., Bu, H., Hu, T., Taketo, M.M., van Es, J.H., Clevers, H., et al. (2009). HDAC1 and HDAC2 regulate oligodendrocyte differentiation by disrupting the b-catenin-TCF interaction. Nat. Neurosci. 12, 829–838. Yona, S., Kim, K.-W., Wolf, Y., Mildner, A., Varol, D., Breker, M., Strauss-Ayali, D., Viukov, S., Guilliams, M., Misharin, A., et al. (2013). Fate mapping reveals origins and dynamics of monocytes and tissue macrophages under homeostasis. Immunity 38, 79–91. Zhao, S., Guo, Y., Sheng, Q., and Shyr, Y. (2014). Heatmap3: an improved heatmap package with more powerful and convenient features. BMC Bioinformatics 15, P16.

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

STAR+METHODS KEY RESOURCES TABLE

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Antibodies Mouse anti-beta-Amyloid antibody (clone 6F/3D)

DAKO

Cat# M0872 RRID:AB_2056966

Mouse anti-APP antibody (clone 22C11)

Millipore

Cat# MAB348 RRID:AB_94882

Rabbit anti-Iba1 antibody

Wako

Cat# 019-19741 RRID:AB_839504

Mouse anti-NeuN antibody (clone A60)

Millipore

Cat# MAB377 RRID:AB_2298772

Rat anti-CD31 antibody

BD PharMingen

Cat# 550274 RRID:AB_393571

Rabbit anti-GFAP antibody

DAKO

Cat# IR524

Goat anti-Olig2 antibody

R&D Systems

Cat# AF2418 RRID:AB_2157554

Mouse anti-Hdac1 antibody (clone 4C4G7)

Proteintech

Cat# 66085 RRID:AB_11232033

Mouse anti-Hdac1 antibody (clone 10E2)

Santa Cruz

Cat# sc-81598 RRID:AB_2118083

Mouse anti-Hdac2 antibody (clone HDAC2-62)

Abcam

Cat# ab12169 RRID:AB_2118547

Sheep anti-serum albumin antibody

Abcam

Cat# ab8940 RRID:AB_306875

Mouse anti-b-Amyloid, 1-16 Antibody (Clone 6E10)

Covance

Cat# SIG-39320, RRID: AB_2564653

Rat anti-Lamp2 antibody

Abcam

Cat# ab13524 RRID:AB_2134736

Mouse Anti-GAPDH antibody (clone 6C5)

Millipore

Cat# MAB374 RRID:AB_2107445

APC conjugated anti mouse CD11b (clone M1/70)

BD PharMingen

Cat# 553312 RRID:AB_398535

eFlour 450 conjugated anti mouse CD45 (Clone 30-F11)

eBioscience

Cat# 48-0451-82 RRID:AB_1518806

Anti-Histone H3 (acetyl K9) antibody

Abcam

Cat# ab10812 RRID:AB_297491

Anti-Histone H3 (acetyl K27) antibody

Abcam

Cat# ab4729 RRID:AB_2118291

Rabbit anti-Homer1 antibody

Millipore

Cat# ABN37 RRID:AB_11214387

Tamoxifen

Sigma-Aldrich

Cat# T5648

Progesterone

Sigma-Aldrich

Cat# P0130

Human Aß42 ultrasensitive ELISA kit

Life Technologies

Cat# KHB3544

Human Aß40 ultrasensitive ELISA kit

Life Technologies

Cat# KHB3481

Click-iT EdU Flow Cytometry Assay Kit

Invitrogen

Cat# C10425

Click-iT EdU Imaging Kit

Invitrogen

Cat# C10340

Picopure RNA extraction kit

Life Technologies

Cat# KIT0204

SMART-Seq v4 Ultra Low Input RNA Kit for Sequencing

Clontech Laboratories

Cat# 634888

Nextera XT DNA Library Preparation Kit

Illumina

Cat# FC-131-1024

KAPA SYBR FAST One-Step

Kapa Biosystems

Cat# KK4835

TruSeq SR Cluster Kit v3

Illumina

Cat# GD-401-3001

TruSeq SBS v3

Illumina

Cat# FC-401-3001

Chemicals, Peptides, and Recombinant Proteins

Critical Commercial Assays

True MicroChIP Kit

Diagenode

Cat# C01010130

KAPA Library Preparation Kit with Real-time PCR Library Amplification

Kapa Biosystems

Cat# KK8221

High Capacity RNA to cDNA kit

Life Technologies

Cat# 4387406

RNA 6000 Pico Kit

Agilent

Cat# 5067-1513

Ovation Pico WTA System V2

Nugen

Cat# 3302-12

Encore Biotin Module

Nugen

Cat# 4200-12

Affymetrix Mouse Gene 2.0 ST

Thermo Fisher Scientific

Cat# 902119

This Paper

GEO (GSE73125, GSE98823)

Deposited Data Raw and analyzed data

(Continued on next page)

Immunity 48, 1–16.e1–e6, March 20, 2018 e1

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

Continued REAGENT or RESOURCE

SOURCE

IDENTIFIER

The Jackson Laboratory

MMRRC_034848-JAX

Mouse: B6.Cg-Hdac1tm1Pmt/J

Wilting et al., 2010

N/A

Mouse: B6.Cg-Hdac2tm1Pmt/J

Wilting et al., 2010

N/A

Experimental Models: Organisms/Strains Mouse: 5x FAD (B6SJLTg(APPSwFlLon,PSEN1*M146L*L286V)

Mouse: B6J.B6N(Cg)-Cx3cr1tm1.1(cre)Jung/J

Yona et al., 2013

N/A

Mouse: B6.129P2(C)-Cx3cr1tm2.1(cre/ERT2)Jung/J

Yona et al., 2013

N/A

Mouse: B6.Cg-Tg(Nes-cre)1Kln/J

The Jackson Laboratory

Stock# 003771

Mouse: crosses of the above

In this paper

N/A

Table S1

N/A

IMARIS

Bitplane

http://www.bitplane.com/imaris/imaris

CASAVA

Illumina

https://support.illumina.com/

STAR

Dobin et al., 2013

https://github.com/alexdobin/STAR

featureCounts

Liao et al., 2014

http://subread.sourceforge.net/

limma

Law et al., 2014

http://bioconductor.org/packages/release/ bioc/html/limma.html

R

The R Project for Statistical Computing

https://www.r-project.org/

Ingenuity Pathway Analysis

QIAGEN

https://analysis.ingenuity.com/

Picard tools

Broad Institute

http://broadinstitute.github.io/picard

diffReps

Shen et al., 2013

https://github.com/shenlab-sinai/diffreps

deepTools

Ramı´rez et al., 2014

https://github.com/deeptools/deepTools

IGV

Robinson et al., 2011

http://software.broadinstitute.org/software/igv/

Affymetrix GeneChip Command Console software

Thermo Fisher Scientific

https://www.thermofisher.com/de/de/home/ life-science/microarray-analysis/microarrayanalysis-instruments-software-services/ microarray-analysis-software/affymetrixgenechip-command-console-software.html

BRB-ArrayTools

Dr. Richard Simon and the BRB-ArrayTools Development Team

https://brb.nci.nih.gov/BRB-ArrayTools/

Definiens Developer XD, Version 2.7.0.

Definiens Inc.

http://www.definiens.com/

Oligonucleotides Primers Software and Algorithms

CONTACT FOR REAGENT AND RESOURCE SHARING Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Marco Prinz ([email protected]). EXPERIMENTAL MODEL AND SUBJECT DETAILS Mice All animal experiments were approved by the Federal Ministry for Nature, Environment and Consumers’ Protection of the state of €rttemberg and were performed in accordance with the respective national, federal and institutional regulations. Generation Baden-Wu of Hdac1fl/fl, Hdac2fl/fl and 5xFAD mice have been described elsewhere (Oakley et al., 2006; Wilting et al., 2010). To generate microglia specific Hdac1 and Hdac2 deleted animals we intercrossed Hdac1fl/fl and Hdac2fl/fl animals with either Cx3cr1Cre (constitutive deletion) or Cx3cr1CreERT2 (inducible deletion) line. To obtain neuro-ectodermal cell specific deletion of Hdac1 and Hdac2, we crossed Hdac1fl/fl and Hdac2fl/wt animals with NestinCre animals (Hagelkruys et al., 2014). The Cx3cr1CreERT2 Hdac1fl/fl Hdac2fl/fl and NestinCre Hdac1fl/flHdac2fl/wt animals were further crossed with 5xFAD line to generate microglia and neuro-ectodermal cell specific Hdac1 and Hdac2 deleted FAD animals respectively. All mice were bred in specific-pathogen-free facility and were group housed up to five per cage with 12 h light/dark cycle. Food and water were available ad libitum. For the inducible deletion e2 Immunity 48, 1–16.e1–e6, March 20, 2018

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

line, 8 mg of Tamoxifen (TAM, Sigma-Aldrich) dissolved in corn oil was applied subcutaneously to 6 week old mice for 2 days (day0 and 2). Only female mice were used in all experiments. All animal experiments were approved by local administration and were performed in accordance to the respective national, federal and institutional regulations. Time mating and embryo analysis Time matings were started in the afternoon and the animals were separated on the following morning after plug check. Embryonic development was estimated considering the day of vaginal plug formation as E0.5. Plug positive Cx3cr1CreERT2 Hdac1fl/fl Hdac2fl/fl females were injected with 3 mg TAM (20 mg/ml TAM + 10 mg/ml Progesterone; Sigma-Aldrich) i.p.at E13. Primary cell culture Primary cultures were prepared as described previously (Mildner et al., 2007). Briefly, cortical hemispheres from E16 embryos (for neuronal culture) were freed from meninges and homogenized in phosphate buffer saline (PBS) with gentle trituration followed by centrifugation at 400 g for 10 min. The cells were then plated with Eagle’s basal medium (BME, GIBCO) supplemented with 1X B27 (GIBCO) and 1% Fetal bovine serum (FBS, GIBCO) and kept at 37 C incubator with 5% CO2. Cells were kept in culture for 9-14 days. For microglia, astrocytes and oligodendrocytes, cortical hemispheres from newborn mice (P1-3) were similarly freed from meninges, homogenized and plated in 25cm2 tissue culture flasks with Dulbecco modified eagle medium (DMEM, GIBCO) supplemented with 10% FBS. Cells were kept for 9-13 days with change of medium every alternate day until the base of the flask was confluent. Next, for isolation of oligodendrocytes, fresh DMEM was added to the flask and the flask was stricken against the palm of the hand 10-15 times to dislodge loosely adhered oligodendrocyte precursor cell. The supernatant was then centrifuged and the pellet was plated with oligodendrocyte differentiating medium (DMEM containing 1% N1 supplement, 1% BSA, 0.05mg/ml Insulin, 0.01% T3 hormone and 7.5 mg/ml D(+) Galactose, all supplements from Sigma Aldrich) and kept for 1-2 days. The mother flasks containing other cell types was then treated with MCSF (30ng/ml final concentration, R&D systems) to differentiate microglia precursors. Two days later, the flasks were shaken for 2-3 hr at 130 rpm at 37 C. The supernatant was centrifuged and plated with DMEM and 10% FBS for microglia. To obtain astrocytes, the mother flasks were further treated with chlodronate (20 mg/ml final concentration) for 2-3 days to kill all microglia precursors. Then the cells from flasks were trypsinized and plated again with DMEM and 10%FBS to astrocytes. METHOD DETAILS Ex vivo isolation of microglia and flow cytometry Mice were anesthetized and transcardially perfused with ice cold PBS. Brain hemispheres were homogenized and subjected to 37% / 70% percoll density gradient separation. Microglia cells were collected from the interface of 37% / 70% percoll and washed extensively with FACS buffer (0.5% FBS in PBS) followed by staining with anti-CD11b (clone M1/70, eBioscience) and anti-CD45 (clone 30-F11, eBioscience) antibodies. Cells were sorted by FACSAriaIII (Becton Dickinson) and used for further analysis. For embryonic and P0 microglia isolation, E16.5 embryos or P0 pups were killed by decapitation followed by brain homogenization and staining. Immunohistochemistry and immunofluorescence Brain hemispheres isolated from perfused mice were post-fixed with 4% paraformaldehyde (PFA). Fixed tissue was then used for either paraffin or cryo embedding. Four mm thick paraffin sections were used to stain with anti-beta-Amyloid (DAKO; 1:200) and anti APP antibody (Millipore; 1:2000) to assess plaque load. For immunofluorescence studies, 12 (cell counting) or 30 (morphometric study) mm thick cryosections were used. Sagittal brain slices were permeabilized (0.5% Triton X-100, 5% BSA in PBS) for 30 min, blocked (0.1% Triton X-100, 5% BSA in PBS) for 1 h at room temperature and incubated overnight at 4 C with the relevant primary antibodies diluted in 0.1% Triton X-100, 5% BSA in PBS: 1:500 rabbit anti-Iba1 (Wako), 1:100 mouse anti-NeuN (Millipore), 1:300 rat anti-CD31, 1:5000 rabbit anti-GFAP (Dako), 1:40 goat anti-Olig2 (R&D Systems), anti 1:100 mouse anti-Hdac1 (Abcam), 1:100 mouse anti-Hdac2 (Abcam), 1:1000 sheep serum albumin (Abcam), 1:1000 6E10 (Covance), 1:1000 anti Homer1 (Millipore) and 1:200 rat anti-Lamp2 (Abcam). Corresponding secondary antibodies conjugated to Alexa Fluor 488, 555 or 647 (1:500, Life Technologies) were used. Nuclei were counterstained with DAPI (1:5000, Sigma). Slices were mounted with ProLong Diamond Antifade Mountant (Life Technologies) or Mowiol (Sigma). Bright field and fluorescence images were acquired using Olympus BX-61 microscope and cellSens Dimension 1.14 software (Olympus). Fluorescence images for microglial morphometric analysis were acquired using Olympus confocal laser scanning microscope (FluoView 1000MPE). 3D reconstruction of microglia 3D reconstruction of microglia was performed as described previously (Goldmann et al., 2013). Briefly, free-floating 30-mm sagittal brain sections (cryo) were stained overnight with anti-Iba-1 antibody (1:500) at 4 C followed by Alexa Fluor 555–conjugated secondary antibody at a dilution of 1:500 for overnight at 4 C. Nuclei were counterstained with DAPI. Imaging was performed on an Olympus Fluoview 1000 confocal laser scanning microscope using a 20 3 0.95 NA objective. Z stacks with 1.14-mm steps in the z direction, 1024 3 1024 pixel resolution, were recorded and analyzed using Imaris software (Bitplane).

Immunity 48, 1–16.e1–e6, March 20, 2018 e3

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

Microglial proliferation assay Two week old pups were injected with 200 mg EdU in 100 ml PBS i.p every alternate day for 10 days (5 pulses). The animals were killed at 6 weeks and microglia cells were isolated from one brain hemisphere. EdU incorporation in CD11bhi CD45lo microglia was determined by FACS using Click-iT EdU Flow Cytometry Assay Kit (Invitrogen, catalog no.C10425) according to manufacturer’s protocol. The other brain hemisphere was post fixed in 4% PFA and cryoembedded. 12mm thick sagittal sections were then stained with Click-iT EdU Imaging Kit (Invitrogen, catalog no. C10340) according to manufacturer’s protocol followed by staining with anti Iba1 antibody and DAPI. Iba1+ EdU+ cells were quantified from 5-8 different regions in the cortex and 3 brain slices per animal. ELISA For the quantification of Abeta 40 and 42 species in the soluble and insoluble brain extracts, cortical, hippocampal and cerebellar tissue was homogenized and sequentially extracted with PBS (soluble fraction), PBS + 0.1% Triton X-100 (membrane bound fraction) and finally with 8M guanidine hydrochloride solution. Protein concentration in each fraction was measured with Bradford reagent (Roth) and ELISA was performed using Human Ab42 ultrasensitive ELISA kit (Life Technologies, catalog no. KHB3544) and Human Ab40 ELISA kit (Life Technologies, catalog no. KHB3481) according to manufacturer’s protocol. 3D reconstruction of amyloid within microglia Quantification of amyloid within plaque associated microglia was performed as recently published (Paolicelli et al., 2017). Two micrometer thick confocal z stacks were rendered using Olympus Fluoview 1000 confocal laser scanning microscope with a 60 x objective. 3D surfaces were reconstructed using the IMARIS surface reconstruction tool. Similarly 6E10 immunoreactive plaque surfaces were reconstructed. Filtering for 6E10 signal contained wholly within the Iba1 surface yielded the volume of 6E10 immunoreactive amyloid contained within the microglia. Microarray study Total RNA was extracted from FACS sorted microglia cells using Picopure RNA extraction kit (Life Technologies) according to manufacturer’s protocol. Isolated RNA was controlled for quantity and determination of a RNA integrity score (RIN) using RNA pico chips on a Bioanalyzer 2100 (Agilent). Sample preparation for microarray hybridization was carried out as described in the NuGEN Ovation Pico WTA System V2 and NUGEN Encore Biotin Module manuals (NuGEN Technologies, Inc, San Carlos, CA, USA). In brief, between 0.4 and 2.5 ng of total RNA was reverse transcribed into double-stranded cDNA in a two-step process, introducing a SPIA tag sequence. Bead purified cDNA was amplified by a SPIA amplification reaction followed by an additional bead purification. 3.0 mg of SPIA cDNA was fragmented, terminally biotin-labeled and hybridized to an Affymetrix Mouse Gene 2.0 ST Array Plate. For hybridization, washing, staining and scanning an Affymetrix GeneTitan system, controlled by the Affymetrix GeneChip Command Console software v4.2, was used. Sample processing was performed at an Affymetrix Service Provider and Core Facility, ‘‘KFB - Center of Excellence for Fluorescent Bioanalytics’’ (Regensburg, Germany; www.kfb-regensburg.de). Differential Gene Expression Analyses were performed using BRB-ArrayTools developed by Dr. Richard Simon and the BRB-ArrayTools Development Team. Pathway Analysis was performed using Ingenuity Pathway Analysis (IPA, QIAGEN). RNA sequencing and qRT PCR Total RNA was extracted from FACS sorted microglia cells using Picopure RNA extraction kit (Life Technologies) according to manufacturer’s protocol. The SMARTer Ultra Low Input RNA Kit for Sequencing v4 (Clontech Laboratories, Inc., Mountain View, CA, USA) was used to generate first strand cDNA from 500 to 750 pg total-RNA. Double stranded cDNA was amplified by LD PCR (11 cycles) and purified via magnetic bead clean-up. Library preparation was carried out as described in the Illumina Nextera XT Sample Preparation Guide (Illumina, Inc., San Diego, CA, USA). 150 pg of input cDNA were tagmented (tagged and fragmented) by the Nextera XT transposome. The products were purified and amplified via a limited-cycle PCR program to generate multiplexed sequencing libraries. For the PCR step 1:5 dilutions of index 1 (i7) and index 2 (i5) primers were used. The libraries were quantified using the KAPA SYBR FAST ABI Prism Library Quantification Kit (Kapa Biosystems, Inc., Woburn, MA, USA). Equimolar amounts of each library were pooled, and the pools were used for cluster generation on the cBot with the Illumina TruSeq SR Cluster Kit v3. The sequencing run was performed on an HiSeq 1000 instrument using the indexed, 50 cycles single-read (SR) protocol and the TruSeq SBS v3 Reagents according to the Illumina HiSeq 1000 System User Guide. Image analysis and base calling resulted in .bcl files, which were converted into .fastq files with the CASAVA1.8.2 software. Library preparation and RNaseq were performed at the Genomics Core Facility ‘‘KFB - Center of Excellence for Fluorescent Bioanalytics’’ (University of Regensburg, Regensburg, Germany; www.kfb-regensburg.de). Fastq Files were quality controlled using FastQC and reads were mapped to the GRCm38 mouse genome using the Star aligner (Dobin et al., 2013). Read counts were obtained by the featureCounts package and differential gene expression analysis was performed using the limma/voom pipeline in R (Ritchie et al., 2015). Heatmaps were generated using the R library heatmap.3 (Zhao et al., 2014). Pathway analysis was performed using Ingenuity Pathway Analysis (IPA, QIAGEN). For gene expression analysis by Real time PCR, isolated RNA was reversed transcribed to cDNA using High Capacity RNA to cDNA kit (Life Technologies) and gene expression was analyzed using gene specific primers (Table S1).

e4 Immunity 48, 1–16.e1–e6, March 20, 2018

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

ChIP-sequencing Microglia were isolated as described above, however prior to antibody labeling for FACS sorting, cells were fixed in 1% formaldehyde in PBS for 5 minutes at room temperature, followed by incubation with 125 mM Tris buffer at pH 7.5 for 10 minutes at room temperature. FACS-sorted cells were resuspended in cell lysis buffer (10mM Tris pH 8.0, 10 mM NaCl, 0.2% NP40, 1x complete protease inhibitor (Sigma Aldrich)), nuclei were pelleted through centrifugation and resuspended in nuclei lysis buffer (50mM Tris pH 8.1, 10 mM EDTA, 1% SDS, 1x complete protease inhibitor). Chromatin was sheared using a Covaris M220 System. Further chromatin IP was using the True MicroChIP Kit (Diagenode) as per the manufacturer’s protocol. Sequencing library preparation was performed using the KAPA Library Preparation Kit with Real-time PCR Library Amplification for Illumina Platforms Kit (Kapa Biosystems) as per the manufacturer’s protocol. The Sequencing runs were performed on an Illumina HiSeq 2000 System as 50 bp paired end runs. Fastq files were quality controlled using FastQC and mapped to the GRCm38 mouse genome using the Star aligner. Duplicate reads were removed using Picard tools (http://broadinstitute.github.io/picard). Peak calling for replicate experiments was subsequently carried out using diffReps (Shen et al., 2013). Visualization was performed on Bigwig files generated using the DeepTools BAM Coverage script (Ramı´rez et al., 2014) and the Integrative Genome Viewer (IGV) (Robinson et al., 2011). Western blot Total protein from brain tissue or microglia cells were extracted in RIPA buffer (25 mM Tris-HCl, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, pH 7.5). Samples were separated by SDS-PAGE and immunoblotted using antibodies against Hdac1 (1:3000, Santa Cruz), Hdac2 (1:3000, Abcam), APP (1:4000, Millipore), and Lamp2 (1:3000, Abcam). GAPDH (1:12000, Millipore) was used as loading control. Morris water maze (MWM) test MWM was performed as described previously (Sellner et al., 2016). Briefly, animals were trained to locate an escape platform (circle, 10 cm diameter) that was placed 1.0 cm below (hidden) the water surface in a pool of 120 cm diameter. The training consisted of 7 consecutive days with four trials per day. If a mouse failed to find the escape platform within the maximum time (60 s), the animal was placed on the platform for 10 s by the experimenter. The platform location was kept constant and the starting position for the mouse varied between four constant locations at the pool rim. On the 8th day, the platform was removed, and the mice were allowed to swim for 60 s to determine their search bias. On day 9, mice were trained to find a visible platform, which had a 10 cm high pole with a white flag and was changed every trial to a new position. Timing of the latency to find the visible platform was started and ended by the experimenter. Data was acquired and analyzed using BIOBSERVE software (BIOBSERVE). All behavioral experiments were carried out in a double-blind fashion and mice were tested in random order. T-maze test Spatial working memory was tested using the continuous spontaneous alternation task in a T-maze as previously published (Spowart-Manning and van der Staay 2004). In short, animals were set into the base of a T-maze and allowed to explore the maze. Once one arm of the T-maze was entered, the other arm was blocked until the animal returned to the base arm. The exit of the base arm was then blocked for 5 s and the animal again allowed to explore. The experiment was stopped after 14 free choice arm entries. Arm entries were scored as alternations if an animal chose the opposing arm compared to the arm visited immediately prior to the scored instance. Repetitive arm entries were scored as entering the same arm for the third or more consecutive time. QUANTIFICATION AND STATISTICAL ANALYSIS CNS cell quantification For quantification of cells, 12mm thick sagittal cryo sections were immunostained with anti Iba1 (microglia) or anti GFAP (astrocytes) antibody and nuclei were counterstained with DAPI. Cells were counted from 5 - 8 different regions in the cortex and 3 brain slices per animal. For neuronal cell quantification, 12 mm thick sagittal cryosections were immunotained with anti NeuN antibody. Two micrometer thick confocal Z stacks were taken from 3 different regions of cortex, 3 brain slices per animal. NeuN+ neuronal nuclei were then quantified by Imaris software. Quantification of amyloid plaques and co-localization with microglia For quantification of amyloid plaques in the brain tissue, 4 mm thick sagittal paraffin sections were stained with anti-beta A4 antibody followed by peroxidase conjugated secondary antibody and subsequently visualized with 3,30 -Diaminobenzidine (DAB) staining. Beta A4 immunoreactive amyloid plaques were then counted from 5 - 8 different regions in the cortex and from whole hippocampus. Three brain slices per animal were used. Additionally, 12 mm thick sagittal cryosections were co-stained with anti Iba1 and 6E10 antibody for visualization of microglia and amyloid plaques. Two micrometer thick confocal z stacks were rendered with Olympus Fluoview 1000 confocal laser scanning microscope using a 20 3 0.95 NA objective. Area covered by 6E10 immunreactive plaques and Iba1+ microglia as well as the fluorescence intensities of the corresponding channel were analyzed by ImageJ software. Co-localization of amyloid plaques with microglia was analyzed by Coloc tool from IMARIS software (Bitplane). Coloc channel is a virtual channel created by the software to demonstrate the overlapping Iba1 and 6E10 signal.

Immunity 48, 1–16.e1–e6, March 20, 2018 e5

Please cite this article in press as: Datta et al., Histone Deacetylases 1 and 2 Regulate Microglia Function during Development, Homeostasis, and Neurodegeneration in a Context-Dependent Manner, Immunity (2018), https://doi.org/10.1016/j.immuni.2018.02.016

Quantification of synapses The density of synaptic marker Homer1 was analyzed using a custom rule-set in Definiens Developer XD (r). Confocal stacks were imported and quantification restricted to 3 levels in the center of the stack. Specific signals were enhanced and dim background signals removed by applying rank leveling. Stained structures were detected by thresholding and split using watershed segmentation from ImageJ. Number and area of detected objects was exported. Statistical analysis Data are represented as mean ± SEM. Statistical analysis are done using Graph Pad Prism. Comparison of multiple groups is done by one-way or two-way ANOVA with post tests. Comparison of two groups is performed using two tailed unpaired t test. Values are considered significant with p < 0.05. DATA AND SOFTWARE AVAILABILITY All sequencing data (RNA and ChIP-seq) are available at Gene Expression Omnibus (GEO: GSE98823). The microarray dataset is available at GEO (GSE73125).

e6 Immunity 48, 1–16.e1–e6, March 20, 2018