Abstract

Autism spectrum disorder (ASD) is thought to emerge during early cortical development. However, the exact developmental stages and associated molecular networks that prime disease propensity are elusive. To profile early neurodevelopmental alterations in ASD with macrocephaly, we monitored subject-derived induced pluripotent stem cells (iPSCs) throughout the recapitulation of cortical development. Our analysis revealed ASD-associated changes in the maturational sequence of early neuron development, involving temporal dysregulation of specific gene networks and morphological growth acceleration. The observed changes tracked back to a pathologically primed stage in neural stem cells (NSCs), reflected by altered chromatin accessibility. Concerted over-representation of network factors in control NSCs was sufficient to trigger ASD-like features, and circumventing the NSC stage by direct conversion of ASD iPSCs into induced neurons abolished ASD-associated phenotypes. Our findings identify heterochronic dynamics of a gene network that, while established earlier in development, contributes to subsequent neurodevelopmental aberrations in ASD.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Data availability

RNA-seq datasets are available at EMBL-EBI ArrayExpress, with the accession code E-MTAB-6018. These raw data are associated with the following figures: Figs. 1, 2, 3, and 7d–f; and Supplementary Figs. 3c, 4, 7a–e, 8f–g, and 10. Additional data generated or analyzed during this study are included in this published article and its Supplementary information files. Supplementary tables are available for the following figures: Figs. 1, 2, 3, and 7d–f; and Supplementary Figs. 3c, 4a, 7a–d, 8f–g, and 10, as well as for additional differential expression analysis (Supplementary Tables 8 and 10). Supplementary Table 1 (Supplementary Data and Notes) provides additional information on all subject-derived iPSC lines used in this study. ATAC-seq datasets are available from the corresponding author upon request.

Additional information

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

References

  1. 1.

    de la Torre-Ubieta, L., Won, H., Stein, J. L. & Geschwind, D. H. Advancing the understanding of autism disease mechanisms through genetics. Nat. Med. 22, 345–361 (2016).

  2. 2.

    Parikshak, N. N., Gandal, M. J. & Geschwind, D. H. Systems biology and gene networks in neurodevelopmental and neurodegenerative disorders. Nat. Rev. Genet. 16, 441–458 (2015).

  3. 3.

    Parikshak, N. N. et al. Integrative functional genomic analyses implicate specific molecular pathways and circuits in autism. Cell 155, 1008–1021 (2013).

  4. 4.

    Willsey, A. J. et al. Coexpression networks implicate human midfetal deep cortical projection neurons in the pathogenesis of autism. Cell 155, 997–1007 (2013).

  5. 5.

    Courchesne, E., Carper, R. & Akshoomoff, N. Evidence of brain overgrowth in the first year of life in autism. J. Am. Med. Assoc. 290, 337–344 (2003).

  6. 6.

    Courchesne, E. et al. Mapping early brain development in autism. Neuron 56, 399–413 (2007).

  7. 7.

    Emerson, R. W. et al. Functional neuroimaging of high-risk 6-month-old infants predicts a diagnosis of autism at 24 months of age. Sci. Transl. Med. 9, eaag2882 (2017).

  8. 8.

    Hazlett, H. C. et al. Early brain development in infants at high risk for autism spectrum disorder. Nature 542, 348–351 (2017).

  9. 9.

    Marchetto, M. C. et al. Altered proliferation and networks in neural cells derived from idiopathic autistic individuals. Mol. Psychiatry 22, 820–835 (2017).

  10. 10.

    Mariani, J. et al. FOXG1-dependent dysregulation of GABA/glutamate neuron differentiation in autism spectrum disorders. Cell 162, 375–390 (2015).

  11. 11.

    Shcheglovitov, A. et al. SHANK3 and IGF1 restore synaptic deficits in neurons from 22q13 deletion syndrome patients. Nature 503, 267–271 (2013).

  12. 12.

    Stein, J. L. et al. A quantitative framework to evaluate modeling of cortical development by neural stem cells. Neuron 83, 69–86 (2014).

  13. 13.

    Allan Institute for Brain Science. Atlas of the developing human brain BrainSpan http://www.brainspan.org/ (2016).

  14. 14.

    Yuan, S. H. et al. Cell-surface marker signatures for the isolation of neural stem cells, glia and neurons derived from human pluripotent stem cells. PLoS One 6, e17540 (2011).

  15. 15.

    Zhang, B. & Horvath, S. A general framework for weighted gene co-expression network analysis. Stat. Appl. Genet. Mol. Biol. 4, e17 (2005).

  16. 16.

    Langfelder, P. & Horvath, S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics 9, 559 (2008).

  17. 17.

    Sakoe, H. Dynamic programming algorithm optimization for spoken word recognition. IEEE Trans. Acoust. Speech Signal Process. 26, 43–49 (1978).

  18. 18.

    Velichko, V. M. & Zagoruyko, N. G. Automatic recognition of 200 words. Int. J. Man. Mach. Stud. 2, 223–234 (1970).

  19. 19.

    Basu, S. N., Kollu, R. & Banerjee-Basu, S. AutDB: a gene reference resource for autism research. Nucleic Acids Res. 37, D832–D836 (2009).

  20. 20.

    Voineagu, I. et al. Transcriptomic analysis of autistic brain reveals convergent molecular pathology. Nature 474, 380–384 (2011).

  21. 21.

    Parikshak, N. N. et al. Genome-wide changes in lncRNA, splicing, and regional gene expression patterns in autism. Nature 540, 423–427 (2016).

  22. 22.

    Iossifov, I. et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature 515, 216–221 (2014).

  23. 23.

    Lancaster, M. A. et al. Cerebral organoids model human brain development and microcephaly. Nature 501, 373–379 (2013).

  24. 24.

    Qian, X. et al. Brain-region-specific organoids using mini-bioreactors for modeling ZIKV exposure. Cell 165, 1238–1254 (2016).

  25. 25.

    Molyneaux, B. J., Arlotta, P., Menezes, J. R. L. & Macklis, J. D. Neuronal subtype specification in the cerebral cortex. Nat. Rev. Neurosci. 8, 427–437 (2007).

  26. 26.

    Toma, K., Wang, T.-C. & Hanashima, C. Encoding and decoding time in neural development. Dev. Growth Differ. 58, 59–72 (2016).

  27. 27.

    Erhardt, J. A. et al. A novel F box protein, NFB42, is highly enriched in neurons and induces growth arrest. J. Biol. Chem. 273, 35222–35227 (1998).

  28. 28.

    Atkin, G. et al. Loss of F-box only protein 2 (Fbxo2) disrupts levels and localization of select NMDA receptor subunits, and promotes aberrant synaptic connectivity. J. Neurosci. 35, 6165–6178 (2015).

  29. 29.

    Vitucci, D. et al. Rasd2 modulates prefronto-striatal phenotypes in humans and ‘schizophrenia-like behaviors’ in mice. Neuropsychopharmacology 41, 916–927 (2016).

  30. 30.

    Abudureyimu, S. et al. Essential role of Linx/Islr2 in the development of the forebrain anterior commissure. Sci. Rep. 8, 7292 (2018).

  31. 31.

    Mandai, K., Reimert, D. V. & Ginty, D. D. Linx mediates interaxonal interactions and formation of the internal capsule. Neuron 83, 93–103 (2014).

  32. 32.

    Zhang, Y. et al. Rapid single-step induction of functional neurons from human pluripotent stem cells. Neuron 78, 785–798 (2013).

  33. 33.

    Kundaje, A. et al. Integrative analysis of 111 reference human epigenomes. Nature 518, 317–330 (2015).

  34. 34.

    Wilczynski, B., Liu, Y.-H., Yeo, Z. X. & Furlong, E. E. M. Predicting spatial and temporal gene expression using an integrative model of transcription factor occupancy and chromatin state. PLoS. Comput. Biol. 8, e1002798 (2012).

  35. 35.

    Jin, F. et al. A high-resolution map of the three-dimensional chromatin interactome in human cells. Nature 503, 290–294 (2013).

  36. 36.

    Chow, M. L. et al. Age-dependent brain gene expression and copy number anomalies in autism suggest distinct pathological processes at young versus mature ages. PLoS. Genet. 8, e1002592 (2012).

  37. 37.

    Gupta, S. et al. Transcriptome analysis reveals dysregulation of innate immune response genes and neuronal activity-dependent genes in autism. Nat. Commun. 5, 5748 (2014).

  38. 38.

    Sun, W. et al. Histone acetylome-wide association study of autism spectrum disorder. Cell 167, 1385–1397 (2016).

  39. 39.

    Lim, E. T. et al. Rates, distribution and implications of postzygotic mosaic mutations in autism spectrum disorder. Nat. Neurosci. 20, 1217–1224 (2017).

  40. 40.

    Hazlett, H. C. et al. Early brain overgrowth in autism associated with an increase in cortical surface area before age 2 years. Arch. Gen. Psychiatry 68, 467–476 (2011).

  41. 41.

    Schumann, C. M. et al. Longitudinal magnetic resonance imaging study of cortical development through early childhood in autism. J. Neurosci. 30, 4419–4427 (2010).

  42. 42.

    Bauman, M. L. & Kemper, T. L. Neuroanatomic observations of the brain in autism: a review and future directions. Int. J. Dev. Neurosci. 23, 183–187 (2005).

  43. 43.

    Belmonte, M. K. et al. Autism and abnormal development of brain connectivity. J. Neurosci. 24, 9228–9231 (2004).

  44. 44.

    Carper, R. A. & Courchesne, E. Localized enlargement of the frontal cortex in early autism. Biol. Psychiatry 57, 126–133 (2005).

  45. 45.

    Casanova, M. F., Buxhoeveden, D. P., Switala, A. E. & Roy, E. Minicolumnar pathology in autism. Neurology 58, 428–432 (2002).

  46. 46.

    Casanova, M. F. et al. Abnormalities of cortical minicolumnar organization in the prefrontal lobes of autistic patients. Clin. Neurosci. Res. 6, 127–133 (2006).

  47. 47.

    Weir, R. K., Bauman, M. D., Jacobs, B. & Schumann, C. M. Protracted dendritic growth in the typically developing human amygdala and increased spine density in young ASD brains. J. Comp. Neurol. 526, 262–274 (2018).

  48. 48.

    Chomiak, T., Karnik, V., Block, E. & Hu, B. Altering the trajectory of early postnatal cortical development can lead to structural and behavioural features of autism. BMC Neurosci. 11, 102 (2010).

  49. 49.

    Chomiak, T. et al. Auditory-cued sensorimotor task reveals disengagement deficits in rats exposed to the autism-associated teratogen valproic acid. Neuroscience 268, 212–220 (2014).

  50. 50.

    Vierbuchen, T. et al. Direct conversion of fibroblasts to functional neurons by defined factors. Nature 463, 1035–1041 (2010).

  51. 51.

    Fishman, V. S. et al. Cell divisions are not essential for the direct conversion of fibroblasts into neuronal cells. Cell Cycle 14, 1188–1196 (2015).

  52. 52.

    Briggs, J. A. et al. Mouse embryonic stem cells can differentiate via multiple paths to the same state. eLife 6, e26945 (2017).

  53. 53.

    Heger, A., Webber, C., Goodson, M., Ponting, C. P. & Lunter, G. GAT: a simulation framework for testing the association of genomic intervals. Bioinformatics 29, 2046–2048 (2013).

  54. 54.

    Takahashi, K. et al. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 131, 861–872 (2007).

  55. 55.

    Marchetto, M. C. N. et al. A model for neural development and treatment of Rett syndrome using human induced pluripotent stem cells. Cell 143, 527–539 (2010).

  56. 56.

    Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).

  57. 57.

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

  58. 58.

    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).

  59. 59.

    Yip, A. M. & Horvath, S. Gene network interconnectedness and the generalized topological overlap measure. BMC Bioinformatics 8, 22 (2007).

  60. 60.

    Langfelder, P., Zhang, B. & Horvath, S. Defining clusters from a hierarchical cluster tree: the Dynamic Tree Cut package for R. Bioinformatics 24, 719–720 (2008).

  61. 61.

    Aach, J. & Church, G. M. Aligning gene expression time series with time warping algorithms. Bioinformatics 17, 495–508 (2001).

  62. 62.

    Giorgino, T. Computing and visualizing dynamic time warping alignments in R: the dtw package. J. Stat. Softw. 31, 1–24 (2009).

  63. 63.

    Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 2498–2504 (2003).

  64. 64.

    Huang, W., Sherman, B. T. & Lempicki, R. A. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 4, 44–57 (2009).

  65. 65.

    Ashburner, M.et al. The Gene Ontology Consortium. Gene ontology: tool for the unification of biology. Nat. Genet. 25, 25–29 (2000).

  66. 66.

    Hosack, D. A., Dennis, G. Jr., Sherman, B. T., Lane, H. C. & Lempicki, R. A. Identifying biological themes within lists of genes with EASE. Genome Biol. 4, R70 (2003).

  67. 67.

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

  68. 68.

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

  69. 69.

    Ross-Innes, C. S. et al. Differential oestrogen receptor binding is associated with clinical outcome in breast cancer. Nature 481, 389–393 (2012).

  70. 70.

    Hahne, F. & Ivanek, R. Visualizing genomic data using Gviz and Bioconductor. Methods Mol. Biol. 1418, 335–351 (2016).

  71. 71.

    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).

Download references

Acknowledgements

We thank the subjects who participated in this study. We also thank M.L. Gage and M. Pena for editorial comments; E. Courchesne for his contribution to fibroblast collection; L. Moore, A. Mendes, B. Miller, K. Rehder, A. Mar and K. Niedrig for technical assistance; K. Diffenderfer and the Salk STEM core facility for technical support; C. Fitzpatrick and C. O’Connor for help with flow cytometry; and T. Toda for helpful discussions. This work was supported by the Flow Cytometry Core Facility and the NGS core facility of the Salk Institute with funding from NIH-NCI CCSG: P30 014195, the Chapman Foundation and the Helmsley Charitable Trust. We also acknowledge support from The James S. McDonnell Foundation, G. Harold & Leila Y. Mathers Charitable Foundation, JPB Foundation, the March of Dimes Foundation, NIH (grant nos MH095741 and MH090258 to F.H.G. and RO3 MH115426-01A1 to M.C.M.), The Engman Foundation, Annette C. Merle-Smith, The Paul G. Allen Family Foundation and The Leona M. and Harry B. Helmsley Charitable Trust (grant no. 2017-PG-MED001). S.T.S. was funded by a fellowship from the German Research Foundation and was recently awarded the NARSAD Young Investigator Grant from the Brain & Behavior Research Foundation.

Author information

Affiliations

  1. Laboratory of Genetics, The Salk Institute for Biological Studies, La Jolla, CA, USA

    • Simon T. Schafer
    • , Apua C. M. Paquola
    • , Shani Stern
    • , Monique Pena
    • , Thomas J. M. Kuret
    • , Marvin Liyanage
    • , Abed AlFatah Mansour
    • , Baptiste N. Jaeger
    • , Maria C. Marchetto
    • , Jerome Mertens
    •  & Fred H. Gage
  2. Lieber Institute for Brain Development, Baltimore, MD, USA

    • Apua C. M. Paquola
  3. Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD, USA

    • Apua C. M. Paquola
  4. Centre de Recherche du Centre Hospitalier Universitaire de Québec–Université Laval, Département de Médecine Moléculaire, Faculté de Médecine, Université Laval, Québec, Canada

    • David Gosselin
  5. Next Generation Sequencing Core, The Salk Institute for Biological Studies, La Jolla, CA, USA

    • Manching Ku
  6. Division of Pediatric Hematology and Oncology, Department of Pediatric and Adolescent Medicine, Faculty of Medicine, University of Freiburg, Freiburg, Germany

    • Manching Ku
  7. Laboratory of Neural Plasticity, Faculties of Medicine and Science, Brain Research Institute, University of Zurich, Zurich, Switzerland

    • Baptiste N. Jaeger
  8. Department of Cellular and Molecular Medicine, University of California, San Diego, La Jolla, CA, USA

    • Christopher K. Glass
  9. Department of Genomics, Stem Cell Biology and Regenerative Medicine, Institute of Molecular Biology & CMBI, University of Innsbruck, Innsbruck, Austria

    • Jerome Mertens

Authors

  1. Search for Simon T. Schafer in:

  2. Search for Apua C. M. Paquola in:

  3. Search for Shani Stern in:

  4. Search for David Gosselin in:

  5. Search for Manching Ku in:

  6. Search for Monique Pena in:

  7. Search for Thomas J. M. Kuret in:

  8. Search for Marvin Liyanage in:

  9. Search for Abed AlFatah Mansour in:

  10. Search for Baptiste N. Jaeger in:

  11. Search for Maria C. Marchetto in:

  12. Search for Christopher K. Glass in:

  13. Search for Jerome Mertens in:

  14. Search for Fred H. Gage in:

Contributions

S.T.S. designed, performed, analyzed and contributed to all experiments. A.P. aligned the RNA-seq dataset, performed statistical analysis and helped with data interpretation. S.S. conducted and analyzed the electrophysiological recording experiments. M.K. performed RNA-seq experiments and helped with interpretation of results. D.G. performed ATAC-seq experiments. D.G. and C.K.G. analyzed data and helped with interpretation of results. M.P. conducted flow cytometry, imaging experiments, analyzed data and assisted with tissue culture and organoid experiments. M.L. and T.K. conducted structural and morphological analysis and performed cloning experiments. A.M. helped with organoid experiments and analyzed data. B.J. contributed to flow cytometry analysis. M.C.M. provided iPSC lines and helped with data interpretation. J.M. provided constructs, assisted with establishment of iPSC-iN protocol and helped with data interpretation. F.H.G. supervised the experimental design and analysis, interpreted results and provided funding. S.T.S. and F.H.G. wrote the manuscript and conceptualized the study.

Competing interests

The authors declare no competing interests.

Corresponding author

Correspondence to Fred H. Gage.

Integrated supplementary information

  1. Supplementary Figure 1 Characterization of iPSC-derived neural stem cells (NSCs).

    a, Quantitative real time PCR assay for assessing retroviral transgene silencing for c-MYC, OCT3/4, KLF4 and SOX2 in ASD and control iPSC lines. Values represent means ± s.e.m. from n = 3 independent cell culture replicates. b, Schema showing directed differentiation of subject-derived iPSCs into functional neurons. c, Representative confocal images of iPSC-derived NSCs that show homogeneous expression of the markers Nestin and Pax6. Scale bar 50 μm. Experiments were performed in all 13 subject lines and 4 cell culture replicates each with similar results. d, Immunofluorescence images of six-week-old neurons stained with βIII-tubulin (Tuj1), Map2ab and vGlut1. Scale bar 100 μm (left) and 10 μm (right). The staining was repeated in all 13 subject lines at least once with similar results. e, Stereological counts of iPSC-derived NSCs expressing Nestin and PAX6. Values represent means ± s.e.m. from n = 3 independent cell culture replicates. f, Quantitative real time PCR assay for assessing the expression of the anterior marker PAX6. Values represent means ± s.e.m. from n = 3 independent cell culture replicates. g, Representative confocal images of iPSC-derived NSCs that show expression of Nestin and the forebrain-associated marker FoxG1. Scale bar 50 μm. Immunohistochemistry was performed in all 13 subject lines with 4 cell culture replicates each. h, Quantification of FoxG1-positive iPSC-derived NSCs across all subject cell lines. Box plots show median (center line) and interquartile range (IQR), with whiskers representing the minimum and maximum; n = 4 independent cell culture replicates.

  2. Supplementary Figure 2 Characterization of iPSC-derived neurons.

    a, Representative confocal images of three-week-old iPSC-derived neurons that show expression for TBR1, a marker for early glutamatergic projection neurons of the cerebral cortex. Scale bar 50 μm. Immunohistochemistry was performed in all 13 subject lines and 3 cell culture replicates each with similar results. b, Stereological quantifications of iPSC-derived Tuj1-positive neurons that express the cortical marker TBR1 at three weeks of differentiation. Values represent means ± s.d. from n = 3 independent cell culture replicates. c, Regional enrichment analysis for two-week-old NSC-derived neurons. Shown are the results of transcriptome correlation analyses between two-week-old NSC-derived neurons (ASD: n = 8, control: n = 5) and post-mortem human brain samples from the BrainSpan dataset (n = 524). The x-axis shows agglomerated brain regions of the BrainSpan post-mortem brain samples. The y-axis shows the relative percentage among brain regions that were classified for each iPSC-derived neuron sample. The classification was based on a sample-specific correlation threshold (Supplementary Methods). Values represent means ± s.d.; NCX, neocortex (see d); AMY, amygdala; HIP, hippocampus; CGE, caudal ganglionic eminence; LGE, lateral ganglionic eminence; MGE, medial ganglionic eminence; DTH, dorsal thalamus; CB, cerebellum. d, Enriched regions of the neocortex (NCX). Shown are percentage values for different cortical regions represented as NCX in f. MFC, medial prefrontal cortex; DFC, dorsolateral prefrontal cortex; OFC, orbital frontal cortex; VFC, ventrolateral prefrontal cortex; Ocx, occipital neocortex; Pcx, parietal neocortex; STC, superior temporal cortex; M1C-S1C, primary motor-sensory cortex. e, Developmental stages of samples corresponding to the regional classifications in d. pcw, post-conceptual week. Values represent means ± s.d.; ASD (n = 8), control (n = 5); n refers to biologically independent subject lines. f, Electrophysiological recordings of iPSC-derived neurons at different time points during in vitro differentiation. Values represent mean ± s.e.m.; 14 days (14 DIF, n = 5 neurons), 22 days (22 DIF, n = 6 neurons), 42 days (42 DIF, n = 6 neurons).

  3. Supplementary Figure 3 Time series-based RNA sequencing captures developmental gene expression dynamics.

    a, Representative images of retroviral lineage-tracing experiments in progressively maturing neurons. Retrovirus (RV)-labeled cells express eGFP and co-expression of PSA-NCAM is indicative of the neuronal lineage. Scale bar 20 μm. Immunohistochemistry was performed in all 13 subject lines and 2 cell culture replicates each with similar results (also see Supplementary Figure 5a). b, Schematic of FACS-based purification of NSCs and defined subpopulations of eGFP+/PSA-NCAM+ neurons over the time course of in vitro differentiation. c, Heatmaps of time-course RNAseq expression analysis showing the top 300 upregulated and top 200 downregulated genes (by fold change) during neuronal differentiation. Expression values are shown for one representative control cell line. Corresponding GO term analyses for differentially expressed genes (Wald test implemented in the DESeq2 package with Benjamini-Hochberg adjusted P < 0.01) are depicted in the table to the right (Supplementary Table 2 and 3); GO terms upregulated: n = 2220 genes, GO terms downregulated: n = 1344 genes.

  4. Supplementary Figure 4 Gene network analysis of NSC-derived neurons.

    a, Module-trait associations. Detailed analysis showing Pearson’s correlation coefficients and student p-values for associations between covariates (x-axis) and module eigengenes; n = 65 time series samples; ASD: n = 8 independent cell lines at 5 time points, control: n = 5 independent cell lines at 5 time points; All p-values are shown where associations passed multiple test correction, with false discovery rate (FDR) adjusted P < 0.05. b, DTW alignment of timing modules TM2 (MElightcyan1) and TM3 (MEsaddlebrown) between control and ASD time series. Cost density plots show average per-step cost densities with labeled contours superimposed. c, GO term analysis of the genes assigned to TM2 (MElightcyan1) and TM3 (MEsaddlebrown). P-values were calculated with a modified Fisher’s exact test (EASE, see methods) and are shown only for those terms with Benjamini-Hochberg adjusted P < 0.01. d, WGCNA cluster dendrogram of an alternatively constructed gene network, in which lowly expressed genes were filtered prior to network construction. e, Preservation statistics for TMs identified. Note that all three gene modules are highly preserved in the alternatively constructed network in d. Module preservation statistics were calculated using the modulePreservation function implemented in WGCNA; TM1 (n = 1,530), TM2 (n = 294), TM3 (n = 401), n refers to the number of genes. f, Gene set enrichment analysis comparing the 44 co-expression modules with multiple gene sets from post-mortem ASD cortex microarray (Voineagu et al. 2011) and RNAseq (Parikshak et al. 2016), human cortical development (Parikshak et al. 2013), genes enriched for ASD-associated rare variants (SFARI) and genes with de novo variants associated with intellectual disability (ID) (Iossifov et al. 2014). Boxes are filled if the enrichment is P < 0.05 (Fisher’s exact test). The WGCNA modules on the x-axis are ordered from left to right based on the absolute time-based module significance measure; SFARI (n = 859), M2 (n = 1,042), M3 (n = 997), M13 (n = 871), M16 (n = 492), M17 (n = 1,042), ID (n = 227), M12 (n = 380), CTX.4 (n = 234), CTX.16 (n = 277), CTX.19 (n = 273), CTX.20 (n = 332), CTX.22 (n = 215), CTX.24 (n = 1,219), background set (n = 22,682); n refers to the number of gene IDs.

  5. Supplementary Figure 5 Morphometric assessment of subject-derived neurons throughout maturation.

    a, Representative confocal images of the morphological development of ASD and control neurons at different days post retroviral infection (dpi) expressing GFP, early neuronal genes (DCX, PSA-NCAM) and axonal filaments (Smi312). Scale bars 50 μm. Immunohistochemistry was performed in all 13 subject lines and 2 cell culture replicates per time point with similar results. b, ASD neurons were significantly more complex at 4, 7 and 14 dpi (two-way ANOVA with Sidak correction, 4dpi: ***P = 0.0006, 7dpi: *P = 0.0147, 14dpi: ****P < 0.0001). Values show means ± s.d.; ASD (n = 8; 320 technical tracing replicates total), control (n = 5; 244 technical tracing replicates total); n refers to biologically independent subject lines. c, Sholl analysis of neurite length from ASD and control neurons at 4 dpi. Total sholl neurite length complexity was larger in the ASD group as compared with controls (two-way ANOVA with Sidak correction, *P < 0.05, **P < 0.001, ***P < 0.001, ****P < 0.0001). Values show means ± s.e.m.; ASD (n = 8), control (n = 5); n refers to biologically independent subject lines. d, Flow cytometry-based gating strategy for the assessment of hDCX:GFP reporter activity during defined neurodevelopmental stages. Experiments were repeated in all 13 subject lines at 5 different time points with similar results; numeric values represent percentages. e, ASD neurons showed altered dynamics in sustaining premature activation of the inserted hDCX promoter with significantly lower fold-changes required to reach maximum activity at 2 and 5 dpi (two-way ANOVA with Sidak correction, 3 dpi: ***P = 0.0002, 5 dpi: **P = 0.0043). The data were normalized to the NSC stage activity for each line as shown in Fig. 2k. Values represent means ± s.e.m.; ASD (n = 8), control (n = 5); n refers to biologically independent subject lines.

  6. Supplementary Figure 6 Aberrant maturational features in developing ASD neurons.

    a, Schematic of the retroviral reporter system for the assessment of Wnt signaling dynamics. Flow cytometry-based analysis of this reporter allows for quantification of transient signaling events through measuring mean fluorescence intensities (MFI). b and c, NSCs were infected with retroviral reporter constructs and analyzed through the time-course of differentiation by gating on defined subpopulations and measuring MFIs. TCF/LEF signals were normalized to the mutant TCF/LEF reporter (mut TCF/LEF); numerical values represent percentages. d and e, Representative TCF/LEF:GFP histograms and relative TCF/LEF signal induction in differentiating neurons. ASD neurons showed accelerated dynamics in downregulating responsiveness to Wnt3a-induced canonical TCF/LEF signaling at 4 and 7 dpi (two-way ANOVA with Sidak correction, 4 dpi: **P = 0.0097, 7 dpi: *P = 0.0108). The values were normalized to the NSC stage activity for each line and represent means ± s.e.m.; ASD (n = 8; 16 technical replicates total), control (n = 5; 10 technical replicates total); n refers to biologically independent subject lines. f, Representative confocal images of five-week-old cerebral organoids. SOX2 immunoreactivity allows for identification of ‘ventricular zone (VZ)-like’ regions. The outer layer was defined as the area from outside the VZ to the nearest pial surface, mostly harboring early immature neurons (DCX+). The white lines indicate the boarders of the aforementioned regions. Scale bar 50 μm. Immunohistochemistry was repeated in 6 subject lines with similar results. g, Schematic showing the stereological approach to assess early neuronal differentiation in ‘VZ-like’ regions of cerebral organoids (see Supplementary Methods) h, Quantitative assessment of early neuronal differentiation in cerebral organoids shows significantly increased numbers of DCX + cells in ASD (%DCX/DAPI: ***P = 0.0002, ratio DCX/SOX2: ***P = 0.0003, Mann-Whitney U-test). Box plots show median (center line) and IQR, with whiskers representing the minimum and maximum; ASD (n = 3; 26 ‘VZ-like’ regions from 8 organoid replicates total), control (n = 3; 17 ‘VZ-like’ regions from 7 organoid replicates total); n refers to independent subject lines. i, Representative confocal images of six-week-old forebrain organoids labeled for TBR1, a marker for deep-layer cortical neurons. The white lines indicate the boarders of VZ- and cortical plate (CP)-like regions. Scale bar 50 μm. Immunohistochemistry was repeated in 6 subject lines and 2 different organoid batches with similar results. j, TBR1+ CP-like regions are significantly larger in ASD forebrain organoids (****P < 0.0001, Mann-Whitney U-test). Box plots show median (center line) and IQR, with whiskers representing the minimum and maximum; ASD (n = 3; 27 CP-like regions from 9 organoid replicates total), control (n = 3; 27 CP-like regions from 8 organoid replicates total); n refers to independent subject lines. k, Representative confocal images of retrovirally labeled neurons at 6wpi in CP-like regions expressing TBR1. Scale bar 25 μm. Immunohistochemistry was repeated in 6 subject lines and 2 different organoid batches with similar results. l, Percentage of GFP/TBR1 + cells in forebrain organoids. Values represent mean ± s.e.m.; RV-GFP 2wpi (n = 3; 12 organoid replicates total), RV-GFP 6wpi (n = 2; 10 organoid replicates total); n refers to independent organoid batches.

  7. Supplementary Figure 7 Aberrant overrepresentation of TM1 genes causes maturational acceleration.

    a, Distribution of individual signed gene significance (GS) values for TM1 genes with kME > 0.7 for different developmental stages. Center measure shows mean from n = 620 TM1 genes. GS values were defined as the Student t-test statistic (two-sided) for testing differential expression between time points and conditions. b, Number and percentage of TM1 genes that feature significantly negative correlations (two-sided Student t-test with FDR < 0.05, n = 65 time series samples) at different developmental stages. Note that control NSCs have considerably more suppressed network genes than ASD NSCs (Supplementary Table 9). c, Left: Schematic Venn diagram showing the overlap of genes in TM1 that are significantly negatively correlated with the NSC stage in ASD and control individuals. Right: Distribution of Gene Significance (GS) values of the 189 genes specifically suppressed in control NSCs between conditions. Violin plots show median (center line), IQR (box), 95% confidence interval and the kernel probability density at different values; n = 189 genes. d, Gene-gene connections for the top 30 negatively correlated TM1 genes in control NSCs. Gradual increasing point size of nodes corresponds to GS values in control NSCs (two-sided Student t-test, ASD NSCs: n = 8 samples; control NSCs: n = 5 samples; Supplementary Table 9). e, Comparative gene expression plots of quantitative real time PCR (qRT-PCR) assays and RNA-seq normalized counts for TM1 genes differentially expressed at the NSC stage; qRT-PCR: two-sided Student-t test, FBXO2: ***P = 0.0001, RASD2: ***P = 0.0009, ISLR2: *P = 0.0127, INSM2: **P = 0.0011; RNA-seq: Wald test as implemented in DESeq2, q-values shown were adjusted by the Benjamini-Hochberg procedure). Box plots show median (center line) and interquartile range (IQR), with whiskers representing the minimum and maximum; ASD NSCs: n = 8, control NSCs: n = 5. f and g, Gating strategy for flow cytometry-based quantification of FBXO2 protein levels in ASD and control NSCs (SOX2+ cells); Experiments were performed in all 13 subject lines with similar results; numeric values represent percentages. h, ASD NSCs show elevated FBXO2 protein levels as compared with control NSCs (*P = 0.0101, Mann-Whitney U-test); ASD NSCs: n = 8, control NSCs: n = 5. i and j, Schematic showing the experimental design for mimicking aberrant TM1 gene expression dynamics. NSCs were infected with an inducible lentiviral construct designed to control the expression levels of FBXO2 (FBXO2-GFP) alone or in combination with TM1 genes FBXO44 and CEND1 (FFC-GFP). Overexpression of GFP was used as a control (UGFP). One day later, transgene expression was induced by adding doxycycline, and neuronal differentiation of NSCs was induced the following day. k, Representative confocal images of the morphological development of control neurons (GFP) and those that started with an aberrant overrepresentation of FBXO2 (FBXO2-GFP) alone or in combination with TM1 genes FBXO44 and CEND1 (FFC-GFP). Scale bars 50 μm. Immunohistochemistry was repeated in 2 independent control lines with similar results. l, Left: Total neurite length of NSC-derived neurons at 4 dpi. ASD neurons had significantly longer neurites at 4 dpi as compared with controls (control: 114.23 ± 9.7 μm, ASD: 226.04 ± 12.22 μm, **P = 0.0016, Mann-Whitney U-test). Values represent means ± s.e.m.; ASD: n = 8 biologically independent cell lines, control: n = 5 biologically independent cell lines. Right: Neurons overexpressing FBXO2-GFP or FCC-GFP had significantly longer neurites at 4 dpi as compared with GFP expressing control neurons (GFP control: 106.74 ± 7.76 μm, n = 22 cells; FBXO2-GFP: 175.83 ± 11.58 μm, n = 31 cells; FFC-GFP: 272.9 ± 23.15 μm, n = 34 cells; one-way ANOVA with Dunnett’s multiple comparisons test, GFP vs FBXO2-GFP: *P = 0.0199; GFP vs FFC-GFP: ****P = 0.0001). Values represent means ± s.e.m.; n = 2 biologically independent control cell lines. m, Left: ASD neurons show significantly higher numbers of PSA-NCAM + cells as compared with controls at 4 dpi (**P = 0.0016, Mann-Whitney U-test). Values represent means ± s.e.m.; ASD: n = 8 biologically independent cell lines, control: n = 5 biologically independent cell lines. Right: Neurons overexpressing FBXO2-GFP or FCC-GFP had significantly higher numbers of PSA-NCAM + cells at 4 dpi as compared with GFP expressing control neurons (one-way ANOVA with Dunnett’s multiple comparisons test, GFP vs FFC-GFPlow: *P = 0.0231; GFP vs FBXO2-GFPmed: *P = 0.0348; GFP vs FFC-GFPhigh: ***P = 0.0010). Values represent means ± s.e.m.; n = 5 biologically independent control cell lines for each of the three groups.

  8. Supplementary Figure 8 Characterization of iPSC-iNs.

    a, Schematic showing the paradigm for direct conversion of iPSCs into induced neurons (iPSC-iNs) by forced expression of an inducible Ngn2 transgene (see Methods). b, Morphological development of iPSC-iNs over the time course of differentiation. Experiments were repeated with 13 independent subject lines with similar results. c, iPSC-iNs express hTau after two weeks of differentiation. The staining was repeated in all 13 subject lines at least once with similar results. d, Regional enrichment analysis for two-week-old iPSC-iNs. Shown are the results of transcriptome correlation analyses between two-week-old iPSC-iNs (ASD: n = 8, control: n = 5) and post-mortem human brain samples from the BrainSpan dataset (n = 524). The x-axis shows agglomerated brain regions of the BrainSpan post-mortem brain samples. The y-axis shows the relative percentage among brain regions that were classified for each iPSC-iN sample. The classification was based on a sample-specific correlation threshold (Supplementary Methods). Values represent means ± s.d.; NCX, neocortex (see e); AMY, amygdala; HIP, hippocampus; CGE, caudal ganglionic eminence; LGE, lateral ganglionic eminence; MGE, medial ganglionic eminence; DTH, dorsal thalamus; CB, cerebellum; STR, striatum. e, Enriched regions of the neocortex (NCX). Shown are percentage values for different cortical regions represented as NCX in f. MFC, medial prefrontal cortex; DFC, dorsolateral prefrontal cortex; OFC, orbital frontal cortex; VFC, ventrolateral prefrontal cortex; Ocx, occipital neocortex; Pcx, parietal neocortex; STC, superior temporal cortex; V1C, primary visual cortex; M1C-S1C, primary motor-sensory cortex. f, Heatmap showing normalized expression values for selected genes across all 14-day-old iPSC-iN samples. IPSC-iNs showed high expression levels of pan-neuronal markers such as RBFOX1, NCAM1 and MAP2 as well as telencephalic markers such as BRN2(POU3F2), CUX1 and SATB2. g, Heatmap showing changes in gene expression between iPSC and iPSC-iNs 14 days after conversion for selected genes and across all lines (see also Supplementary Table 8). Compared with their corresponding iPSC stages, iPSC-iNs showed significant upregulation of other cortical transcription factors including TBR1, whereas other prominent forebrain transcription factors such as CTIP2/BCL11B were lacking. h, Representative immunofluorescence image of two-week-old iPSC-iNs stained with βIII-tubulin (Tuj1) and vGlut1. Scale bar represents 5 μm. The staining was repeated in all 13 subject lines at least once with similar results. i, Electrophysiological characterization over the time course of iPSC-iN maturation. Values represent mean ± s.e.m.; 8 days (iPSC N 8, n = 8 neurons), 15 days (iPSC N 15, n = 9 neurons), 28 days (iPSC N 28, n = 4 neurons).

  9. Supplementary Figure 9 Direct conversion of iPSCs into neurons circumvents the NSC state.

    a, Representative immunofluorescence images of iPSC-iNs that express the immature neuronal marker during an early multipolar stage. Scale bar represents 50 μm. Experiments were repeated with 13 independent subject lines with similar results. b, Quantification of NSC-specific gene expression during iPSC-iN conversion in comparison to directed differentiation (NSC). During these early time points, none of the tested NSC-specific genes, including Nestin, SOX2 and PAX6, changed expression as compared with the iPSC stage, whereas neuronal markers such as TUBB-3 markedly increased. Conversely, NSCs derived through directed differentiation from the same iPSC lines revealed a dramatic increase in expression of NSC stage-specific genes. Gene expression is normalized to respective iPSC states and depicted as Fold Change; Values represent mean ± s.e.m.; n = 3 independent control cell lines. c, Schematic of combined lentiviral system to monitor nestin promoter activity during iPSC-iN conversion (nestin::eGFP-TRE::N2AR). d, Gating procedure for reporter-based lineage tracing of iPSC-iNs; numerical values represent percentages. e, Histograms of flow cytometry-based assessment of nestin promoter activity during iPSC-iN conversion (iN12 hrs, 48 hrs and 60 hrs) and directed differentiation (NSC). Experiments were repeated once with similar results; numerical values represent percentages. f, Sholl analysis of neurite length from ASD and control iPSC-iNs at 4, 7 and 14 days. Total sholl neurite length complexity was similar in both groups (two-way ANOVA with Sidak correction, not significant). Values show means ± s.e.m.; ASD iPSC-iNs (n = 8), control (n = 5); n refers to biologically independent subject lines. g, Sholl intersections from ASD and control iPSC-iNs at 4, 7 and 14 days was similar in both groups (two-way ANOVA with Sidak correction, not significant). Values show means ± s.e.m.; ASD iPSC-iNs (n = 8), control (n = 5); n refers to biologically independent subject lines.

  10. Supplementary Figure 10 Gene network analysis of maturing iPSC-iNs.

    a, Schematic of FACS-based purification of iPSCs (SSEA4+/TRA1–81+) and iPSC-iNs (eGFP+/PSA-NCAM+) over the time course of in vitro differentiation. b, Module-trait associations. Detailed analysis showing Pearson’s correlation coefficients and student p-values for associations between covariates (x-axis) and module eigengenes; n = 52 iPSC-iN time series samples; ASD iPSC-iNs: n = 8 independent cell lines at 4 time points, control iPSC-iNs: n = 5 independent cell lines at 4 time points. c, Scatterplots showing ranked expression (left) and connectivity (right) between the two temporal datasets (NSC-derived neurons (NSC-Ns) and iPSC-iNs). The highly significant Pearson correlations between the NSC-N and the iPSC-iN network (P < 1 × 10−200, two-sided Student t-test) suggests that the datasets are comparable; n = 22,111 genes. d, Gene set enrichment analysis comparing the 18 co-expression modules of the iPSC-iN network with the TM1, TM2 and TM3 gene sets from the network of NSC-derived neurons. Boxes are filled if the enrichment P < 0.05 and odd ratios are presented (Fisher’s exact test); TM1: n = 1,530 genes, TM2: n = 294 genes, TM3: n = 401 genes. e, Scatterplot between Gene Significance (GS time; days of iN conversion) and intramodular connectivity (kME) of TM1 genes within the TM1-equivalent module MEblue in the iPSC-iN network. Each point corresponds to a gene in the blue module. GS time and kME exhibit a very significant Pearson correlation, implying that the hub genes of the blue module also tend to be highly correlated with time in the iPSC-iN network (P < 1 × 10−200, two-sided Student t-test); n = 1,530 genes. f, Module eigengene (ME) dynamics of the iPSC-iN module blue averaged across developmental time. Upper panel: Heatmap showing average MEs of iPSC-iN module blue across time. Blue: low expression, Red: high expression. Lower panels: Plots of iPSC-iN module blue showing trajectories of module eigengenes for control (black) and ASD (red) across time, with dots representing values for each individual. g, GS values of the top 60 hub genes across developmental time and between conditions. Box plots show median (center line) and IQR, with whiskers representing the minimum and maximum; n = 60 TM1 genes.

  11. Supplementary Figure 11 ATAC-seq quality metrics.

    a, Fragment length distribution of ATAC-seq reads from all size-selected libraries. Most of the reads fall into the NFR ( < 100 bp) and a clear mono-nucleosomal peak can be seen in all libraries. b, Fraction of reads per sample that fall into the NFR ( < 100 bp). All libraries show a similar high amount of NFR c, Distribution of peak width between conditions called on NFR reads with the parameters defined in Supplementary Methods. d, Principal component analysis based on DA peaks shows a clear separation between the ASD and control group; ASD: n = 8, control: n = 5. e, Enrichment (log2 FoldChange) for DA peaks that fall into gene-distal regions of the identified TMs. f, Fraction of TM hub genes (kME > 0.8) with DA peaks in gene-distal regions. g, Coverage maps of normalized ATAC-seq signals from ASD and control NSCs around the promoter region of NCDN, a TM1-associated gene, on chromosome 1. Individual sample coverages (ASD: n = 8, control: n = 5) are displayed and annotations below are based on previously annotated chromatin states (see Supplementary Methods).

Supplementary information

  1. Supplementary text and figures

    Supplementary Figures 1–11, Supplementary Table 1 and Supplementary Methods

  2. Reporting Summary

  3. Supplementary Table 2

    Differentially expressed genes (DEGs) in the course of neuronal maturation (Wald test implemented in the DESeq2 package with Benjamini–Hochberg adjusted P values – padj); n = 5 independent control subject lines at the NSC stage and at day 14 (related to Supplementary Fig. 3).

  4. Supplementary Table 3

    GO term analyses for DEGs (Wald test implemented in the DESeq2 package with Benjamini–Hochberg adjusted P<0.01); GO terms upregulated: n = 2,220 genes, GO terms downregulated: n = 1,344 genes; EASE score, a modified Fisher’s exact test as implemented in DAVID Bioinformatics Resources 6.8 (related to Supplementary Fig. 3).

  5. Supplementary Table 4

    Gene network for time-series of NSC-derived neurons (NSC-N). Module assignment and module membership for each gene. WGCNA results; n = 65 NSC-N time series samples.

  6. Supplementary Table 5

    ASD-risk genes used for calculations of module enrichment. Related to Fig. 1.

  7. Supplementary Table 6

    GO term enrichment analysis for genes assigned to TM1; n = 1,530 genes (statistical overrepresentation test is based on a binominal test implemented in the PANTHER tool). Related to Fig. 2.

  8. Supplementary Table 7

    Table listing all iPSC lines and clones used for each experiment.

  9. Supplementary Table 8

    Differential expression of TM1-specific genes between the NSC stage (D0) and 4 d of differentiation. Individual tabs list the genes and test parameters for the control and ASD groups (Wald test as implemented in DESeq2, padj refers to P values adjusted by the Benjamini–Hochberg procedure; ASD NSCs: n=8, control NSCs: n=5). Related to Supplementary Fig. 7.

  10. Supplementary Table 9

    Gene significance (stage-specific significance of Pearson correlation) for TM1 genes (kME > 0.7) that show significantly negative correlations at the NSC stage in controls as compared with all other stages (two-sided Student’s t test with FDR adjusted P values, n = 65 time-series samples). Related to Supplementary Fig. 7.

  11. Supplementary Table 10

    Differential gene expression between control iPSCs and 14-d-old iPSC-iNs (Wald test as implemented in DESeq2; padj refers to P values adjusted by the Benjamini–Hochberg procedure; ASD NSCs: n = 8, control NSCs: n = 5). Related to Figure 4.

  12. Supplementary Table 11

    Gene network for iPSC-iN time-series. Module assignment and module membership for each gene. WGCNA results; n = 53 iPSC-iN time series samples. Related to Figure 4.

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/s41593-018-0295-x