Abstract

The rapid elimination of dying neurons and nonfunctional synapses in the brain is carried out by microglia, the resident myeloid cells of the brain. Here we show that microglia clearance activity in the adult brain is regionally regulated and depends on the rate of neuronal attrition. Cerebellar, but not striatal or cortical, microglia exhibited high levels of basal clearance activity, which correlated with an elevated degree of cerebellar neuronal attrition. Exposing forebrain microglia to apoptotic cells activated gene-expression programs supporting clearance activity. We provide evidence that the polycomb repressive complex 2 (PRC2) epigenetically restricts the expression of genes that support clearance activity in striatal and cortical microglia. Loss of PRC2 leads to aberrant activation of a microglia clearance phenotype, which triggers changes in neuronal morphology and behavior. Our data highlight a key role of epigenetic mechanisms in preventing microglia-induced neuronal alterations that are frequently associated with neurodegenerative and psychiatric diseases.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    Dickstein, D. L., Weaver, C. M., Luebke, J. I. & Hof, P. R. Dendritic spine changes associated with normal aging. Neuroscience 251, 21–32 (2013).

  2. 2.

    O’Banion, M. K., Coleman, P. D. & Callahan, L. M. Regional neuronal loss in aging and Alzheimer’s disease: a brief review. Semin. Neurosci. 6, 307–314 (1994).

  3. 3.

    Woodruff-Pak, D. S. et al. Differential effects and rates of normal aging in cerebellum and hippocampus. Proc. Natl Acad. Sci. USA 107, 1624–1629 (2010).

  4. 4.

    Morterá, P. & Herculano-Houzel, S. Age-related neuronal loss in the rat brain starts at the end of adolescence. Front. Neuroanat. 6, 45 (2012).

  5. 5.

    Mensah, P. L. Stability of large cell-medium cell clusters in the mature neostriatum. Exp. Brain Res. 47, 57–60 (1982).

  6. 6.

    Neumann, H., Kotter, M. R. & Franklin, R. J. M. Debris clearance by microglia: an essential link between degeneration and regeneration. Brain 132, 288–295 (2009).

  7. 7.

    Mass, E. et al. A somatic mutation in erythro-myeloid progenitors causes neurodegenerative disease. Nature 549, 389–393 (2017).

  8. 8.

    Zhao, X. et al. Noninflammatory changes of microglia are sufficient to cause epilepsy. Cell Rep. 22, 2080–2093 (2018).

  9. 9.

    Mass, E. et al. Specification of tissue-resident macrophages during organogenesis. Science 353, aaf4238 (2016).

  10. 10.

    Matcovitch-Natan, O. et al. Microglia development follows a stepwise program to regulate brain homeostasis. Science 353, aad8670 (2016).

  11. 11.

    Mathys, H. et al. Temporal tracking of microglia activation in neurodegeneration at single-cell resolution. Cell Rep. 21, 366–380 (2017).

  12. 12.

    Wang, Y. et al. TREM2 lipid sensing sustains the microglial response in an Alzheimer’s disease model. Cell 160, 1061–1071 (2015).

  13. 13.

    Krasemann, S. et al. The TREM2-APOE pathway drives the transcriptional phenotype of dysfunctional microglia in neurodegenerative diseases. Immunity 47, 566–581.e9 (2017).

  14. 14.

    Butovsky, O. et al. Identification of a unique TGF-β-dependent molecular and functional signature in microglia. Nat. Neurosci. 17, 131–143 (2014).

  15. 15.

    Buttgereit, A. et al. Sall1 is a transcriptional regulator defining microglia identity and function. Nat. Immunol. 17, 1397–1406 (2016).

  16. 16.

    Margueron, R. & Reinberg, D. The polycomb complex PRC2 and its mark in life. Nature 469, 343–349 (2011).

  17. 17.

    Cunningham, C. L., Martínez-Cerdeño, V. & Noctor, S. C. Microglia regulate the number of neural precursor cells in the developing cerebral cortex. J. Neurosci. 33, 4216–4233 (2013).

  18. 18.

    Song, J., Olsen, R. H. J., Sun, J., Ming, G. L. & Song, H. Neuronal circuitry mechanisms regulating adult mammalian neurogenesis. Cold Spring Harb. Perspect. Biol. 8, a018937 (2016).

  19. 19.

    Heiman, M. et al. A translational profiling approach for the molecular characterization of CNS cell types. Cell 135, 738–748 (2008).

  20. 20.

    Madry, C. & Attwell, D. Receptors, ion channels, and signaling mechanisms underlying microglial dynamics. J. Biol. Chem. 290, 12443–12450 (2015).

  21. 21.

    Blander, J. M. The many ways tissue phagocytes respond to dying cells. Immunol. Rev. 277, 158–173 (2017).

  22. 22.

    Galluzzi, L. et al. Molecular definitions of autophagy and related processes. EMBO J. 36, 1811–1836 (2017).

  23. 23.

    Heckmann, B. L., Boada-Romero, E., Cunha, L. D., Magne, J. & Green, D. R. LC3-associated phagocytosis and inflammation. J. Mol. Biol. 429, 3561–3576 (2017).

  24. 24.

    Elmore, M. R. P. et al. Colony-stimulating factor 1 receptor signaling is necessary for microglia viability, unmasking a microglia progenitor cell in the adult brain. Neuron 82, 380–397 (2014).

  25. 25.

    Fourgeaud, L. et al. TAM receptors regulate multiple features of microglial physiology. Nature 532, 240–244 (2016).

  26. 26.

    Swigut, T. & Wysocka, J. H3K27 demethylases, at long last. Cell 131, 29–32 (2007).

  27. 27.

    De Santa, F. et al. The histone H3 lysine-27 demethylase Jmjd3 links inflammation to inhibition of polycomb-mediated gene silencing. Cell 130, 1083–1094 (2007).

  28. 28.

    Kruidenier, L. et al. A selective jumonji H3K27 demethylase inhibitor modulates the proinflammatory macrophage response. Nature 488, 404–408 (2012).

  29. 29.

    Montgomery, N. D. et al. The murine polycomb group protein Eed is required for global histone H3 lysine-27 methylation. Curr. Biol. 15, 942–947 (2005).

  30. 30.

    Jadhav, U. et al. Acquired tissue-specific promoter bivalency is a basis for PRC2 necessity in adult cells. Cell 165, 1389–1400 (2016).

  31. 31.

    von Schimmelmann, M. et al. Polycomb repressive complex 2 (PRC2) silences genes responsible for neurodegeneration. Nat. Neurosci. 19, 1321–1330 (2016).

  32. 32.

    Nestler, E. J. The neurobiology of cocaine addiction. Sci. Pract. Perspect. 3, 4–10 (2005).

  33. 33.

    Friedman, D., Honig, L. S. & Scarmeas, N. Seizures and epilepsy in Alzheimer’s disease. CNS Neurosci. Ther. 18, 285–294 (2012).

  34. 34.

    Hong, S. et al. Complement and microglia mediate early synapse loss in Alzheimer mouse models. Science 352, 712–716 (2016).

  35. 35.

    Goldmann, T. et al. Origin, fate and dynamics of macrophages at central nervous system interfaces. Nat. Immunol. 17, 797–805 (2016).

  36. 36.

    Rothhammer, V. et al. Microglial control of astrocytes in response to microbial metabolites. Nature 557, 724–728 (2018).

  37. 37.

    Rehli, M. et al. Transcription factor Tfec contributes to the IL-4-inducible expression of a small group of genes in mouse macrophages including the granulocyte colony-stimulating factor receptor. J. Immunol. 174, 7111–7122 (2005).

  38. 38.

    Rőszer, T., Menéndez-Gutiérrez, M. P., Cedenilla, M. & Ricote, M. Retinoid X receptors in macrophage biology. Trends Endocrinol. Metab. 24, 460–468 (2013).

  39. 39.

    Dzhagalov, I. L., Chen, K. G., Herzmark, P. & Robey, E. A. Elimination of self-reactive T cells in the thymus: a timeline for negative selection. PLoS Biol. 11, e1001566 (2013).

  40. 40.

    Stowell, R. D. et al. Cerebellar microglia are dynamically unique and survey Purkinje neurons in vivo. Dev. Neurobiol. 78, 627–644 (2018).

  41. 41.

    Grabert, K. et al. Microglial brain region-dependent diversity and selective regional sensitivities to aging. Nat. Neurosci. 19, 504–516 (2016).

  42. 42.

    Wei, S. et al. Functional overlap but differential expression of CSF-1 and IL-34 in their CSF-1 receptor-mediated regulation of myeloid cells. J. Leukoc. Biol. 88, 495–505 (2010).

  43. 43.

    Przanowski, P. et al. The signal transducers Stat1 and Stat3 and their novel target Jmjd3 drive the expression of inflammatory genes in microglia. J. Mol. Med. (Berl.) 92, 239–254 (2014).

  44. 44.

    Stephan, A. H. et al. A dramatic increase of C1q protein in the CNS during normal aging. J. Neurosci. 33, 13460–13474 (2013).

  45. 45.

    Györffy, B. A. et al. Local apoptotic-like mechanisms underlie complement-mediated synaptic pruning. Proc. Natl. Acad. Sci. USA 115, 6303–6308 (2018).

  46. 46.

    Paolicelli, R. C. & Ferretti, M. T. Function and dysfunction of microglia during brain development: consequences for synapses and neural circuits. Front. Synaptic Neurosci. 9, 9 (2017).

  47. 47.

    Tee, W.-W., Shen, S. S., Oksuz, O., Narendra, V. & Reinberg, D. Erk1/2 activity promotes chromatin features and RNAPII phosphorylation at developmental promoters in mouse ESCs. Cell 156, 678–690 (2014).

  48. 48.

    Fujii, S. et al. RAS oncogenic signal upregulates EZH2 in pancreatic cancer. Biochem. Biophys. Res. Commun. 417, 1074–1079 (2012).

  49. 49.

    Souroullas, G. P. et al. An oncogenic Ezh2 mutation induces tumors through global redistribution of histone 3 lysine 27 trimethylation. Nat. Med. 22, 632–640 (2016).

  50. 50.

    Dietmar, R., Rüb, U., Orantes, M. & Braak, H. Phases of A beta-deposition in the human brain and its relevance for the development of AD. Neurology 58, 1791–1800 (2002).

  51. 51.

    Stanley, S. et al. Profiling of glucose-sensing neurons reveals that GHRH neurons are activated by hypoglycemia. Cell Metab. 18, 596–607 (2013).

  52. 52.

    Parkhurst, C. N. et al. Microglia promote learning-dependent synapse formation through brain-derived neurotrophic factor. Cell 155, 1596–1609 (2013).

  53. 53.

    Yona, S. et al. Fate mapping reveals origins and dynamics of monocytes and tissue macrophages under homeostasis. Immunity 38, 79–91 (2013).

  54. 54.

    GENSAT. The Gene Expression Nervous System Atlas (GENSAT) Project, NINDS Contracts N01NS02331 & HHSN271200723701C to The Rockefeller University (New York, NY). http://www.gensat.org/index.html (2003).

  55. 55.

    de Boer, J. et al. Transgenic mice with hematopoietic and lymphoid specific expression of Cre. Eur. J. Immunol. 33, 314–325 (2003).

  56. 56.

    Deng, L. et al. A novel mouse model of inflammatory bowel disease links mammalian target of rapamycin-dependent hyperproliferation of colonic epithelium to inflammation-associated tumorigenesis. Am. J. Pathol. 176, 952–967 (2010).

  57. 57.

    Qian, B.-Z. et al. CCL2 recruits inflammatory monocytes to facilitate breast-tumour metastasis. Nature 475, 222–225 (2011).

  58. 58.

    Doyle, J. P. et al. Application of a translational profiling approach for the comparative analysis of CNS cell types. Cell 135, 749–762 (2008).

  59. 59.

    Srinivas, S. et al. Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev. Biol. 1, 4 (2001).

  60. 60.

    Jung, S. et al. Analysis of fractalkine receptor CX(3)CR1 function by targeted deletion and green fluorescent protein reporter gene insertion. Mol. Cell. Biol. 20, 4106–4114 (2000).

  61. 61.

    Schumacher, A., Faust, C. & Magnuson, T. Positional cloning of a global regulator of anterior-posterior patterning in mice. Nature 384, 648 (1996).

  62. 62.

    Denisenko, O., Shnyreva, M., Suzuki, H. & Bomsztyk, K. Point mutations in the WD40 domain of Eed block its interaction with Ezh2. Mol. Cell. Biol. 18, 5634–5642 (1998).

  63. 63.

    Cardona, A. E. et al. Control of microglial neurotoxicity by the fractalkine receptor. Nat. Neurosci. 9, 917–924 (2006).

  64. 64.

    Bohlen, C. J. et al. Diverse requirements for microglial survival, specification, and function revealed by defined-medium cultures. Neuron 94, 759–773.e8 (2017).

  65. 65.

    Saura, J., Tusell, J. M. & Serratosa, J. High-yield isolation of murine microglia by mild trypsinization. Glia 44, 183–189 (2003).

  66. 66.

    Kriaucionis, S. & Heintz, N. The nuclear DNA base 5-hydroxymethylcytosine is present in Purkinje neurons and the brain. Science 324, 929–930 (2009).

  67. 67.

    Heiman, M., Kulicke, R., Fenster, R. J., Greengard, P. & Heintz, N. Cell type-specific mRNA purification by translating ribosome affinity purification (TRAP). Nat. Protoc. 9, 1282–1291 (2014).

  68. 68.

    Kim, D. & Salzberg, S. L. TopHat-Fusion: an algorithm for discovery of novel fusion transcripts. Genome Biol. 12, R72 (2011).

  69. 69.

    Anders, S., Pyl, P. T. & Huber, W. HTSeq–a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169 (2015).

  70. 70.

    Purushothaman, I. & Shen, L. SPEctRA: First stable release. https://zenodo.org/record/60547#.WyweGlVKipp (2016).

  71. 71.

    Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).

  72. 72.

    Howe, E. A., Sinha, R., Schlauch, D. & Quackenbush, J. RNA-seq analysis in MeV. Bioinformatics 27, 3209–3210 (2011).

  73. 73.

    Chen, E. Y. et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinforma. 14, 128 (2013).

  74. 74.

    Kuleshov, M. V. et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 44, W90–7 (2016). W1.

  75. 75.

    Huntley, R. P. et al. The GOA database: gene Ontology annotation updates for 2015. Nucleic Acids Res. 43, D1057–D1063 (2015).

  76. 76.

    Holtman, I. R. et al. Induction of a common microglia gene expression signature by aging and neurodegenerative conditions: a co-expression meta-analysis. Acta Neuropathol. Commun. 3, 31 (2015).

  77. 77.

    Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).

  78. 78.

    Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930 (2014).

  79. 79.

    Robinson, M. D. & Smyth, G. K. Small-sample estimation of negative binomial dispersion, with applications to SAGE data. Biostatistics 9, 321–332 (2008).

  80. 80.

    Robinson, M. D. & Smyth, G. K. Moderated statistical tests for assessing differences in tag abundance. Bioinformatics 23, 2881–2887 (2007).

  81. 81.

    Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).

  82. 82.

    McCarthy, D. J., Chen, Y. & Smyth, G. K. Differential expression analysis of multifactor RNA-seq experiments with respect to biological variation. Nucleic Acids Res. 40, 4288–4297 (2012).

  83. 83.

    Zhou, X., Lindsay, H. & Robinson, M. D. Robustly detecting differential expression in RNA sequencing data using observation weights. Nucleic Acids Res. 42, e91 (2014).

  84. 84.

    Bennett, M. L. et al. New tools for studying microglia in the mouse and human CNS. Proc. Natl. Acad. Sci. USA 113, E1738–E1746 (2016).

  85. 85.

    Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386 (2014).

  86. 86.

    Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47 (2015).

  87. 87.

    Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25 (2009).

  88. 88.

    Li, H. et al. The sequence alignment/map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).

  89. 89.

    Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008).

  90. 90.

    Shao, N.-Y., Loh, Y.-H.E. and Shen, L. Region-analysis: a Python program for genomic interval annotations. https://github.com/shenlab-sinai/region_analysis (2016).

  91. 91.

    Shen, L., Shao, N., Liu, X. & Nestler, E. ngs.plot: Quick mining and visualization of next-generation sequencing data by integrating genomic databases. BMC Genom. 15, 284 (2014).

  92. 92.

    Liu, Z. et al. Fscn1 is required for the trafficking of TGF-β family type I receptors during endoderm formation. Nat. Commun. 7, 12603 (2016).

  93. 93.

    Liu, X. et al. Salidroside provides neuroprotection by modulating microglial polarization after cerebral ischemia. J. Neuroinflamm. 15, 39 (2018).

  94. 94.

    Willén, K. et al. Aβ accumulation causes MVB enlargement and is modelled by dominant negative VPS4A. Mol. Neurodegener. 12, 61 (2017).

  95. 95.

    Schaefer, A. et al. Control of cognition and adaptive behavior by the GLP/G9a epigenetic suppressor complex. Neuron 64, 678–691 (2009).

  96. 96.

    Paolino, A., Fenlon, L. R., Kozulin, P., Richards, L. J. & Suárez, R. Multiple events of gene manipulation via in pouch electroporation in a marsupial model of mammalian forebrain development. J. Neurosci. Methods 293, 45–52 (2018).

  97. 97.

    Miyamoto, A. et al. Microglia contact induces synapse formation in developing somatosensory cortex. Nat. Commun. 7, 12540 (2016).

  98. 98.

    Tufail, Y. et al. Phosphatidylserine exposure controls viral innate immune responses by microglia. Neuron 93, 574–586.e8 (2017).

  99. 99.

    Schaefer, A. et al. Cerebellar neurodegeneration in the absence of microRNAs. J. Exp. Med. 204, 1553–1558 (2007).

  100. 100.

    LeComte, M. D., Shimada, I. S., Sherwin, C. & Spees, J. L. Notch1-STAT3-ETBR signaling axis controls reactive astrocyte proliferation after brain injury. Proc. Natl. Acad. Sci. USA 112, 8726–8731 (2015).

  101. 101.

    Moyon, S. et al. Functional characterization of DNA methylation in the oligodendrocyte lineage. Cell Rep. 15, 748–760 (2016).

  102. 102.

    Bisht, K. et al. Dark microglia: a new phenotype predominantly associated with pathological states. Glia 64, 826–839 (2016).

  103. 103.

    Sun, H. et al. Attenuation of atherosclerotic lesions in diabetic apolipoprotein E-deficient mice using gene silencing of macrophage migration inhibitory factor. J. Cell. Mol. Med. 19, 836–849 (2015).

  104. 104.

    Kohno, H. et al. Photoreceptor proteins initiate microglial activation via Toll-like receptor 4 in retinal degeneration mediated by all-trans-retinal. J. Biol. Chem. 288, 15326–15341 (2013).

  105. 105.

    Tan, C. L. et al. MicroRNA-128 governs neuronal excitability and motor behavior in mice. Science 342, 1254–1258 (2013).

  106. 106.

    Sullivan, J. M. et al. Autism-like syndrome is induced by pharmacological suppression of BET proteins in young mice. J. Exp. Med. 212, 1771–1781 (2015).

  107. 107.

    Racine, R. J. Modification of seizure activity by electrical stimulation. II. Mot. Seizure Electroencephalogr. Clin. Neurophysiol. 32, 281–294 (1972).

  108. 108.

    Maze, I. et al. Essential role of the histone methyltransferase G9a in cocaine-induced plasticity. Science 327, 213–216 (2010).

Download references

Acknowledgements

We thank D. Littman (NYU School of Medicine, NY) for Cx3cr1CreErt2/+(Litt) mice; N. Heintz (Rockefeller University, NY) for Tg(Prox1-Cre)SJ39 and Tg(Aldh1l1-eGFPL10a) mice; A. Domingos (Instituto Gulbenkian de Ciência, PT) and J. Friedman (Rockefeller University, NY) for Eef1a1LSL.eGFPL10a/+ mice; G. Lemke (Salk Institute, CA) for brain tissue from Axl−/−Mertk−/− mice; the Flow Cytometry CoRE at Icahn School of Medicine for cytometry assistance; I. Lemischka (Icahn School of Medicine, NY) for Fluidigm C1 use; O. Butovsky (Harvard Medical School, MA) for P2RY12 antibody; and J.W. Murray for ImageJ automation. We are very grateful to D. Schafer, E. Marcora, A. Tarakhovsky, J.W. Murray, and all Schaefer lab and Ronald M. Loeb Center members for helpful discussions and comments on the manuscript. This work was supported by the National Institutes of Health (NIH) Director New Innovator Award DP2 MH100012-01 (A.S.), R01NS091574 (A.S.), R21 MH115353 (A.S.), 1 RF1 AG054011-01 (A.M.G. and A.S.), JPB Foundation (A.M.G.), NARSAD Young Investigator Award #25065 (P.A.), Boehringer Ingelheim and the Austrian Industrial Research Promotion Agency FFG-852936 (M.B.), T32AG049688 (A.B.), F30MH106293 (J.R.S), and Ruth L. Kirschstein NRSA fellowship F31MH111147 (J.M.S).

Author information

Affiliations

  1. Fishberg Department of Neuroscience, Department of Psychiatry, Friedman Brain Institute, Icahn School of Medicine at Mount Sinai, New York, NY, USA

    • Pinar Ayata
    • , Ana Badimon
    • , Hayley J. Strasburger
    • , Mary Kaye Duff
    • , Sarah E. Montgomery
    • , Yong-Hwee E. Loh
    • , Anna A. Pimenova
    • , Brianna R. Ramirez
    • , Andrew T. Chan
    • , Josefa M. Sullivan
    • , Immanuel Purushothaman
    • , Alison M. Goate
    • , Li Shen
    •  & Anne Schaefer
  2. Ronald M. Loeb Center for Alzheimer’s Disease, Icahn School of Medicine at Mount Sinai, New York, NY, USA

    • Pinar Ayata
    • , Anna A. Pimenova
    • , Alison M. Goate
    •  & Anne Schaefer
  3. Research Institute of Molecular Pathology, Vienna Biocenter, Vienna, Austria

    • Anja Ebert
    •  & Meinrad Busslinger
  4. Department of Genetics & Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, USA

    • Joseph R. Scarpa
    • , Alison M. Goate
    •  & Bojan Losic
  5. Tisch Cancer Institute, Cancer Immunology Program, Icahn School of Medicine at Mount Sinai, New York, NY, USA

    • Bojan Losic

Authors

  1. Search for Pinar Ayata in:

  2. Search for Ana Badimon in:

  3. Search for Hayley J. Strasburger in:

  4. Search for Mary Kaye Duff in:

  5. Search for Sarah E. Montgomery in:

  6. Search for Yong-Hwee E. Loh in:

  7. Search for Anja Ebert in:

  8. Search for Anna A. Pimenova in:

  9. Search for Brianna R. Ramirez in:

  10. Search for Andrew T. Chan in:

  11. Search for Josefa M. Sullivan in:

  12. Search for Immanuel Purushothaman in:

  13. Search for Joseph R. Scarpa in:

  14. Search for Alison M. Goate in:

  15. Search for Meinrad Busslinger in:

  16. Search for Li Shen in:

  17. Search for Bojan Losic in:

  18. Search for Anne Schaefer in:

Contributions

P.A. and A.S. designed the study. P.A. did in vivo molecular & imaging experiments. A.B. and P.A ran behavioral experiments. A.A.P., P.A., A.S., and A.M.G. designed the in vitro experiments, which were performed by H.J.S., M.K.D., A.A.P., A.B., and P.A. S.E.M. and A.B. did morphological analyses. H.J.S., M.K.D., B.R.R., S.E.M., J.M.S., and A.T.C. performed immunoblotting. Y.-H.E.L., I.P., and L.S. performed bioinformatics for bulk RNA and ChIP sequencing. J.R.S. and B.L. analyzed single nuclei sequencing data. A.E. and M.B. generated the EedloxP/loxP mice. A.S. and P.A. wrote the manuscript. All authors discussed results and provided input on the manuscript.

Competing interests

A.M.G. receives financial compensation as an advisory board member for Denali Therapeutics.

Corresponding author

Correspondence to Anne Schaefer.

Integrated supplementary information

  1. Supplementary Figure 1 cbMg display cellular phenotype associated with active cell clearance.

    (a) cbMg show a less ramified / less surveillant morphology. Microglia morphology was evaluated in striatum / cerebellum of brain sections from 4/6mo wild-type mice. Left: Representative projection of 3D z-stacked images showing single IBA1+ microglia (IBA1: green, DAPI: blue) with manual 2D tracing of microglial processes. Scale: 10 μm. Right: Representative microglia reconstruction (Neurolucida) with concentric circles for Sholl analysis. Right: Quantification of regional microglia morphology. Numbers of primary processes: stMg: mean=5.325, SEM=0.2393; cbMg: mean=3.100, SEM=0.1871; p=0.0014, F=1.636, t4=7.849. Sholl intersections at 10 μm interval: stMg: mean 19.83, SEM=2.181; cbMg: mean= 11.08, SEM=2.316; p=0.0195, F=1.127, t4=3.774. Number of branch endings: stMg: mean=23.44, SEM=2.718; cbMg: mean= 11.92, SEM=1.728; p=0.0135, F=2.476, t4=4.219. Combined length of processes (μm): stMg: mean 262.4, SEM=31.47; cbMg: mean= 149.2, SEM=40.91; p=0.033, F=1.69, t4=3.198. Branch order: stMg: mean=85.79, SEM=14.08; cbMg: mean=40.67, SEM=8.296; p=0.0262, F=2.882, t4=3.446. n=35 cells from n=5 mice. Longest branch (μm): stMg: mean=37.29, SEM=1.560; cbMg: mean=36.80, SEM=1.136; p=0.7206, F=1.885, t2=0.4115. Soma size (μm2): stMg: mean=46.47, SEM=2.369; cbMg: mean=47.10, SEM=2.164; p=0.9007, F=1.198, t2=0.1411. n=25 cells from n=3 mice. All t-tests were two-tailed paired. (b,c) Microglia in different brain regions display distinct morphological features. (b) CD68+ lysosome content (green) in P2RY12+ microglia (red) was determined using immunofluorescence analysis from brain sections of 4/6mo wild-type mice (DAPI: blue). Scale: 10 μm. Representative image is shown. (c) Quantification of regional microglial morphological properties. Microglial density (mm-1): olfactory bulb microglia (olfMg): mean=130.6, SEM=14.14; striatum microglia (stMg): mean=166.0, SEM=19.05; cortex microglia (cxMg): mean=122.4, SEM=4.713; dentate gyrus microglia (dgMg): mean=138.1, SEM=12.6; cerebellum microglia (cbMg): mean=73.47, SEM=4.713; p=0.0048, F=7.41. Percent P2RY12+ area covered per microglia: olfMg: mean=1.2, SEM=0.168; stMg: mean=0.909, SEM=0.0375; cxMg: mean=1.16, SEM=0.0878; dgMg: mean=0.924, SEM=0.0323; cbMg: mean=0.3646, SEM=0.07586; p=0.0006, F=12.59. Percent CD68+ lysosomal area / microglia: olfMg: mean=0.02393, SEM=0.001433; stMg: mean=0.01349, SEM=0.0025; cxMg: mean=0.01315, SEM=0.00119; dgMg: mean=0.03087, SEM=0.00295; cbMg: mean=0.03889, SEM=0.00147; p< 0.0001, F=30.22. Ratio of CD68+ lysosomal area / microglia area: olfMg: mean=0.01576, SEM=0.0049; stMg: mean=0.015, SEM= 0.003; cxMg: mean=0.014, SEM=0.002; dgMg: mean=0.0190, SEM=0.0025; cbMg: mean=0.126, SEM=0.0225; p<0.0001, F=22.02. One-way ANOVA with Tukey’s Multiple Comparison post-hoc analysis to compare all pairs of columns. 3-5 sections of whole brain regions from n=3 wild-type mice. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. Bar graphs with individual data points show mean ± SEM. (d) cbMg lysosomes contain DAPI+ material. Representative immunofluorescence images of microglia (left) and their 3D reconstruction (right, Imaris, related to Fig.1) from striatum and the granule cell layer and molecular layer of the cerebellum. IBA1+ microglia: green; CD68+ lysosomes:red; and DAPI:blue. Scale: 5 μm. Side scatters of microglia containing DAPI+/CD68+ lysosome are shown underneath each forward scatter for cbMg with the DAPI channel masked for the microglia surface. 3D axes are shown. Arrows point to the box enclosing DAPI+/CD68+ lysosome inside IBA1+ microglia with a zoomed in view. Experiment was repeated 3 times independently with similar results.

  2. Supplementary Figure 2 Generation of microglia-specific TRAP mice.

    (a) Analysis of previously described microglia-specific Cre mouse lines revealed non-exclusive and/or incomplete targeting of microglia in the adult mouse brain for many of the tested Cre-lines (summarized in table). GFP/YFP: green; microglial marker, IBA1: red; and DAPI: blue from cerebellum and striatum of Tg(Prox1-Cre)SJ39; Eef1a1LSLeGFPL10a/+, Tg(Csf1r-iCre);R26LSL.eYFP/+, Tg(Csf1r-Cre/Esr1*);Eef1a1LSLeGFPL10a/+, and Tg(Vav1-iCre);Eef1a1LSLeGFPL10a/+ mice are shown. Scale: 20 μm. Tg(Csf1r-iCre) and Tg(Vav1-iCre) lines were excluded due to neuronal targeting in brain regions of interest. Tg(Csf1r-Cre/Esr1*) line was excluded due to neuronal targeting and insufficient labeling of microglia. Tg(Prox1-Cre)SJ39 line exhibited exclusive and complete microglia targeting in a limited number brain regions (cerebellum, striatum, and cortex) and was used as an alternative line to target microglia in these regions. Arrows indicate non-microglial targeting. (b) Microglia-specific GFP expression in the striatum, cerebellum, brain stem, and dentate gyrus of Cx3cr1CreErt2/+(Litt);Eef1a1LSLeGFPL10a/+ mice. Scale: 20 μm. Top: GFP/YFP: green; IBA1+ microglia: red; and DAPI: blue; Bottom: GFP/YFP+ cells: green; NeuN+ neurons (left): red or GFAP+ astrocytes/Bergmann glia (right): red; and DAPI: blue. (c) Expression of eGFP-L10a protein across the microglial cell body, including protrusions, is visualized in Cx3cr1CreErt2/+(Jung);Eef1a1LSLeGFPL10a/+ mice. GFP: green; IBA1+ microglia: red; and DAPI: blue from striatum and cerebellum of Cx3cr1CreErt2/+(Jung);Eef1a1LSLeGFPL10a/+ mice. Scale: 10 μm. Representative immunofluorescence images from > 2 independent experiments are shown.

  3. Supplementary Figure 3 Validation of microglia-specific TRAP mice.

    (a) Schematic showing TRAP-sequencing from microglia-specific and astrocyte-specific TRAP mice. (b) Volcano plot for differential expression analysis (DESeq2, n=2/genotype/region) of each gene. y-axis: -log10 (adjusted p-value), and x-axis: log2 (fold change) between astrocyte- and microglia-TRAP. 885 genes (green) are significantly enriched in microglia TRAP; 970 genes (blue) are significantly enriched in astrocyte TRAP. (c) Heatmap with hierarchical clustering distances shows the variation in the expression levels (z-scored log2 RPKM) of 885 microglia genes and 970 astrocyte genes. (d) Selected Gene Ontology (GO) annotations enriched (Enrichr) for 885 microglia-enriched genes (green) and 970 astrocyte-enriched genes (blue). y-axis: -log10 (p-value). Dotted lines: p-value=0.05. (e) Top 50 microglia genes and top 50 astrocyte genes ranked by highest expression, listed in bar graph with RPKM of each gene (y-axis) in microglia- (green) and astrocyte-TRAP (blue). Bar graphs show mean ± SEM (n=4/genotype). (f) cbMg and stMg are heterogeneous independent of the specific microglia-TRAP mouse line used. Principal component analysis (PCA) based on the expression of top 500 differentially expressed genes of TRAP from the striatum and cerebellum of n=6 Cx3cr1CreErt2/+;Eef1a1LSLeGFPL10a/+ (circles) and n=2 Tg(Prox1-Cre)SJ39; Eef1a1LSLeGFPL10a/+ (squares) mice and from the striatum and cerebellum of n=2 Tg(Aldh1l1-eGFP/L10a) astrocyte-specific TRAP mice is shown. (g) Enrichment of microglia-specific genes in microglia-specific TRAP vs astrocyte-specific TRAP or TRAP from wild-type controls. Left: Heatmap with hierarchical clustering distances shows the variation in the expression levels of 885 microglia-enriched genes in TRAP samples from the cerebellum and striatum of Cx3cr1CreErt2/+;Eef1a1LSLeGFPL10a/+, Tg(Prox1-Cre)SJ39; Eef1a1LSLeGFPL10a/+, Tg(Aldh1l1-eGFP/L10a), and wild-type control mice. Right: Box-and-whisker plots show relative expression of these genes in pooled TRAP samples from the two microglia-specific TRAP lines, Cx3cr1CreErt2/+;Eef1a1LSLeGFPL10a/+ (n=12) and Tg(Prox1-Cre)SJ39;Eef1a1LSLeGFPL10a/+ (n=4), from astrocyte-specific TRAP line, Tg(Aldh1l1-eGFP/L10a) (n=4), and wild-type control mice to assess/exclude nonspecific mRNA binding (n=3). Cx3cr1CreErt2/+: min=0.17, 25%=0.97, median=1.42, 75%=1.80, max=2.83; Tg(Prox1-Cre)SJ39: min=-0.58, 25%=0.66, median=1.12, 75%=1.53, max= 2.99 Tg(Aldh1l1): min=-6.12, 25%=-4.22, median=-3.41, 75%=-2.31, max=-0.91; ctrl: min=-8.48, 25%=-5.12, median=-3.87, 75%=-2.20, max=0.025; p<0.0001, F=3888, Kruskal-Wallis statistic (KWS)=2166, One-way ANOVA (Kruskal-Wallis test) with Dunn’s Multiple Comparison post-hoc analysis to compare all pairs of columns. Box-and-whisker plots (Tukey) and heatmaps show normalized expression values (z-scored log2 RPKM). ***p ≤ 0.001.

  4. Supplementary Figure 4 Microglia-specific TRAP analysis reveals region-specific expression patterns and prevents ex vivo isolation-induced activation of microglia.

    (a) Schematic showing FACS-based microglia isolation performed in parallel with microglia-specific TRAP-seq. (b) Left: Representative flow cytometry data illustrates isolation of Percoll-enriched microglia (Methods) from the striatum / cerebellum of Cx3cr1CreErt2/+; Eef1a1LSLeGFPL10a/+ mice. x-axis: GFP signal, y-axis: Aqua Live/Dead signal to select for live cells. Gate for GFP-positive cells. Percentage of the populations gated compared to total scatter is shown. Right: Quantification of percentages of GFP+ events in the striatum (mean=50.23, SEM=4.491) and cerebellum (mean=40.97, SEM=7.441) p=0.0928, F=2.746, t2=3.050. Bar graphs with individual data points show mean ± SEM. Two-tailed paired t-test (n=3 independent experiments). (c) mRNAs obtained by TRAP and by microglia isolation display a high correlation by linear (Pearson) correlation. Representative scatter plots of genic expression with correlation values between individual (n=1 replicate) TRAP and FACS samples from stMg and cbMg. Scatter plots of genic expression were reproduced with similar results for all replicates. (d) MA plot for differential expression analysis (DESeq2, n=2/region/method) of each gene shows log2 (mean expression) across all samples (x-axis) vs log2 (fold change) by FACS over TRAP (y-axis). 683 genes (red) are significantly induced by FACS. Genes in blue are down-regulated by FACS and genes in green are equally enriched. (e) Heatmap shows the variation in the expression levels (z-scored log2 RPKM) of 40 pro-inflammatory and immediate early genes in FACS and TRAP samples. (f) Genome browser views (IGV) of selected genes in TRAP and FACS samples.

  5. Supplementary Figure 5 cbMg display a more immature and neurodegeneration-associated phenotype as compared to a more mature homeostatic phenotype in stMg.

    (a) MA plot for differential expression analysis (DESeq2, n=2/region) for each gene shows log2 (mean expression) across all samples (x-axis) and log2 (fold change) of expression of stMg over cbMg from 9mo Cx3cr1CreErt2/+;Eef1a1LSLeGFPL10a/+ mice (y-axis). 733 genes (purple) are significantly enriched in stMg, 297 genes (orange) are significantly enriched in cbMg, and genes in green are equally enriched. Pie charts show the GO-based categories of 733 stMg-enriched genes (top) and 297 cbMg-enriched genes (bottom). (b) Heatmaps show the variation in the expression levels of genes involved in engulfment, metabolism, transcription regulation, and chromatin modification from the 297 cbMg-enriched genes identified by microglia-TRAP from 9mo mice (DESeq2, n=2). Also shown: selected proinflammatory genes that are not significantly enriched. z-scored log2 RPKM of genes in cbMg and stMg of 3-,6-, and 9-month-old mice are shown. (c) Left: Venn diagrams show overlap between cbMg (orange) and stMg (purple) enriched genes with immature microglia markers10 (red; stMg: p=0.0083, OR= 1.809, χ2=7.003; cbMg: p < 0.0001, OR= 8.083, χ2=167.0) and mature microglia markers10 (blue; stMg: p < 0.0001, OR=10.45, χ2=548.1; cbMg: p=0.2416, OR=1.518, χ2=1.371). Right: Venn diagrams show overlap between cbMg (orange) and stMg (purple) enriched genes with genes up- (red; stMg: p=0.9149, OR=0.9773, χ2=0.3533; cbMg: p < 0.0001, OR=4.113, χ2=69.38) and down-regulated (blue; stMg: p < 0.0001, OR=11.41, χ2=755.2; cbMg: p=0.93, OR=1.034, χ2=0.007709) in Alzheimer’s Disease (AD)-associated microglia in mice12. (d) Venn diagrams show overlap between cbMg (orange) and stMg (purple) enriched genes with microglia associated with early response to AD11 (cluster 3; red; stMg: p =0.543, OR= 0.8897, χ2=0.37; cbMg: p < 0.0001, OR= 3.374, χ2=54.44) (cluster 7; red; stMg: p =0.7595, OR= 1.056, χ2=0.09368; cbMg: p < 0.0001, OR= 2.579, χ2=26.02) and late response to AD11 (cluster 6; red; stMg: p =0.6705, OR= 0.9227, χ2=0.181; cbMg: p < 0.0001, OR= 3.903, χ2=77.47). (c,d) Genes found in significant overlaps are indicated. p values and odds ratios of Venn diagrams (underneath each overlap) were calculated using the non-parametric χ2 test from a total of 22,706 protein-coding genes analyzed in DESeq2.

  6. Supplementary Figure 6 Single-microglial-nuclei RNA-sequencing confirms the distinct molecular signatures of cbMg and stMg.

    (a) Schematic showing microglia-TRAP, microglial cell isolation by FACS and microglial nuclei isolation performed in parallel. (b) Representative flow cytometry data illustrates isolation of microglial nuclei singlets from the striatum and cerebellum of Cx3cr1CreErt2/+;Eef1a1LSLeGFPL10a/+ mice. Left: x-axis: DyeCycle Ruby to quantify the number of nuclei, y-axis: GFP. Gate for singlets of nuclei is depicted. Right: x-axis: GFP signal, y-axis: Pacific Blue signal to check for autofluorescence. Gate for GFP-positive singlets. Percentage of the populations gated compared to total scatter is shown. Quantification of percentages of low DyeCycle Ruby events (indicating nuclei singlets) in the striatum (mean=82.70, SEM=3.821) and cerebellum (mean=82.47, SEM= 2.682) p=0.9551, F=2.03, t2=0.0635. Quantification of percentages of GFP+ events (indicating microglial nuclei singlets) in the striatum (mean=3.367, SEM=0.5925) and cerebellum (mean=0.2, SEM= 0.0577) p=0.0275, F=105.3, t2=5.903. Bar graphs with individual data points show mean ± SEM. *P ≤ 0.05, T-tests were two-tailed paired (n=3 independent experiments). (c) Linear (Pearson) correlation of RNA obtained by nuclear isolation and TRAP. Representative scatter plots of genic expression (cpm) with correlation values from cbMg and stMg (n=1 replicate). Scatter plots of genic expression were reproduced with similar results for all replicates. (d) The variation in the expression levels (z-scored log2 RPKM) of 40 selected immediate early and proinflammatory genes. Left: heatmap with hierarchical clustering distances; right: box-and-whisker plots (Tukey) for TRAP, sorted microglial nuclei, and sorted microglial cells. TRAP: min=-5.775, 25%=-2.195, median=-1.36, 75%=-0.86, max=0.127; nuclei: min=-5.97, 25%=-3.088, median=-2.458, 75%=-1.29, max=-0.5081; cells: min=2.038, 25%=3.123, median=3.707, 75%=4.97, max=7.555; p < 0.0001, F=271.3, KWS=82.65, one-way ANOVA (Kruskal-Wallis test) with Dunn’s Multiple Comparison post-hoc analysis to compare all pairs of columns (n=4/method). ***p ≤ 0.001. (e) Linear (Pearson) correlation among individual nuclei samples and between total nuclear RNA, TRAP and microglial cell RNA. Representative scatter plots of genic expression (cpm) with correlation values from cbMg and stMg (n=1 replicate). Scatter plots of genic expression were reproduced with similar results for all replicates. (f) Number of counts per gene (cpm, top) and number of genes (bottom) obtained from each of 34 stMg (purple) and 29 cbMg (orange) single nuclei after quality control filtering. (g) Modified MA plot for regional contrast provides overview of transcriptional landscape with percentage of cells expressing the gene (x-axis) and log2 (fold change) of expression (y-axis) from n=34 stMg/29 cbMg single nuclei (edgeR-QL from 4 Cx3cr1CreErt2/+;Eef1a1LSLeGFPL10a/+ mice). stMg- (purple), cbMg-enriched (orange) genes are shown. Genes in green are equally enriched. Scatter plot shows -log10 p of the GO term enrichment for the stMg-DEGs (y-axis) vs cbMg-DEGs (x-axis). stMg-enriched (purple), cbMg-enriched (orange), equally enriched (green) GOs are shown. Selected GOs are named. Dotted lines: p = 0.001. (h) Scatter plot compares the log2 (fold change) by TRAP (x-axis) vs by microglial single-nuclei sequencing (y-axis) for the 297 cbMg- and 734 stMg-enriched genes identified by TRAP. Genes that are enriched in cbMg (orange) and stMg (purple) by both methods are shown. GO annotations (Enrichr) enriched for 314 stMg-enriched and 82 cbMg-enriched genes by both methods. y-axis: -log10 p. Dotted lines: p=0.05. Pie charts show the GO-based categories of 314 stMg-enriched genes (top) and 82 cbMg-enriched genes (bottom).

  7. Supplementary Figure 7 Confirmation of differentially expressed genes in stMg and cbMg at RNA and protein level.

    (a,b) Microglia cells that are positive for the indicated RNAs in striatum and cerebellum of brain sections from 4mo Cx3cr1GFP/+ mice were determined using a combination of immunohistochemistry for GFP (green) with RNA in situ hybridization (ISH, red). ISH: Mrc1 (a, first row), Axl (a, second row), Apoe (a, third row), Tfec (a, fourth row), Ecscr (b, first row), Gpr56 (b, second row). (DAPI: blue). Scale: 10 μm. Representative image (left) and quantification of percentage of microglia cells that are positive for the indicated RNA ISH (right). Mrc1: stMg: mean=1.852, cbMg: mean=20.81. Axl: stMg: mean=22.36, cbMg: mean=61.11. Tfec: stMg: mean=4.688, cbMg: mean=25.73. Apoe: stMg: mean=35.62, cbMg: mean=75.75. Ecscr: stMg: mean=86.84, cbMg:mean=9.723. Gpr56: stMg: mean=67.59, cbMg: mean=14.17. > 50 cells from n=2 mice. Dotted circles indicate microglial soma. (c) Brain sections from 4mo Cx3cr1GFP/+ mice were tested for RNA in situ hybridization (ISH, red) to positive and negative controls provided by manufacturer (RNAScope. DAPI: blue). Scale: 10 μm. Dotted circles indicate cells positive for ISH signal. The experiment was repeated independently twice with similar results. (d) MHCII, CD74, AXL, and ApoE expression (red) in GFP+/P2RY12+/IBA1+ microglia (green) were determined using immunofluorescence analysis from striatum and cerebellum of brain sections from 4mo Cx3cr1GFP/+ or wild-type mice (DAPI: blue). Scale: 20 μm. Representative image (left) and quantification of ratio protein area / microglia area (right) is shown. MHCII: stMg: mean=0.02166, cbMg: mean=0.3239; 8 images/region from n=2 mice. CD74: stMg: mean=0.023, cbMg: mean=0.16; 8 images/region from n=2 mice. AXL: stMg: mean=0.2550, cbMg: mean=0.7506; 10 images/region from n=2 mice. ApoE: stMg: mean=5.585, SEM=2.097; cbMg: mean=22.81, SEM=1.829; p=0.0035, F=1.315, t4=6.189; 11 images/region from n=3 mice. A two-tailed paired t-test was performed for all quantifications. Bar graphs with individual data points show mean ± SEM. (e) Apoptosis of human Jurkat cells was induced with 1 μM Staurosporine for 3 hours, and induction of apoptosis was verified using FITC-Annexin V / propidium iodide kit. Quantification of necrotic (mean=0.4067, SEM=0.1225), late apoptotic (mean=5.877, SEM=0.3910), live (mean= 17.48, SEM=1.589), and early apoptotic (mean=76.23, SEM=1.492), p<0.0001, F=990.1. One-way ANOVA with Tukey’s Multiple Comparison post-hoc analysis to compare all pairs of columns (n=3 wells of Jurkat cells). ***p ≤ 0.001. (f) Representative flow cytometry histogram (related to Fig. 2d) overlay of stMg / cbMg isolated from 3/4mo Cx3cr1GFP/+ mice illustrates that cbMg are more efficient in engulfment of pHrodo-labeled early apoptotic Jurkat cells. x-axis: pHrodo intensity. Histograms show cbMg / stMg (1) alone, (2) with early apoptotic cells, and (3) with early apoptotic cells and Cytochalasin D (phagocytosis inhibitor). The experiment was repeated independently 4 times with similar results.

  8. Supplementary Figure 8 cbMg are involved in the clearance of dying cells.

    (a) Schematic showing strategy for depletion of microglia in wild-type mice using PLX diet. (b) PLX treatment for 3 weeks ablates ~99% of all microglia in the adult brain. Immunofluorescence images with from the striatum / cerebellum of 4mo wild-type mice (3 weeks of PLX or control diet) IBA1+ microglia: green, DAPI: blue. Scale: 50 μm. Representative immunofluorescence images from at least 10 independent experiments are shown. (c) Absence of microglia reveals accumulation of apoptotic cells in the cerebellum. Representative immunofluorescence images are shown from striatum and cerebellum of brain sections from 4/5mo PLX-treated and control mice (6-8 weeks of PLX treatment). NeuN+ neurons: red; DAPI: blue. Scale: 100 μm. Arrows and dotted circles indicate cCASP3+ apoptotic cells (green). The experiment was repeated independently 4 times with similar results. (d) Staining controls show specific cCASP3 staining. Top: secondary antibody staining of the cerebellum of an adult wild-type mouse. Bottom: Representative immunofluorescence image from the developing cerebellum of brain sections from a P7 mouse. Arrows and dotted circles indicate cCASP3+ apoptotic cells (green). NeuN+ neurons: red; DAPI: blue. Scale: 100 μm. The experiment was repeated independently twice with similar results. (e) Confirmation of cerebellum-specific cell death by TUNEL staining (green) in the cerebellum of brain sections from 4mo PLX-treated mice. NeuN: red, DAPI: blue. Scale: 10 μm. Dotted circles indicate TUNEL+/NeuN+ cell. Number of TUNEL+ cells per cm2 of control (striatum: mean=0.0, SEM=0.0; cerebellum: mean= 23.15, SEM= 23.15) or PLX-treated mice (striatum: mean= 11.57, SEM= 11.57; cerebellum: mean=118.1, SEM=10.61); p=0.0012, F=14.95; 3-5 cerebellum / striatum sections from n=3 mice/group. One-way ANOVA with Tukey’s Multiple Comparison post-hoc analysis to compare all pairs of columns. Bar graphs with individual data points show mean ± SEM, **P ≤ 0.01.

  9. Supplementary Figure 9 Characterization of apoptotic cells in the adult brain.

    (a) 15% of the dying cells are of neuronal origin. Left: Immunolfluorescence images to assess overlap between Nissl+ neurons / GFAP+ astrocytes / Bergmann glia / OLIG2+ oligodendrocytes / CD11b+ macrophages/microglia (green), with cCASP3+ apoptotic cells (red) in the cerebellum of PLX-treated mice. Arrows point to cCASP3+ cells. Dotted circles indicate CD11b+ cells in the meninges. DAPI: blue. Scale: 100 μm. Right: Quantification of overlap of cCASP3+ cells with Nissl+ (mean=11.58, SEM=0.9391), NeuN (mean=12.78, SEM=1.470), GFAP+ (mean=0, SEM=0), OLIG2+ (mean=0, SEM=0), CD11b+ (mean=0, SEM=0), and GFP+ (mean=0, SEM=0) cells. p<0.0001, F=78.31; 3-5 cerebellum/striatum sections from n=3 wild-type mice. PLX-treated Cx3cr1GFP/+ mice were used to test for the existence of GFP+ apoptotic macrophages/microglia. ***p ≤ 0.001. (b) Apoptotic cells in the cerebellum are distributed within all layers. Top: Representative immunofluorescence images are shown from the Purkinje cell layer and the molecular layer of cerebellum of brain sections from 4mo PLX-treated mice. Nissl+ neurons: green; DAPI: blue. Scale: 10 μm. Arrows point to cCASP3+ apoptotic cells (red). Bottom: Bar graph with individual data points shows percent localization of cCASP3+ cells. Granule cell layer: mean=0.4950, SEM=0.07649, molecular layer: mean= 0.1868, SEM= 0.04645, white matter: mean= 0.2042, SEM= 0.06745, and Purkinje cell layer: mean=0.1140, SEM=0.02625; p=0.0072, F=8.516; 3-5 cerebellum / striatum sections from n=3 PLX-treated wild-type mice. (c) Apoptotic cells numbers in the cerebellum are comparable to neurogenic regions. Quantification of cCASP3+ apoptotic cells (cm-2) in neurogenic and non-neurogenic regions. olf: mean=101.3, SEM=18.20; st: mean= 3.333, SEM=1.667, cx: mean=4.938, SEM=4.938, dg: mean=500.0, SEM=19.24; cb: mean=110.0, SEM=25.17; p<0.0001, F=155.0. Ratio of cCASP3+ apoptotic cells per 100 microglia: olf: mean=0.73, SEM=0.1311; st: mean= 0.0196, SEM=0.0196, cx: mean= 0.0403, SEM= 0.0403, dg: mean= 3.387, SEM= 0.1304; cb: mean=2.450, SEM=0.2917; p<0.0001, F=95.10. Ratio of cCASP3+ apoptotic cells per percent area covered by microglia olf: mean=0.8832, SEM=0.235; st: mean=0.0337, SEM=0.01985, cx: mean=0.02604, SEM=0.02604, dg: mean=1.988, SEM=0.2696; cb: mean=7.201, SEM=0.6095; p<0.0001, F=64.29. 12 whole brain sections from n=3 mice. (d) Cell division rates in the cerebellum are not comparable to neurogenic regions. Quantification of density of Ki67+ cells (cm-2) undergoing cell division in neurogenic and non-neurogenic regions. olf: mean=486.6, SEM=41.30; st: mean= 55.32, SEM= 8.583, cx: mean= 64.76, SEM= 8.330, dg: mean= 2050, SEM= 132.3; cb: mean=125.0, SEM=8.660; p<0.0001, F=187.6. 16 whole brain sections from n=4 mice. Bar graphs with individual data points show mean ± SEM. One-way ANOVA with Tukey’s Multiple Comparison post-hoc analysis to compare all pairs of columns. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

  10. Supplementary Figure 10 Exposure to dying cells triggers the expression of the cbMg-enriched clearance genes in forebrain microglia in vitro.

    (a) Schematic showing exposure of primary microglia to early apoptotic cells in vitro. (b-g) Dot plots show relative expression of selected cbMg-enriched clearance-associated, stMg-enriched surveillance-associated, or pro-inflammatory genes in primary microglia exposed to vehicle or early apoptotic cells. (b) Anxa2 (4hr: p<0.0001, t4=11.13; 12hr: p<0.0001, t4=36.96), Colec12 (4hr: p=0.072, t4=2.433; 12hr: p=0.002, t4=7.498), Lilrb4 (4hr: p=0.019, t4=3.799; 12hr: p<0.0001, t4=14.04), Apoe (4hr: p=0.926, t4=0.09957; 12hr:p=0.002, t4=6.964), Cd74 (4hr: p=0.146, t4=1.801; 12hr: p=0.001, t4=8.070), Ptch1 (4hr:p=0.008, t4=4.920; 12hr: p=0.037, t4=3.066), Clec7a (4hr: p=0.001, t4=9.045; 12hr: p=0.015, t4=4.064), Msr1 (4hr: p=0.436, t4=0.8643; 12hr: p=0.011, t4=4.448), Lyz2 (4hr: p=0.092, t4=2.203; 12hr: p=0.024, t4=3.529), Ptplad2 (4hr: p=0.004, t4=6.109; 12hr: p=0.009, t4=4.809). (c) Pparg (4hr: p=0.916, t4=0.1120; 12hr: p=0.002, t4=7.144), Jdp2 (4hr: p<0.0001, t4=16.66; 12hr: p<0.0001, t4=15.16), Rarg (4hr: p<0.0001, t4=20.03; 12hr: p=0.001, t4=10.16), Tfec (4hr: p=0.002, t4=7.261; 12hr: p<0.0001, t4=11.47); Ahr (4hr: p<0.0001, t4=14.08; 12hr: p=0.004, t4=6.033), En2 (4hr: p=0.468, t4=0.8004; 12hr: p=0.025, t4=3.493), Tead4 (4hr: p=0.624, t4=0.5296; 12hr: p=0.002, t4=7.013). (d) Slc2a5 (4hr: p=0.088, t4=2.496; 12hr: p<0.0001, t4=15.30), Asb2 (4hr: p<0.0001, t4=13.82; 12hr: p<0.0001, t4=25.35), Tmem119 (4hr: p=0.001, t4=9.393; 12hr: p=0.003, t4=6.721), Fscn1 (4hr: p=0.501, t4=0.7381; 12hr: p<0.0001, t4=18.93), P2ry12 (4hr: p=0.175, t4=1.647; 12hr: p=0.733, t4=0.3659), Fcrls (4hr: p=0.043, t4=2.937; 12hr: p=0.001, t4=10.31). (e) Hhex (4hr: p=0.961, t4=0.05233; 12hr: p=0.025, t4=3.501), Esr1 (4hr: p=0.362, t4=1.029; 12hr: p=0.018, t4=3.885), Irf8 (4hr: p=0.023, t4=3.610; 12hr: p=0.001, t4=9.580), Sall1 (4hr: p=0.034, t4=3.163; 12hr: p=0.139, t4=1.842), Sall3 (4hr: p=0.695, t4=0.4223; 12hr: p=0.449, t4=0.8381). (f) Ccl2 (4hr: p=0.002, t4=7.739; 12hr: p=0.022, t4=3.632), Cxcl10 (4hr: p=0.218, t4=1.459; 12hr: p=0.020, t4=3.757), Mx1 (4hr: p=0.609, t4=0.5541; 12hr: p=0.879, t4=0.1659), Tnf (4hr: p=0.175, t4=1.648; 12hr: p=0.462, t4=0.8128), Ifnb1 (4hr: p=0.475, t4=0.7882; 12hr: p=0.864, t4=0.1830). (g) Kdm6b (4hr: p=0.005, t4=5.645; 12hr: p<0.0001, t4=15.33), and Kdm6a (4hr: p=0.524, t4=0.6979; 12hr: p=0.004, t4=5.871). Dot plots show mean ± SEM. Unpaired two-tailed t-test.

  11. Supplementary Figure 11 H3K27me3 is differentially enriched in stMg and cbMg and genetic deletion of Eed in adult microglia leads to progressive changes.

    (a) H3K27me3 enrichment at the transcriptional start site (TSS) of genes in adult cxMg is negatively correlated with cxMg gene expression. Heatmap shows (left) the abundance of H3K27me3 (MACS, n=1) ranked by log2 fold change (log2fc) of H3K27me3 ChIP over input in cxMg nuclei at the TSS ± 5 kb of individual genes; (right) log2fc (DESeq2, n=2) of mRNAs from cxMg-TRAP over its unbound fraction. p< 2.2e-16, r=-0.3241428; Pearson correlation. (b) A large number of the H3K27me3+ genes are shared between cbMg and stMg. Scatter plot shows log2fc of H3K27me3 ChIP in cbMg over its input (x-axis) versus log2fc of H3K27me3 ChIP in stMg over its input (y-axis). 7312 genes are H3K27me3+ both in cbMg and stMg (red). GO annotations (Enrichr) enriched for 7312 H3K27me3+ genes shared in stMg and cbMg. y-axis: -log10 (p-value). (c) H3K27me3 is differentially enriched in cbMg and stMg. Profile plots depict mean Reads Per million Mapped Reads of H3K27me3 in stMg and cbMg with coverage over the TSS ± 5 kb. Left: 6571 genes equally enriched in stMg and cbMg. Middle: 1382 genes with higher H3K27me3 in stMg. Right: 592 genes with higher H3K27me3 in cbMg. (d) Schematic showing the generation of EedloxP/loxP mouse and the Cre-dependent deletion of Eed. Brackets indicate the two homology arms mediating recombination in embryonic stem cells. loxP (red) and frt (yellow) sites are indicated by arrow heads. Representative genotyping gel image of Eed+/+, EedloxP/+, and EedloxP/loxP mice is shown from > 100 independent experiments. (e) Deletion of Eed results in progressive up-regulation of direct PRC2-target genes. Box-and-whisker plots show relative expression of H3K27me3+ genes up-regulated in Eed-deficient stMg, cbMg, and cxMg related to Fig. 4. stMg: control: min=-5.736, 25%=-2.92, median=-1.165, 75%=-0.506, max=1.94; 3mo: min=-3.79, 25%=-0.809, median=-0.308, 75%=0.29, max=2.83; 6mo: min=-6.51, 25%=-2.009, median=-0.33, 75%=0.086, max=1.598; and 9mo: min=-0.05, 25%=1.09, median=1.97, 75%=5.376, max=8.77; p<0.0001; F=135.8, KWS=208.1. cbMg: control: min=-6.455, 25%=-3.629, median=-0.91, 75%=-0.287, max=0.33; 3mo: min=-3.975, 25% -0.79, median=-0.28, 75%=0.376, max=2.14; 6mo: min=-5.08, 25%=-0.898, median=-0.32, 75%=0.1988, max=3.18; and 9mo: min=0.487, 25%=1.07, median=2.03, 75%=3.38, max=7.47; p<0.0001; F=86.18, KWS=139.7. cxMg: control: min=-2.945, 25%=-1.09, median=-0.737, 75%=-0.4977, max=0.0658; 3mo: min=-2.34, 25%=-0.98, median=-0.60, 75%=-0.30, max=0.839; 6mo: min=-1.98, 25%=-0.4999, median=0.117, 75%=0.519, max=1.379; and 9mo: min=0.42, 25%=0.86, median=1.59, 75%=2.96, max=5.056; p<0.0001; F=282.5, KWS=302.8. One-way ANOVA (Kruskal-Wallis test) with Dunn’s Multiple Comparison post-hoc analysis to compare all pairs of columns. Box-and-whisker plots (Tukey) show normalized expression values in z-scored log2 RPKM. **p ≤ 0.01, ***p ≤ 0.001. (f) Deletion of Eed does not lead to a loss of microglial identity or pro-inflammatory activation but leads to activation of transcription factors and phagocytic genes in stMg. Heatmaps show the variation in the expression levels (z-scored log2 RPKM) of pro-inflammatory / immediate early genes and, pan-microglia genes in control, 3 & 9mo Eed-deficient stMg / cbMg; and, phagocytosis/metabolism-related, and transcription factor-encoding genes in in control, 3 & 9mo Eed-deficient stMg (DESeq2, n=2/group). (g) Genome browser views (IGV) of selected genes in stMg TRAP from control and Cx3cr1CreErt2/+; Eef1a1LSLeGFPL10a/+;EedloxP/loxP mice. (h) Venn diagrams show overlap between 215 up-regulated (red) and 643 down-regulated (blue) genes in Eed-deficient stMg in 9mo mice with genes up- (Down: p=0.0416, OR=1.472, χ2=4.153; red; Up: p=0.007, OR=2.088, χ2=7.283); and down-(blue; Down: p < 0.0001, OR=5.365, χ2=181.9; Up: p=0.6152, OR=1.232, χ2=0.2526) regulated in Alzheimer’s Disease (AD)-associated microglia in mice12. p values and odds ratios (underneath each overlap) were calculated using the non-parametric χ2 test from a total 22,706 protein-coding genes analyzed in DESeq2. (i) Quantification of regional density and morphology of stMg Cx3cr1CreErt2/+; Eef1a1LSLeGFPL10a/+; EedloxP/loxP (mutant) and Cx3cr1CreErt2/+; Eef1a1LSLeGFPL10a/+; EedloxP/+ (control) mice. Cell body densities (control: mean=163.0, SEM=20.62; mutant: mean=127.4, SEM=10.00; p=0.1958, F=4.250, t4=1.551). Sholl intersections at 10 μm interval (control: mean=17.91, SEM=2.889; mutant: mean= 15.21, SEM=1.220; p=0.4373, F=5.605, t4=0.8619). Number of branch endings (control: mean=23.18, SEM=3.881; mutant: mean=22.90, SEM=2.362; p=0.9544, F=2.70, t4=0.060). 20-25 cells from n=3 mice/genotype. Bar graphs with individual data points show mean ± SEM. All t-tests were two-tailed unpaired.

  12. Supplementary Figure 12 Genetic deletion of Eed in adult microglia leads to global changes in the adult brain.

    (a) The number of neurons in the striatum with Eed-deficient microglia remains unchanged. Quantification of the DAPI+ cells (control: mean=223.1, SEM=14.90; mutant: mean=242.9, SEM=9.861; p=0.4079; F=3.425, t3=0.9599) and NeuN+ neurons (control: mean=67.04, SEM=12.72; mutant: mean=107.4, SEM=7.274; p=0.1020; F=4.585, t3=2.331). n=3 and 2 for control and mutant, respectively. (b) The survival of neurons in the striatum with Eed-deficient microglia remains unchanged. Quantification of the percentage of cCASP3+ neurons in control (mean=8.063, SEM=1.174) and mutant mice (mean=10.11, SEM=1.455). p=0.2997, F=1.537, t10=1.094. (c) Cx3cr1CreErt2/+;EedloxP/loxP mice display normal behavior in the open field paradigm at 3 months (n=5/genotype) and 15 months of age (n=12/genotype). Bar graphs with individual data points show total distance moved in cm (left: EedloxP/loxP control: mean=2402, SEM=174.7; mutant: mean=2181, SEM=372.0; p=0.6052, F=4.534, t8=0.5380; right: control: mean=2144, SEM=235.9; mutant: mean=2003, SEM=192.1; p=0.6479, F=1.508, t22=0.4630). (d) Cx3cr1CreErt2/+;EedloxP/loxP mice show reduced response to cocaine. Total distance traveled is shown in 5 minute bins (control: black; mutant:grey), during 30-minute habituation period and the 60 minutes following cocaine injection on day 1 (control: mean=533.0, SEM=34.97; mutant: mean=458.4, SEM=41.79; p=0.5791, F=0.3170) and day 7 (control: mean=1688, SEM=217.6; mutant: mean=997.0, SEM=191.7; p=0.0453, F=4.504). Two-way ANOVA with repeated measures (n=12 mice/genotype). (e) Eed-deficiency in cbMg does not cause a change in Purkinje cell spine density. Control: mean=2.783, SEM=0.1092; mutant: mean=2.457, SEM=0.1672; p=0.1905, F=6.247, t4=1.574 (n=3/genotype). (f) Cx3cr1CreErt2/+;EedloxP/loxP mice display normal motor learning or balance by rotarod test at 3 month (control: mean=233.2, SEM=13.00; mutant: mean=242.0, SEM=11.86; p=0.6655, F=0.19) and 15 month of age (control: mean=177.0, SEM=8.765; mutant: mean=162.2, SEM=9.176; p=0.6655, F=0.19). Two-way ANOVA with repeated measures (young: n=5/genotype; old: n=11 control, 7 mutant mice). (g) Loss of H3K27me3 in cxMg leads to a progressive gain of cbMg signature in Cx3cr1CreErt2/+(Litt); Eef1a1LSL.eGFPL10a/+; EedloxP/loxP mice. Left: Box-and-whisker plots show mean relative expression of the 27 cbMg signature genes that are up-regulated in Eed-deficient cxMg at 9 months. Control: ctrl: min=-2.657, 25%=-0.833, median=-0.68, 75%=-0.45, max=-0.30; 3mo: min=-2.46, 25%=-0.68, median=-0.439, 75%=0.055, max=0.854 ; 6mo: min=-1.68, 25%=-0.039, median=0.231, 75%=0.515, max=1.62; 9mo: min=0.588, 25%=1.014, median=1.19, 75%=2.022, max=2.788; p < 0.0001, F=68.56, KWS=75.96. Right: Heatmap shows the variation in the expression levels (z-scored log2 RPKM) of 27 cbMg-enriched genes in 3,6 & 9mo Eed-deficient cxMg and control cxMg. box-and-whisker plots (Tukey) show mean and variance for normalized expression values in z-scored log2-rpkm. Shapiro-Wilk normality p < 0.0001 for all samples. One-way ANOVA (Kruskal-Wallis test) with Dunn’s Multiple Comparison post-hoc analysis to compare all pairs of columns. **p ≤ 0.01, ***p ≤ 0.001. (h) Morphology of Eed-deficient cxMg reflect a gain of clearance phenotype. CD68+ lysosome content (red) in YFP+ microglia (green) was determined using immunofluorescence analysis from cortex of brain sections from 12-month-old control Cx3cr1CreErt2/+;EedloxP/+ and mutant Cx3cr1CreErt2/+;EedloxP/loxP mice (DAPI: blue). Scale: 10 μm. Representative image (left) and quantification of the lysosomal area / microglia area (right) is shown. Control: mean=0.04767, SEM=0.01230); mutant: mean=0.09713, SEM=0.007464; p=0.0263, F=2.716, t4=3.438 (12 images from n=3/genotype). (i) Microglia-specific Eed deficiency is associated with a reduction in spine density of cortical pyramidal cells. Representative images of Golgi-stained neuronal processes of the layer IV-V pyramidal neurons from 15-month-old control, EedloxP/loxP, and Cx3cr1CreErt2/+;EedloxP/loxP mice. Scale: 2 μm. Bar graph with individual data points show total spine densities (left) of layer IV-V cortical pyramidal neurons of control mice: mean=1.098, SEM=0.03408; and mutant mice: mean=0.8713, SEM=0.06293; p= 0.0337, F=3.411, t4=3.174); and mushroom spine densities of layer IV-V cortical pyramidal neurons of control mice: mean=0.5369, SEM=0.01514; and mutant mice: mean=0.4271, SEM=0.01488; p=0.0066, F=1.035, t4=5.177. 28 dendrites from n=3 mice / genotype. (j) The number of neurons in the cortex with Eed-deficient microglia remains unchanged. Quantification of the DAPI+ cells (control: mean=2773, SEM=193.8; mutant: mean=2844, SEM=116.5; p=0.7700; F=2.764, t4=0.3129) and NeuN+ neurons (control: mean=1133, SEM=42.70; mutant: mean=1209, SEM=44.83; p=0.2838; F=1.102, t4=1.237). n=3/genotype. Bar graphs with individual data points show mean ± SEM, line graphs show mean ± SEM, t-tests are two-tailed unpaired unless otherwise specified.

  13. Supplementary Figure 13 Young Cx3cr1CreErt2/+;EedloxP/loxP mice and aged Cx3cr1 haploinsufficient Cx3cr1CreErt2/+ mice display normal behavior.

    (a) 3-month-old Cx3cr1CreErt2/+;EedloxP/loxP mice display normal thigmotaxis in the open field. Bar graphs with individual data points show that mutant mice spend equal time in the center vs periphery (mean=0.1372, SEM=0.02633) compared with control mice (mean=0.2046, SEM=0.06382) p=0.3764, F=5.876, t5=0.9704, n=6/genotype. (b) Left: 3-month-old Cx3cr1CreErt2/+;EedloxP/loxP mice display normal behavior in the elevated plus maze at young age. 3-month-old mice spend equal time in the open arms (left: control: mean=20.88, SEM=13.27; mutant: mean=35.83, SEM=35.57; p=0.6592, F=2.789, t8=0.4579) and in closed arms (middle: control: mean=171.6, SEM=35.03; mutant: mean=171.0, SEM=16.88; p=0.9897, F=6.459, t8=0.01330); and travel equal distances (right: control: mean=1564, SEM=212.7; mutant: mean=2092, SEM=210.9; p=0.1311, F=1.527, t8=1.682). n=4 control, 5 mutant mice. Right: 15-month-old Cx3cr1CreErt2/+;EedloxP/loxP mice travel equal distances in the elevated plus maze paradigm (control: mean=1758, SEM=163.8; mutant: mean=1430, SEM= 201.8) p=0.2210, F=1.365, t17=1.271; n=10/genotype. (c) Cx3cr1 haploinsufficiency in 15-month-old Cx3cr1CreErt2/+ mice does not affect normal behavior in the elevated plus maze. Mutant and control mice spend equal time in the open arms (left: control: mean=22.17, SEM=4.888; mutant: mean=21.83, SEM=7.682; p=0.9713, F=2.778, t15=0.03657) and in closed arms (middle: control: mean=178.6, SEM=10.98; mutant: mean=171.7, SEM=6.651; p=0.5910, F=2.422, t15=0.5491); and travel equal distances (right: control: mean=1845, SEM=(212.7; mutant: mean=1683, SEM=210.9; p=0.1250, F=1.238, t15=1.625; n=8 control, 9 mutant mice). (d) 15-month-old Cx3cr1CreErt2/+;EedloxP/loxP mice show a deficit in cued memory in the fear conditioning paradigm. Bar graphs with individual data points show that control and mutant mice freeze equal amounts at baseline (control: mean=2.831, SEM=1.131; mutant: mean=8.024, SEM=3.445; p=0.1820, F=1.384, MWU=31.50; Shapiro Wilk p-value (SWp)=0.0039 & 0.0010), in training (control: mean=40.89, SEM=8.172; mutant: mean=67.78, SEM=10.04; p=0.0500, F=1.372, t19=2.093), in the old context (control: mean=36.59, SEM=6.040; mutant: mean=36.48, SEM= 7.900; p=0.9918, F=1.555, t19=0.01041), and in the new context (control: mean=16.00, SEM=4.083; mutant: mean=18.03, SEM=5.824; p=0.7786; F=2.035, t18=0.2854). However, mutant mice freeze less (mean=23.17, SEM=9.154) than control mice (mean=46.24, SEM=7.778) after the tone cue (p=0.0448, F=1.259, MWU=26.00; SWp=0.0111 for mutant, t19=1.932). n=11 control, 10 mutant mice. (e) 15-month-old Cx3cr1CreErt2/+;EedloxP/loxP mice show a deficit in novel object recognition test. Bar graphs with individual data points show that mutant mice spend equal time (object 1: mean=1.742, SEM=1.181; object 2 mean=2.916, SEM=1.113; p=0.3896, F= 1.127, t9=0.9039) sniffing two identical objects during training as 15-month-old control EedloxP/loxP mice (object 1: mean=1.655, SEM=0.5399; object 2 mean=2.856, SEM=0.5316; p=0.0610, F=1.032, t9=2.141) and 15-month-old Cx3cr1 haploinsufficient Cx3cr1CreErt2/+ mice (object 1: mean=3.329, SEM=0.7976; object 2: mean=5.314, SEM=1.218; p=0.0660, F=2.071, t7=2.176). However, control mice spent more time sniffing a novel object (novel: mean=4.338, SEM=1.485; familiar: mean=1.421, SEM= 0.4844; p=0.0222, F=9.396, t9=2.758), as do Cx3cr1 haploinsufficient mice (novel: mean=6.521, SEM=1.094; familiar: mean=2.812, SEM=0.8749; p=0.0137, F=2.823, t8=3.146); whereas, mutant mice (right) do not (novel: mean=2.763, SEM=1.126; familiar: mean=2.176, SEM=0.6447; p=0.4640, F=3.049, t9=0.7648). Two-tailed paired t-tests (n=9 for Cx3cr1CreErt2/+, n=10 for EedloxP/loxP and Cx3cr1CreErt2/+;EedloxP/loxP). Grubbs’ Test was used to identify and exclude an outlier. (f) Cx3cr1 haploinsufficiency is not associated with the development of seizures in Cx3cr1CreErt2/+ mice Kaplan-Meier curve for n=9 mice/genotype; p=0.8491; χ2=0.03618 from log-rank Mantel–Cox test). (g) Cx3cr1CreErt2/+;EedloxP/loxP mice show a normal survival rate compared to control mice. Kaplan-Meier curve for n=28 mice/genotype; p=0.8491; χ2=0.03618 from log-rank Mantel–Cox test). (h) 12-month-old microglia-depleted (PLX) mice show no anxiety-like phenotype in the open field. PLX mice spend equal time in the center vs periphery (mean=0.3400, SEM=0.06352) compared with control mice (mean=0.3973, SEM=0.06133) p=0.4605, F=1.515, t19=0.6165; n=8 PLX-treated / 13 control mice. (i) 12-month-old PLX mice show no anxiety-like phenotype in the elevated plus maze. PLX mice spend equal time in the open arms (mean=22.65, SEM=7.520) compared with control mice (mean=19.83, SEM=6.603) p=0.7814, F=1.297, t18=0.2817; and in closed arms (control: mean=179.6, SEM=12.24; PLX: mean=182.3, SEM=6.801) p=0.8532, F=3.241, t18=0.1877; and travel equal distances (control: mean=2490, SEM=104.5; PLX: mean=2689, SEM=77.91; p=0.1448, F=1.798, t18=1.525; n=10/group. (j) PLX mice are not prone to development of a seizure phenotype. Kaplan-Meier curve (n=7/group; p=1.00; χ2=0.00 (log-rank Mantel–Cox test). Bar graphs with individual data points show mean ± SEM. All t-tests were two-tailed unpaired.

  14. Supplementary Figure 14 Comparison of cbMg gene expression patterns with published data sets.

    (a) Left: Venn diagrams show overlap between cbMg (orange) and stMg (purple) enriched genes by TRAP with cbMg-enriched genes (red; stMg: p < 0.0001, OR=0.2023, χ2=24.68; cbMg: p < 0.0001, OR=3.752, χ2=77.82) and stMg-enriched genes (blue; stMg: p < 0.0001, 4.206, χ2=154.7; cbMg: p=0.7791, OR=1.095, χ2=0.07869) from Grabert et. al41. Right: Venn diagrams show overlap between cbMg (orange) and stMg (purple) enriched genes by single nuclei sequencing with cbMg-enriched genes (red; stMg: p < 0.0001, OR=0.5590, χ2=22.71; cbMg: p=0.0852, OR=1.174, χ2=2.963) and stMg-enriched genes (blue; stMg: p < 0.0001, 2.148, χ2=63.39; cbMg: p=0.1205, OR=1.198, χ2=2.410) from Grabert et. al41. (b) Horizontal bar graphs show log2 (fold change) in the counts of selected cbMg-specific/clearance-related (orange) and stMg-enriched/surveillance-related (purple) transcription factors in (1) expression upon apoptotic cell exposure, (2) enrichment of H3K27me3 in stMg vs cbMg, and (3) expression in Eed-deficient stMg vs control stMg. (c) Venn diagrams show overlap between cbMg (orange) and stMg (purple) enriched genes with genes up-(red; stMg: p=0.906, OR=1.089, χ2=0.01393; cbMg: p < 0.0001, OR=12.63, χ2=71.70) or down-regulated (blue; stMg: p < 0.0001, OR=19.42, χ2=950.8; cbMg: p=0.406, OR=0.558, χ2=0.6903) in microglia upon injection of apoptotic cells from Krasemann et. al13. Selected genes are named. (d) Venn diagram shows overlap between 215 genes up-regulated in PRC2-deficient stMg (purple), 144 genes up-regulated in PRC2-deficient cbMg (orange), and 518 genes up-regulated in PRC2-deficient cxMg (blue). Overlapping genes are indicated. p values and odds ratios (underneath each overlap) were calculated using the non-parametric χ2 test from a total 22,706 protein-coding genes analyzed in DESeq2.

  15. Supplementary Figure 15 Original western blots.

    Cropped KODAK films or ChemiDoc images for Western blots in Fig. 2 and 5 are shown. Membranes were cut prior to antibody staining to allow for simultaneous detection of proteins running at different sizes on the same membrane.

Supplementary information

  1. Supplementary Text and Figures

    Supplementary Figures 1–15

  2. Reporting Summary

  3. Supplementary Table 1

    Astrocyte TRAP vs. microglia TRAP

  4. Supplementary Table 2

    Microglia isolation by FACS vs. microglia TRAP

  5. Supplementary Table 3

    cbMg vs. stMg by TRAP

  6. Supplementary Table 4

    Overlaps between region-specific microglia genes and published datasets

  7. Supplementary Table 5

    Single microglia nuclei gene expression matrix

  8. Supplementary Table 6

    cbMg vs. stMg by single nuclei RNA sequencing

  9. Supplementary Table 7

    H3K27me3 in cbMg, stMg, and cxMg

  10. Supplementary Table 8

    Dysregulated genes in Eed-deficient stMg

  11. Supplementary Table 9

    Dysregulated genes in Eed-deficient cbMg

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/s41593-018-0192-3