Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Chd4 choreographs self-antigen expression for central immune tolerance

Abstract

Autoreactive T cells are eliminated in the thymus to prevent autoimmunity by promiscuous expression of tissue-restricted self-antigens in medullary thymic epithelial cells. This expression is dependent on the transcription factor Fezf2, as well as the transcriptional regulator Aire, but the entire picture of the transcriptional program has been obscure. Here, we found that the chromatin remodeler Chd4, also called Mi-2β, plays a key role in the self-antigen expression in medullary thymic epithelial cells. To maximize the diversity of self-antigen expression, Fezf2 and Aire utilized completely distinct transcriptional mechanisms, both of which were under the control of Chd4. Chd4 organized the promoter regions of Fezf2-dependent genes, while contributing to the Aire-mediated induction of self-antigens via super-enhancers. Mice deficient in Chd4 specifically in thymic epithelial cells exhibited autoimmune phenotypes, including T cell infiltration. Thus, Chd4 plays a critical role in integrating Fezf2- and Aire-mediated gene induction to establish central immune tolerance.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Fig. 1: Distinct features of Fezf2-induced TRA genes expressed by mTECs.
Fig. 2: Distinct expression pattern of Aire- and Fezf2-induced genes in mTECs.
Fig. 3: Identification of a Fezf2-interacting partner Chd4.
Fig. 4: Chd4 and Fezf2 cooperatively regulate chromatin structure and transcription.
Fig. 5: Chd4 controls chromatin accessibility of super-enhancers in mTECs.
Fig. 6: Autoimmune phenotypes in Chd4 cKO mice.

Data availability

The ATAC-Seq, ChIP-Seq and RNA-Seq datasets reported in this manuscript can be accessed in the Gene Expression Omnibus with the accession code GSE144880. Specifically, data on analyses of mTECs are available from accession codes GSE144877 (RNA-Seq), GSE144878 (ChIP-Seq) and GSE144879 (ATAC-Seq). The other referenced publically available datasets are: single-cell RNA-Seq and H3K4me3 and H3K27me3 ChIP-Seq for mTECs (GSE53111); H3K27ac, H3K4me1 and Aire ChIP-Seq in mTECs and ATAC-Seq for wild-type or Aire KO mTECs (GSE92597 and GSE92594); H3K4me2 ChIP-Seq in mTEC-II (GSE103969); H3K4ac and H3K9ac ChIP-Seq in mTECs (GSE114713); and a microarray dataset for various tissues and cells (GSE10246). The microarray dataset for the I-BET 151-treated wild-type or Aire KO mTECs was provided by H. Yoshida and D. Mathis37. Other data that support the findings of this study are available from the corresponding author upon request. Source data are provided with this paper.

Code availability

Scripts needed to reconstruct analysis files will be made available on request.

References

  1. 1.

    Zinkernagel, R. M. & Doherty, P. C. Restriction of in vitro T cell-mediated cytotoxicity in lymphocytic choriomeningitis within a syngeneic or semiallogeneic system. Nature 248, 701–702 (1974).

    PubMed  CAS  Google Scholar 

  2. 2.

    Hogquist, K. A., Baldwin, T. A. & Jameson, S. C. Central tolerance: learning self-control in the thymus. Nat. Rev. Immunol. 5, 772–782 (2005).

    PubMed  PubMed Central  CAS  Google Scholar 

  3. 3.

    Davis, M. M. et al. Ligand recognition by αβ T cell receptors. Annu Rev. Immunol. 16, 523–544 (1998).

    PubMed  CAS  Google Scholar 

  4. 4.

    Cheng, M. & Anderson, M. S. Thymic tolerance as a key brake on autoimmunity. Nat. Immunol. 19, 659–664 (2018).

    PubMed  PubMed Central  CAS  Google Scholar 

  5. 5.

    Goodnow, C. C., Sprent, J., Fazekas de St Groth, B. & Vinuesa, C. G. Cellular and genetic mechanisms of self tolerance and autoimmunity. Nature 435, 590–597 (2005).

    PubMed  CAS  Google Scholar 

  6. 6.

    Klein, L., Kyewski, B., Allen, P. M. & Hogquist, K. A. Positive and negative selection of the T cell repertoire: what thymocytes see (and don’t see). Nat. Rev. Immunol. 14, 377–391 (2014).

    PubMed  PubMed Central  CAS  Google Scholar 

  7. 7.

    Aschenbrenner, K. et al. Selection of Foxp3+ regulatory T cells specific for self antigen expressed and presented by Aire+ medullary thymic epithelial cells. Nat. Immunol. 8, 351–358 (2007).

    PubMed  CAS  Google Scholar 

  8. 8.

    Derbinski, J., Schulte, A., Kyewski, B. & Klein, L. Promiscuous gene expression in medullary thymic epithelial cells mirrors the peripheral self. Nat. Immunol. 2, 1032–1039 (2001).

    PubMed  CAS  Google Scholar 

  9. 9.

    Liston, A., Lesage, S., Wilson, J., Peltonen, L. & Goodnow, C. C. Aire regulates negative selection of organ-specific T cells. Nat. Immunol. 4, 350–354 (2003).

    PubMed  CAS  Google Scholar 

  10. 10.

    Rattay, K., Meyer, H. V., Herrmann, C., Brors, B. & Kyewski, B. Evolutionary conserved gene co-expression drives generation of self-antigen diversity in medullary thymic epithelial cells. J. Autoimmun. 67, 65–75 (2016).

    PubMed  CAS  Google Scholar 

  11. 11.

    The Finnish–German APECED Consortium. An autoimmune disease, APECED, caused by mutations in a novel gene featuring two PHD-type zinc-finger domains. Nat. Genet. 17, 399–403 (1997).

  12. 12.

    Nagamine, K. et al. Positional cloning of the APECED gene. Nat. Genet. 17, 393–398 (1997).

    PubMed  CAS  Google Scholar 

  13. 13.

    Giraud, M. et al. An IRF8-binding promoter variant and AIRE control CHRNA1 promiscuous expression in thymus. Nature 448, 934–937 (2007).

    PubMed  CAS  Google Scholar 

  14. 14.

    Colobran, R. et al. Association of an SNP with intrathymic transcription of TSHR and Graves’ disease: a role for defective thymic tolerance. Hum. Mol. Genet. 20, 3415–3423 (2011).

    PubMed  CAS  Google Scholar 

  15. 15.

    Gimenez-Barcons, M. et al. Autoimmune predisposition in Down syndrome may result from a partial central tolerance failure due to insufficient intrathymic expression of AIRE and peripheral antigens. J. Immunol. 193, 3872–3879 (2014).

    PubMed  CAS  Google Scholar 

  16. 16.

    Anderson, M. S. et al. Projection of an immunological self shadow within the thymus by the Aire protein. Science 298, 1395–1401 (2002).

    CAS  Google Scholar 

  17. 17.

    Mathis, D. & Benoist, C. Aire. Annu Rev. Immunol. 27, 287–312 (2009).

    PubMed  CAS  Google Scholar 

  18. 18.

    Takaba, H. et al. Fezf2 orchestrates a thymic program of self-antigen expression for immune tolerance. Cell 163, 975–987 (2015).

    PubMed  CAS  Google Scholar 

  19. 19.

    Schug, J. et al. Promoter features related to tissue specificity as measured by Shannon entropy. Genome Biol. 6, R33 (2005).

    PubMed  PubMed Central  Google Scholar 

  20. 20.

    Hassan, R. & Ho, M. Mesothelin targeted cancer immunotherapy. Eur. J. Cancer 44, 46–53 (2008).

    PubMed  CAS  Google Scholar 

  21. 21.

    Michaelidou, K., Tzovaras, A., Missitzis, I., Ardavanis, A. & Scorilas, A. The expression of the CEACAM19 gene, a novel member of the CEA family, is associated with breast cancer progression. Int. J. Oncol. 42, 1770–1777 (2013).

    PubMed  CAS  Google Scholar 

  22. 22.

    Sansom, S. N. et al. Population and single-cell genomics reveal the Aire dependency, relief from Polycomb silencing, and distribution of self-antigen expression in thymic epithelia. Genome Res. 24, 1918–1931 (2014).

    PubMed  PubMed Central  CAS  Google Scholar 

  23. 23.

    Brennecke, P. et al. Single-cell transcriptome analysis reveals coordinated ectopic gene-expression patterns in medullary thymic epithelial cells. Nat. Immunol. 16, 933–941 (2015).

    PubMed  PubMed Central  CAS  Google Scholar 

  24. 24.

    Meredith, M., Zemmour, D., Mathis, D. & Benoist, C. Aire controls gene expression in the thymic epithelium with ordered stochasticity. Nat. Immunol. 16, 942–949 (2015).

    PubMed  PubMed Central  CAS  Google Scholar 

  25. 25.

    Giraud, M. et al. An RNAi screen for Aire cofactors reveals a role for Hnrnpl in polymerase release and Aire-activated ectopic transcription. Proc. Natl Acad. Sci. USA 111, 1491–1496 (2014).

    PubMed  CAS  Google Scholar 

  26. 26.

    Abramson, J., Giraud, M., Benoist, C. & Mathis, D. Aire’s partners in the molecular control of immunological tolerance. Cell 140, 123–135 (2010).

    PubMed  CAS  Google Scholar 

  27. 27.

    Tong, J. K., Hassig, C. A., Schnitzler, G. R., Kingston, R. E. & Schreiber, S. L. Chromatin deacetylation by an ATP-dependent nucleosome remodelling complex. Nature 395, 917–921 (1998).

    PubMed  CAS  Google Scholar 

  28. 28.

    Low, J. K. K. et al. CHD4 is a peripheral component of the nucleosome remodeling and deacetylase complex. J. Biol. Chem. 291, 15853–15866 (2016).

    PubMed  PubMed Central  CAS  Google Scholar 

  29. 29.

    Stevens, T. J. et al. 3D structures of individual mammalian genomes studied by single-cell Hi-C. Nature 544, 59–64 (2017).

    PubMed  PubMed Central  CAS  Google Scholar 

  30. 30.

    de Dieuleveult, M. et al. Genome-wide nucleosome specificity and function of chromatin remodellers in ES cells. Nature 530, 113–116 (2016).

    PubMed  PubMed Central  CAS  Google Scholar 

  31. 31.

    Yoshida, T. et al. Chromatin restriction by the nucleosome remodeler Mi-2β and functional interplay with lineage-specific transcription regulators control B-cell differentiation. Genes Dev. 33, 763–781 (2019).

    PubMed  PubMed Central  CAS  Google Scholar 

  32. 32.

    Zhang, J. et al. Harnessing of the nucleosome-remodeling-deacetylase complex controls lymphocyte development and prevents leukemogenesis. Nat. Immunol. 13, 86–94 (2011).

    PubMed  PubMed Central  Google Scholar 

  33. 33.

    Hosokawa, H. et al. Functionally distinct Gata3/Chd4 complexes coordinately establish T helper 2 (Th2) cell identity. Proc. Natl Acad. Sci. USA 110, 4691–4696 (2013).

    PubMed  CAS  Google Scholar 

  34. 34.

    Waterfield, M. et al. The transcriptional regulator Aire coopts the repressive ATF7ip–MBD1 complex for the induction of immunotolerance. Nat. Immunol. 15, 258–265 (2014).

    PubMed  PubMed Central  CAS  Google Scholar 

  35. 35.

    Bansal, K., Yoshida, H., Benoist, C. & Mathis, D. The transcriptional regulator Aire binds to and activates super-enhancers. Nat. Immunol. 18, 263–273 (2017).

    PubMed  PubMed Central  CAS  Google Scholar 

  36. 36.

    Hnisz, D. et al. Super-enhancers in the control of cell identity and disease. Cell 155, 934–947 (2013).

    CAS  Google Scholar 

  37. 37.

    Yoshida, H. et al. Brd4 bridges the transcriptional regulators, Aire and P-TEFb, to promote elongation of peripheral-tissue antigen transcripts in thymic stromal cells. Proc. Natl Acad. Sci. USA 112, E4448–E4457 (2015).

    PubMed  CAS  Google Scholar 

  38. 38.

    Dowen, J. M. et al. Control of cell identity genes occurs in insulated neighborhoods in mammalian chromosomes. Cell 159, 374–387 (2014).

    PubMed  PubMed Central  CAS  Google Scholar 

  39. 39.

    Lodato, S. et al. Gene co-regulation by Fezf2 selects neurotransmitter identity and connectivity of corticospinal neurons. Nat. Neurosci. 17, 1046–1054 (2014).

    PubMed  PubMed Central  CAS  Google Scholar 

  40. 40.

    Seelig, H. P. et al. The major dermatomyositis-specific Mi-2 autoantigen is a presumed helicase involved in transcriptional activation. Arthritis Rheum. 38, 1389–1399 (1995).

    PubMed  CAS  Google Scholar 

  41. 41.

    Minnich, M. et al. Multifunctional role of the transcription factor Blimp-1 in coordinating plasma cell differentiation. Nat. Immunol. 17, 331–343 (2016).

    PubMed  PubMed Central  CAS  Google Scholar 

  42. 42.

    Hoffmeister, H. et al. CHD3 and CHD4 form distinct NuRD complexes with different yet overlapping functionality. Nucleic Acids Res. 45, 10534–10554 (2017).

    PubMed  PubMed Central  CAS  Google Scholar 

  43. 43.

    Torrado, M. et al. Refinement of the subunit interaction network within the nucleosome remodelling and deacetylase (NuRD) complex. FEBS J. 284, 4216–4232 (2017).

    PubMed  PubMed Central  CAS  Google Scholar 

  44. 44.

    Suzuki, H. I., Young, R. A. & Sharp, P. A. Super-enhancer-mediated RNA processing revealed by integrative microRNA network analysis. Cell 168, 1000–1014.e15 (2017).

    PubMed  PubMed Central  CAS  Google Scholar 

  45. 45.

    Lomada, D. et al. Stat3 signaling promotes survival and maintenance of medullary thymic epithelial cells. PLoS Genet. 12, e1005777 (2016).

    PubMed  PubMed Central  Google Scholar 

  46. 46.

    Lewis, M. J. et al. Autoantibodies targeting TLR and SMAD pathways define new subgroups in systemic lupus erythematosus. J. Autoimmun. 91, 1–12 (2018).

    PubMed  CAS  Google Scholar 

  47. 47.

    Clement, C. C. et al. Autoimmune response to transthyretin in juvenile idiopathic arthritis. JCI Insight 1, 85633 (2016).

    PubMed  Google Scholar 

  48. 48.

    Tzartos, J. S. et al. Antibodies to aquaporins are frequent in patients with primary Sjögren’s syndrome. Rheumatology 56, 2114–2122 (2017).

    PubMed  CAS  Google Scholar 

  49. 49.

    Sjowall, C., Eriksson, P., Almer, S. & Skogh, T. Autoantibodies to C-reactive protein is a common finding in SLE, but not in primary Sjogren’s syndrome, rheumatoid arthritis or inflammatory bowel disease. J. Autoimmun. 19, 155–160 (2002).

    PubMed  Google Scholar 

  50. 50.

    Rapoport, B., Chazenbalk, G. D., Jaume, J. C. & McLachlan, S. M. The thyrotropin (TSH) receptor: interaction with TSH and autoantibodies. Endocr. Rev. 19, 673–716 (1998).

    PubMed  CAS  Google Scholar 

  51. 51.

    Han, W. et al. TBR1 directly represses Fezf2 to control the laminar origin and development of the corticospinal tract. Proc. Natl Acad. Sci. USA 108, 3041–3046 (2011).

    PubMed  CAS  Google Scholar 

  52. 52.

    O’Shaughnessy-Kirwan, A., Signolet, J., Costello, I., Gharbi, S. & Hendrich, B. Constraint of gene expression by the chromatin remodelling protein CHD4 facilitates lineage specification. Development 142, 2586–2597 (2015).

    PubMed  PubMed Central  Google Scholar 

  53. 53.

    Kasai, M. et al. Difference in antigen presentation pathways between cortical and medullary thymic epithelial cells. Eur. J. Immunol. 26, 2101–2107 (1996).

    PubMed  CAS  Google Scholar 

  54. 54.

    Nitta, T. et al. The thymic cortical epithelium determines the TCR repertoire of IL-17-producing γδT cells. EMBO Rep. 16, 638–653 (2015).

    PubMed  PubMed Central  CAS  Google Scholar 

  55. 55.

    Hayashi, M. et al. Osteoprotection by semaphorin 3A. Nature 485, 69–74 (2012).

    PubMed  CAS  Google Scholar 

  56. 56.

    Komatsu, N. et al. Pathogenic conversion of Foxp3+ T cells into TH17 cells in autoimmune arthritis. Nat. Med. 20, 62–68 (2014).

    PubMed  CAS  Google Scholar 

  57. 57.

    Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).

    PubMed  PubMed Central  CAS  Google Scholar 

  58. 58.

    Langmead, B., Wilks, C., Antonescu, V. & Charles, R. Scaling read aligners to hundreds of threads on general-purpose processors. Bioinformatics 35, 421–432 (2018).

    PubMed Central  Google Scholar 

  59. 59.

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

    PubMed  PubMed Central  Google Scholar 

  60. 60.

    Ramírez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res. 44, W160–W165 (2016).

    PubMed  PubMed Central  Google Scholar 

  61. 61.

    Liao, Y., Smyth, G. K. & Shi, W. The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote. Nucleic Acids Res. 41, e108 (2013).

    PubMed  PubMed Central  Google Scholar 

  62. 62.

    Bodenhofer, U., Kothmeier, A. & Hochreiter, S. APCluster: an R package for affinity propagation clustering. Bioinformatics 27, 2463–2464 (2011).

    PubMed  CAS  Google Scholar 

  63. 63.

    Krzywinski, M. et al. Circos: an information aesthetic for comparative genomics. Genome Res. 19, 1639–1645 (2009).

    PubMed  PubMed Central  CAS  Google Scholar 

  64. 64.

    Robinson, J. T. et al. Integrative genomics viewer. Nat. Biotechnol. 29, 24–26 (2011).

    PubMed  PubMed Central  CAS  Google Scholar 

  65. 65.

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

    PubMed  PubMed Central  Google Scholar 

  66. 66.

    Heinz, S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38, 576–589 (2010).

    PubMed  PubMed Central  CAS  Google Scholar 

  67. 67.

    Žuklys, S. et al. Foxn1 regulates key target genes essential for T cell development in postnatal thymic epithelial cells. Nat. Immunol. 17, 1206–1215 (2016).

    PubMed  PubMed Central  Google Scholar 

  68. 68.

    Lenda, D. M., Stanley, E. R. & Kelley, V. R. Negative role of colony-stimulating factor-1 in macrophage, T cell, and B cell mediated autoimmune disease in MRL-Fas lpr mice. J. Immunol. 173, 4744–4754 (2004).

    PubMed  CAS  Google Scholar 

  69. 69.

    Koh, A. S. et al. Rapid chromatin repression by Aire provides precise control of immune tolerance. Nat. Immunol. 19, 162–172 (2018).

    PubMed  PubMed Central  CAS  Google Scholar 

  70. 70.

    Akaike, H. Information Theory and an Extension of the Maximum Likelihood Principle (eds Petrov, B. N. & Caski, F.) 267–281 (Akadimiai Kiado, 1973).

Download references

Acknowledgements

We thank P. A. Sharp, Y. Kurikawa and L. Danks for insightful and valuable comments, D. Mathis and H. Yoshida for providing information on super-enhancers and the microarray data on Brd4, B. Hendrich for providing Chd4 floxed mice, B. Malissen for providing Foxp3-ires-GFP mice, T. Tsubokawa for technical assistance, and other laboratory members for helpful discussions. This work was supported by a Takeda Visionary Research Grant to H. Takaba from the Takeda Science Foundation, a grant to H. Takaba from the Astellas Foundation for Research on Metabolic Disorders, a grant for Challenging Exploratory Research (26670231), a Challenging Research (Exploratory) Grant (17K19545), a Young Scientists A Grant (16H06228) to H. Takaba, and a Specially Promoted Research Grant (15H05703) to H. Takayanagi from MEXT/JSPS of Japan. This work was also supported by AMED-CREST (20gm1210008) to H. Takayanagi. Studies conducted by H.I.S. at the Massachusetts Institute of Technology were supported by United States Public Health Service grants R01-GM034277 and R01-CA133404 to P. A. Sharp and P01-CA042063 to T. Jacks from the NIH, as well as a Koch Institute Support (core) grant (P30-CA14051) from the National Cancer Institute, and supported in part by an agreement between the Whitehead Institute for Biomedical Research and Novo Nordisk.

Author information

Affiliations

Authors

Contributions

Y.T. and H. Takaba designed the experiments, performed most of the experiments and bioinformatics analyses, and wrote the manuscript. R.B., C.D.P.M., Y.A., T.O. and Y.M. performed the experiments using mice. H.I.S., A.T. and T.K. performed the next-generation sequencing and bioinformatics analyses. H. Takayanagi supervised the project and wrote the manuscript.

Corresponding author

Correspondence to Hiroshi Takayanagi.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information L. A. Dempsey was the primary editor on this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.

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

Extended data

Extended Data Fig. 1 Fezf2-induced genes have distinct features.

a, Venn diagram displaying number of Fezf2-induced genes and Aire-induced genes (fold-change > 2, q < 0.1). RNA-seq was performed for Fezf2 or Aire knockout (KO) (n=3 biologically independent replicates for each genotype) and the wild-type (WT) (control for Fezf2 KO n=3, control for Aire KO n=3 biologically independent replicates) mTECs. P-values were calculated with DESeq2 algorithm (two-tailed Wald test), and p-values were corrected for multiple testing using the Benjamini-Hochberg method. b, Density plot of the entropy score for all genes. Genes with a low entropy score (< 4.5) were defined as TRAs. c, Entropy score of representative Fezf2-induced TRA genes, Aire-induced TRA genes and house-keeping genes. d, Expression of Fabp9 (Fezf2-induced TRA), Mup4 (Aire-induced TRA) and Gapdh (House-keeping genes) in peripheral tissues and cells (GSE10246). e, Density plot of the log2 expression intensity (mean read count per gene in cells expressing the gene) for Fezf2-induced genes (n=623) versus Aire-induced genes (n=1521) at the single-cell level. p = 4.0e-4 (two tailed Wilcoxon rank-sum test). f, Boxplot of the number of Fezf2-induced or Aire-induced genes in single mTECs (n=174). Boxplot indicating median, boxed interquartile range, whiskers extending to the most extreme points within the range between (lower quantile - 1.5 IQR) and (lower quantile + 1.5 IQR), and outliers beyond the range of whiskers are indicated as dots. p = 9.5e-3 (two tailed Wilcoxon rank-sum test). g-j, Result of the affinity-propagation clustering of the expression matrix for randomly selected genes whose number was matched to Aire-induced genes (n=1521) (g) or Fezf2-induced genes (n=623) (i) in single mTECs. Heatmaps of the inter-gene Pearson correlation which are control for Fig. 2b (g) and Fig. 2d (i) are shown. Net similarity calculated from the result of 1000 times affinity-propagation clustering of the expression matrix for randomly selected genes whose number was matched to Aire-induced genes (h) or Fezf2-induced genes (j) are shown as histogram. The arrow indicates the net similarity calculated from the result of affinity-propagation clustering of the expression matrix for Aire-induced genes (h) or Fezf2-induced genes. k, Plot of the tag density for H3K4me3 in mTECs around the TSS of the Fezf2-induced TRA genes (n=159) versus Aire-induced TRA genes (n=541). Average tag counts are presented as lines and standard errors are presented as shades of line. Boxplot depicting the tag density of H3K4me3 around the TSS of Fezf2-induced TRA genes and Aire-induced TRA genes. Boxplot indicating median, boxed interquartile range, whiskers extending to the most extreme points within the range between (lower quantile - 1.5 IQR) and (lower quantile + 1.5 IQR), and outliers beyond the range of whiskers are indicated as dots. p = 3.4e-13 (two tailed Wilcoxon rank-sum test). l, Plot of tag density for H3K27me3 in mTECs around the TSS of Fezf2-induced TRA genes (n=159) versus Aire-induced TRA genes (n=541). Average tag counts are presented as lines and standard errors are presented as shades of line. Boxplot depicting the tag density of H3K27me3 around the TSS of Fezf2-induced TRA genes and Aire-induced TRA genes. Boxplot indicating median, boxed interquartile range, whiskers extending to the most extreme points within the range between (lower quantile - 1.5 IQR) and (lower quantile + 1.5 IQR), and outliers beyond the range of whiskers are indicated as dots. p = 8.8e-8 (two tailed Wilcoxon rank-sum test). m, Average plot (top) and heatmap (bottom) of ChIP-seq reads for H3K4ac, H3K4me2, H3K9ac and H3K27ac in mTECs around the TSS of Fezf2-induced genes (n=640) versus Aire-induced genes (n=1553). Average tag counts are presented as lines and standard errors are presented as shades of line. n, Average plot (top) and heatmap (bottom) of ChIP-seq reads for H3K4ac, H3K4me2, H3K9ac, H3K27ac and H3K4me1 in mTECs around the enhancer of mTECs whose nearest genes were Fezf2-induced genes (n=814) or Aire-induced genes (n=1166). Average tag counts are presented as lines and standard errors are presented as shades of line. o,p, Plot of the ATAC-seq tag density in mTECs for Fezf2-induced genes (n=640) versus Aire-induced genes (n=1553) (o) or Fezf2-induced TRA genes (n=159) versus Aire-induced TRA genes (n=541) (p).

Extended Data Fig. 2 Construction of 3xFlag-Fezf2 knock-in mice.

a, Thymus sections from WT mice at 3 weeks old stained with anti-Chd4, Fezf2 or Aire antibodies and with DAPI (blue). Scale bars indicate 100 μm. Representative data from two independent experiments is indicated. b, Schematic diagram of 3xFlag-Fezf2 knock-in (KI) mice strategy. The Fezf2 genomic locus following homologous recombination after cleavage with CRISPR/Cas9. Fezf2 exons are represented in grey boxes. Guided RNA (gRNA) was designed to cleave after the first base pair of the start codon of Fezf2. An oligo DNA sequence with 45bp with complementary sites was designed to introduce a triple epitope Flag peptide followed with a serine-4xglycine- serine linker into exon 1 of Fezf2. c, FACS profiles for EpCAM and CD45 (top) or Ly51 and UEA-1 of EpCAM+CD45-TECs (bottom) derived from 3xFlag-Fezf2 KI mice or control mice at 6 weeks old. (n=6 biologically independent samples per each genotype). d, The number of TEC (EpCAM+CD45-), mTEC (UEA-1+Ly51-TEC) or cTEC (UEA-1-Ly51+TEC). P-values were calculated by two-tailed unpaired t-test. Center line indicates mean. Error bars, SEM. e, Immunohistochemistry of WT and KI mice staining for Fezf2, Aire, cytokeratin 5 (Krt5), and DAPI (left) or Flag, Aire, Krt5 and DAPI (right). Independent experiments were repeated twice (n=2, 3 biologically independent samples for each experiment). Scale bar, 100μm.

Extended Data Fig. 3 Analysis of TEC and T cell development in the thymus of Chd4 cKO mice.

a, FACS profiles for CD4 and CD8 in the total thymocytes of the control and Chd4 cKO mice at 4-weeks old (left panels). The frequency and number of CD4 single positive (CD4SP), CD8 single positive (CD8SP), CD4 and CD8 double positive (DP) thymocyte populations are indicated (right panels). (n=6, 5 biologically independent replicates for control and Chd4 cKO mice). b, TCRβ and CD69 expression profiles of DP thymocytes from the control and Chd4 cKO mice at 4-weeks old (left panels). The frequency and number of TCRβ+CD69+ cells in the DP thymocytes are indicated (right panels) (n = 6, 5 for control and Chd4 cKO mice). c, FACS profiles for CD25 and Foxp3 in CD4SP thymocytes from control and Chd4 cKO mice at 4-weeks old (left panels). The frequency and number of CD25+Foxp3+ cells in CD4SP thymocytes are indicated (right panels) (n = 3 biologically independent replicates per each genotypes). d, Structure of the thymus in Chd4 cKO mice or control mice at 4-weeks old (H&E staining). C; cortex, and M; medulla. Scale bars indicate 1mm (top) or 100μm (bottom). Independent experiments were repeated twice (n=1, 2 biologically independent samples for each experiment). e, Thymus sections from Chd4 cKO mice or control mice at 4-weeks old stained with anti-Keratin 5 (Krt5), Keratin 8 (Krt8), Fezf2 and Aire antibodies. Scale bars indicate 100μm. Three technically and biologically independent experiments were performed. f, FACS profiles for Ly51 and UEA-1 of EpCAM+CD45-TECs derived from Chd4 cKO mice or control mice at 3-weeks old (left panels). The frequency and number of Ly51+UEA-1- (cTECs) or Ly51-UEA-1+ (mTECs) among the TECs (right panels). (n=6 biologically independent replicates per each genotypes). g, FACS profiles for CD80 and MHC class II in mTECs derived from Chd4 cKO mice or control mice at 3-weeks old (left panels). The frequency and number of CD80hiMHCIIhi mTECs (mTEChi) or CD80loMHCIIlo mTECs (mTEClo) are indicated (right panels) (n=6 biologically independent replicates per each genotypes). h, FACS profiles of CD4 and CD8 of the total cells in the spleen or lymph node of control and Chd4 cKO mice at 6-7 months old (left panels). The frequency and number of the CD4 single positive (CD4SP) and CD8 single positive (CD8SP) populations are indicated (right panels). (n=8 biologically independent replicates per each genotypes). (***p < 0.0001, unpaired two tailed t-test). Error bars, SEM.

Extended Data Fig. 4 Chd4 and Fezf2 cooperatively regulate specific chromatin structure and genes expression on a genome-wide scale.

a, b, Scatter plot of RNA-seq data of control (a) and Chd4 cKO (b) mTECs. Pearson correlation (number of genes = 21970) between three replicates were indicated. (Two-tailed t-test.) c, Heatmap of ATAC-seq reads for two replicates of WT, Fezf2 cKO and Chd4 cKO mTECs. Read count around promoter proximal or distal ATAC-seq peaks in mTECs were indicated. d, Scatter plot of ATAC-seq tag count in each replicate of WT, Fezf2 cKO and Chd4 cKO mTECs. Pearson correlation between two replicates were indicated (number of bin = 273213). e, Heatmap of ATAC-seq reads from WT, Fezf2 cKO or Chd4 cKO mTECs were indicated for each class of ATAC peaks in mTECs. Promoter proximal or distal ATAC peaks in mTECs were classified according to their dependency on Fezf2 or Chd4. f, Normalized ATAC-seq (WT, Fezf2 cKO and Chd4 cKO mTECs) profiles at promoter regions of genes in mTECs. Numbers to the left indicate the ranges of normalized tag counts. g, Effect of Chd4 on the accessibility of Fezf2-induced (fold-change in tag counts > 2), Fezf2-neutral (> 2 fold-change in tag counts > 0.5) and Fezf2-repressed (0.5 > fold-change in tag counts) peaks (left panels). Effect of Fezf2 on the accessibility of Chd4-induced (fold-change in the tag counts > 2), Chd4-neutral (> 2 fold-change in tag counts > 0.5) and Chd4-repressed (> 0.5 fold-change in tag counts) peaks (right panels). All peaks (top), promoter proximal peaks (middle), and promoter distal peaks (bottom) were analyzed. The cumulative distribution plot is presented as the cumulative probability (vertical axis) versus the log2 fold-change of the tag counts. p-values are provided in the figure (two-tailed Wilcoxon rank-sum test). h, Cumulative distribution plot of the Fezf2-dependent (left panels) or Chd4-dependent (right panels) expression change in genes near the peaks induced by Fezf2 and Chd4 (WT / Chd4 cKO > 2, WT / Fezf2 cKO > 2), or uninduced by Fezf2 and Chd4 (WT / Chd4 cKO < 2, WT / Fezf2 cKO < 2) was analyzed at all peaks (top), proximal promoter peaks (middle), and promoter distal peaks (bottom). Plots are presented as the cumulative probability (vertical axis) versus log2 fold-change of the expression values between the WT and Fezf2 KO mTECs (left panels), or control and Chd4 cKO mTECs (right panels) (horizontal axis). p-values are shown by one-tailed Wilcoxon rank-sum test.

Extended Data Fig. 5 ChIP-seq analysis of Fezf2.

a, Average plot (top) and heatmap (bottom) of Fezf2 ChIP-seq reads for replicate 1 and 2 around the promoter proximal Fezf2-binding sites (n=2018). Average tag counts are presented as lines and standard errors are presented as shades of line. b, Average plot (top) and heatmap (bottom) of ChIP-seq reads for H3K4ac, H3K4me2, H4K4me3, H3K9ac, H3K27ac and H3K27me3 in mTECs around the promoter proximal Fezf2-binding sites (n=2018, ChIP-seq peaks of Fezf2). Average tag counts are presented as lines and standard errors are presented as shades of line. c, Average plot (top) and heatmap (bottom) of mTEC ATAC-seq reads around the promoter proximal Fezf2-binding sites (n=2018). Average tag counts are presented as lines and standard errors are presented as shades of line. d, Plot of the tag density for Fezf2-ChIP around the TSS of the Fezf2-induced genes (n=640) versus Aire-induced genes (n=1553). Average tag counts are presented as lines and standard errors are presented as shades of line. e, Scatter and contour plot of log2 fold-change of ATAC-seq tag count at the Fezf2 (n=2018) or Aire (n=10761) ChIP-seq peaks. Red box; the peaks with log2 (WT / Chd4 cKO) > 1 and log2 (WT / Fezf2 cKO) > 1. Blue box; the peaks with log2 (WT / Chd4 cKO) < -1 and log2 (WT / Fezf2 cKO) < -1. Orange dotted vertical line, fold-change = -1 (left) or = 1 (right); green dotted horizontal line, fold-change = -1 (top) or = 1 (bottom). The frequency of peaks is shown in the red or blue boxes. p-values from two-tailed Fisher’s exact test are indicated.

Extended Data Fig. 6 Chd4 and Aire control chromatin structure at super-enhancers (SEs).

a, Cumulative distribution plot of the Chd4-dependency of the TRA genes induced by the Fezf2 (n=519) or Aire (n=1304) and other TRA genes (n=708), presented as the cumulative probability (vertical axis) versus log2 fold-change of the expression values between the control and Chd4 cKO mTECs (horizontal axis). P-values from two-tailed Wilcoxon rank-sum test are indicated. b-d, Effect of Aire on the accessibility of Chd4-induced (fold-change in the tag counts > 2), Chd4-neutral (> 2 fold-change in the tag counts > 0.5) and Chd4-repressed (> 0.5 fold-change in the tag counts) peaks (left panels). Effect of Chd4 on the accessibility of Aire-induced (fold-change in the tag counts > 2), Aire-neutral (> 2 fold-change in the tag counts > 0.5) and Aire-repressed (> 0.5 fold-change in the tag counts) peaks (right panels) was analyzed at all peaks (b), promoter proximal peaks (c), and promoter distal peaks (d). The cumulative distribution plot is presented as the cumulative probability (vertical axis) versus log2 fold-change of the tag counts. p-values are provided in the figure (two-tailed Wilcoxon rank-sum test). e, Raw mapped read at SEs by Chd4 ATAC-seq and H3K27ac ChIP-seq analyses in mTECs. Numbers are the ranges of normalized tag densities. SE loci; blue bar. f, Workflow of the identification of the co-expression of microclusters among the Chd4-, Aire-, and Fezf2-induced genes and super-enhancer (SE) neighborhood genes. For expression matrices containing genes regulated by Aire, Chd4, or Fezf2, or random genes and SE neighborhood genes, in which the TSS was nearest to the center of the SE, gene-by-gene Pearson correlations were calculated and affinity-propagation clustering was performed based on the Pearson correlations, with no preset number of clusters (SE neighborhood genes, n = 978; Aire-induced genes, n = 1485; Chd4-induced genes, n = 1358; Fezf2-induced genes, n = 566; random genes (matched to Aire-induced genes), n = 1485). Heatmaps of the Pearson correlations of selected co-expression clusters are shown in Fig. 5h. g, Heatmaps of the gene-by-gene Pearson correlations calculated from the expression matrix containing the SE neighborhood genes and genes induced by Aire, Chd4 or Fezf2, or random genes (Aire-induced genes and SE genes, n = 2463; Chd4-induced genes and SE genes, n = 2336; Fezf2-induced genes and SE genes, n = 1544; random genes (matched to Aire-induced genes) and SE genes, n = 2463). h, Distribution of the mean intracluster correlation of the co-expression clusters containing the SE neighborhood genes and genes induced by Aire, Chd4 or Fezf2, or random genes. For each of the Aire-, Chd4-, and Fezf2-induced genes, 1000 random permutations of the same size as the target gene sets were performed (gray histogram). i, Relationship between the mean intracluster correlation of the co-expression clusters and the number of Chd4-, Aire-, and Fezf2-induced genes or SE neighborhood genes included in each co-expression cluster. Note that one significant cluster include about 15-19 (median) Aire- or Chd4-depedent TRA genes and only a few SE neighborhood genes. j, Quantification of the interactions between the SE neighborhood genes and Aire- or Chd4-induced genes, defined as local (< 1 Mb on the same chromosome), intrachromosomal (> 1 Mb on the same chromosome), or interchromosomal (different chromosomes) in the significant co-expression microclusters shown in panel Fig. 5h.

Supplementary information

Supplementary Information

Supplementary Fig. 1 and Supplementary Tables 1 and 2.

Reporting Summary

Source data

Source Data Fig. 3

Western Blots

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Tomofuji, Y., Takaba, H., Suzuki, H.I. et al. Chd4 choreographs self-antigen expression for central immune tolerance. Nat Immunol 21, 892–901 (2020). https://doi.org/10.1038/s41590-020-0717-2

Download citation

Further reading

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing