Single-cell brain organoid screening identifies developmental defects in autism

In the paper

“The dual gRNA is located within the 3′ long terminal repeat (LTR) and is thus transcribed by RNA polymerase II to be captured by scRNA-seq assays17.”

What the sources say

• Partly supported

  In “Direct detection of gRNAs from single-cell transcriptome data” (p. 2), the paper places an hU6-gRNA cassette in the 3′ LTR and says the gRNA becomes part of an RNA polymerase II transcript detectable by poly-A-enriched RNA-seq. The paper describes an individual gRNA in that cassette, however, not a dual-gRNA construct; that qualifier is unsupported. (LLM, conf=0.84)

  cites [17] Datlinger, P. et al. Pooled CRISPR screening with single-cell transcriptome read-out. Nat. Methods 14, 297–301 (2017). sourceExperimental — accuracy not yet measured

In the paper

“Notably, we found a cluster of interneurons expressing MEIS2 and high levels of PAX6 (LGE PAX6+ interneurons), a signature that resembles mouse olfactory bulb precursors that were recently reported to generate neurons redirected to white matter in primates29.”

What the sources say

• Supported

  In “Reuse of OB neurons in primate cerebrum” (page unavailable), the paper reports that the LGE_MEIS2/PAX6 class has a strong transcriptomic correlation with mouse olfactory-bulb granule cells. The same section reports that neurons of this class migrate through primate cortical streams and persist in deep white matter, supporting the cited paraphrase. (LLM, conf=0.90)

  cites [29] Schmitz, M. T. et al. The development and evolution of inhibitory neurons in primate cerebrum. Nature 603, 871–877 (2022). source

In the paper

“Many ASD risk genes associated with transcriptional regulation and chromatin remodelling are crucial for cortical development21,22.”

What the sources say

• Partly supported

  The Conclusions call chromatin remodeling and transcription critical pathways for typical development and link chromatin remodeling to neurogenesis and neural differentiation. The Integrating complementary data section uses mid-fetal neocortex as a developmental context for identifying ASD risk genes, but does not establish that many of the transcriptional and chromatin-remodeling risk genes are individually crucial for cortical development. Page numbers are unavailable in our copy. (LLM, conf=0.82)

  cites [21] Study, T. D. et al. Synaptic, transcriptional, and chromatin genes disrupted in autism. Nature 515, 209–215 (2014). sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites [22] Trevino, A. E. et al. Chromatin and gene-regulatory dynamics of the developing human cerebral cortex at single-cell resolution. Cell 184, 5053–5069.e23 (2021). source

In the paper

“Moreover, coexpression network analyses of ASD genes indicate that the developmental defects associated with ASD may arise during fetal stages5,6, periods that are difficult to investigate.”

What the sources say

• Supported

  The Discussion reports that coexpression networks involving ASD risk genes converge in cortical projection neurons during midfetal human development (10–24 postconceptional weeks). It also describes robust evidence of vulnerability during midfetal development, supporting the citing sentence’s qualified claim that ASD-associated developmental defects may arise during fetal stages. (LLM, conf=0.92)

  cites [5] Willsey, A. J. et al. Coexpression networks implicate human midfetal deep cortical projection neurons in the pathogenesis of autism. Cell 155, 997–1007 (2013). source
• Supported

  The Results section (‘Genes implicated in ASD are highly co-expressed during human cortical development’; page not supplied) reports that ASD gene-network analyses converge on modules involved in prenatal and early postnatal development. The Discussion also reports a signal during fetal development. These findings support the claim’s qualified statement that ASD-associated developmental defects may arise during fetal stages. (LLM, conf=0.90)

  cites [6] Parikshak, N. N. et al. Integrative functional genomic analyses implicate specific molecular pathways and circuits in autism. Cell 155, 1008–1021 (2013). source

In the paper

“Although CRISPR screening coupled with single-cell transcriptomic readout provides unprecedented resolution for phenotypic characterization14–16, such approaches have not been fully explored in organoids.”

What the sources say

• Supported

  The Introduction describes Perturb-seq as combining CRISPR/Cas9 perturbations with single-cell RNA-seq to provide a rich genomic readout. The Results explain that it links each cell’s transcriptional profile to its perturbation, supporting the claim of high-resolution phenotypic characterization. The paper does not use the word “unprecedented,” but contrasts this approach with the low-content readouts of earlier pooled screens. (LLM, conf=0.91)

  cites [14] Dixit, A. et al. Perturb-seq: dissecting molecular circuits with scalable single-cell RNA profiling of pooled genetic screens. Cell 167, 1853–1866.e17 (2016). source
• ?Not checked — no full text of the sourcecites [15] Jaitin, D. A. et al. Dissecting immune circuits by linking CRISPR-pooled screens with single-cell RNA-seq. Cell 167, 1883–1896.e15 (2016). source
• Supported

  The Introduction contrasts conventional pooled screens, which average simple phenotypes across cells, with Perturb-seq, which pairs CRISPR interference with single-cell RNA sequencing to produce rich phenotypic data. It explains that this approach can distinguish responses within individual cells that bulk readouts cannot. The cited passages do not provide page numbers. (LLM, conf=0.91)

  cites [16] Adamson, B. et al. A multiplexed single-cell CRISPR screening platform enables systematic dissection of the unfolded protein response. Cell 167, 1867–1882.e21 (2016). source

In the paper

“Cortical abnormalities are a prominent feature of ASD8,20.”

What the sources say

• Supported

  The review’s “Altered fetal cortical development” section identifies multiple cortical abnormalities in individuals with ASD, including irregular lamination, ectopic cells and dendritic abnormalities. Its “Altered neural circuitry” section also describes changes in cortical connectivity. Together, these support the broad statement, while the review notes that the neuropathological studies had small cohorts. (LLM, conf=0.91)

  cites [8] Torre-Ubieta, L., de la Won, H., Stein, J. L. & Geschwind, D. H. Advancing the understanding of autism disease mechanisms through genetics. Nat. Med. 22, 345–361 (2016). source
• ?Not checked — no full text of the sourcecites [20] Chen, J. A., Peñagarikano, O., Belgard, T. G., Swarup, V. & Geschwind, D. H. The emerging picture of autism spectrum disorder: genetics and pathology. Pathology Mech. Dis. 10, 111–144 (2015). source

In the paper

“Cell type-specific alterations have been observed in patients with ASD and brain organoid disease models11,34.”

What the sources say

• Supported

  The Discussion (page unavailable) reports prior evidence of dysregulated GABAergic and glutamatergic cortical neurons in ASD patients and experimental models. The Results, under “Asynchronous development in SUV420H1 mutants” (page unavailable), identify altered proportions of specific neuronal types in cortical organoids. Together, these support the claim for patients and organoid models. (LLM, conf=0.89)

  cites [11] Paulsen, B. et al. Autism genes converge on asynchronous development of shared neuron classes. Nature 602, 268–273 (2022). source
• Partly supported

  The paper’s Abstract reports cell type-specific molecular changes in cortical tissue from patients with ASD, including changes in upper-layer excitatory neurons and microglia. The sentence also mentions brain organoid disease models, but our copy does not report findings from organoid models (searched: organoid). (LLM, conf=0.88)

  cites [34] Velmeshev, D. et al. Single-cell genomics identifies cell type–specific molecular changes in autism. Science 364, 685–689 (2019). sourceExperimental — accuracy not yet measured

In the paper

“Additionally, the effects of perturbations of ASD risk genes on cell-type abundances can sometimes be transient11,53, and therefore certain abnormalities during development may not be captured.”

What the sources say

• Supported

  In Results (‘Asynchronous development in SUV420H1 mutants’; page unavailable), the paper reports increased proportions of immature deep-layer projection neurons in one-month mutant organoids and says that this phenotype was rescued after three months. It also describes a transient GABAergic phenotype in the same cell line, supporting the claim that some effects on cell-type abundances are transient. (LLM, conf=0.93)

  cites [11] Paulsen, B. et al. Autism genes converge on asynchronous development of shared neuron classes. Nature 602, 268–273 (2022). source
• Supported

  The Results (page not specified in our copy) report a higher proportion of interneurons in CHD8-mutant cerebral organoids specifically at day 60. The Discussion describes the effect on interneuron development as early and temporally specific, supporting the claim that an ASD-risk-gene perturbation can have a transient effect on cell-type abundance. (LLM, conf=0.91)

  cites [53] Villa, C. E. et al. CHD8 haploinsufficiency links autism to transient alterations in excitatory and inhibitory trajectories. Cell Rep. 39, 110615 (2022). source

In the paper

“We find that dorsal intermediate progenitors, ventral progenitors and upper-layer excitatory neurons are among the most vulnerable cell types.”

In the paper

“Here we show that perturbation of 36 high-risk autism spectrum disorder genes related to transcriptional regulation uncovers their effects on cell fate determination.”

In the paper

“We have developed the CRISPR–human organoids–single-cell RNA sequencing (CHOOSE) system, which uses verified pairs of guide RNAs, inducible CRISPR–Cas9-based genetic disruption and single-cell transcriptomics for pooled loss-of-function screening in mosaic organoids.”

In the paper

“Broadly, we found that perturbed cells adopted cell states resembling those found in the unperturbed controls (Fig. 1f).”

In the paper

“Single-cell transcriptome profiling of cerebral organoids at 4 months revealed a large diversity of dorsal and ventral telencephalon cell populations (Fig. 1f,g and Extended Data Fig. 6a–j) (14 independent pools of organoids, 3–7 organoids each pool, 65 organoids in total, three independent batches).”

In the paper

“When comparing eCas9-induced with uninduced cells from scRNA-seq libraries, we found significantly enriched and depleted gRNAs (Fig. 2a,b) (induced cells, n = 14 independent pools of organoids, three batches; uninduced cells, n = 8 independent pools of organoids, two batches).”

In the paper

“Here, we describe the CHOOSE system which combines parallel genetic perturbations with single-cell transcriptomic readout in mosaic cerebral organoids.”

In the paper

“We identified progenitor cells with dorsal (PAX6) or ventral (ASCL1, OLIG2) origins.”

In the paper

“Notably, we found a cluster of interneurons expressing MEIS2 and high levels of PAX6 (LGE PAX6+ interneurons), a signature that resembles mouse olfactory bulb precursors that were recently reported to generate neurons redirected to white matter in primates29.”

In the paper

“RNA velocity analysis30,31 revealed developmental trajectories from neural progenitor cells to neuronal populations in both the dorsal and ventral telencephalons (Extended Data Fig. 6k).”

In the paper

“A time course analysis revealed a deviation from the initial gRNA distribution as early as day 20 (Extended Data Fig. 7b).”

In the paper

“Further analysis of our organoid dataset and a primary developing human brain dataset32 revealed cell type-specific expression patterns for several ASD genes that we targeted in our screen (Extended Data Fig. 6l,m).”

In the paper

“Perturbing members of the BRG1/BRM-associated factor (BAF) chromatin remodelling complex leads to enrichment of ventral telencephalon progenitors.”

In the paper

“We observed a heterogenous gRNA representation in eCas9-induced cells at 4 months.”

In the paper

“In addition to neuronal populations, we identified glial cell populations including astrocytes (S100B, APOE, ALDH1L1) and OPCs (OLIG2, PDGFRA) (Fig. 1f,g).”

In the paper

“Among the 36 genes, one of the most significant changes in cell type composition was identified in the context of ARID1B.”

In the paper

“We found that, similar to ASD risk genes, CHOOSE DEGs were enriched in specific TF modules (Fig. 3f and Supplementary Data 7).”

In the paper

“Using Pando41, we harnessed these multimodal measurements to infer a GRN of the developing telencephalon and extract sets of genes regulated by each transcription factor (TF) module, as well as positive and negative regulatory interactions between the TFs (Extended Data Fig. 14h,i and Supplementary Data 5).”

In the paper

“We detected 2,071 differentially expressed genes (DEGs) across all perturbations (Fig. 3a and Supplementary Data 2).”

In the paper

“Using a FACS-based approach, we further confirmed the enrichment and depletion phenotypes in individually perturbed organoids for the four genes (KMT2C, LEO1, ADNP and WAC) that showed the largest effect sizes (Extended Data Fig. 8).”

In the paper

“For 23 perturbations, we detected changes in the abundance of at least one cell type (CMH test, false discovery rate (FDR) < 0.05) (Fig. 2c).”

In the paper

“Notably, in the dorsal populations, seven perturbations caused CHCHD2 downregulation (Fig. 3a).”

In the paper

“Additionally, we observed an enrichment of v-RGCs and/or ccv-RGCs in 10 perturbations (for example, ARID1B, BCL11A and DEAF1) and an enrichment of INPs in 10 perturbations (for example, ILF2, MED13 and TCF20).”

In the paper

“Perturbation of KMT2C, for example, led to a strong enrichment of ventral cells.”

In the paper

“In the ventral cell populations, the most frequently detected DEG is the adhesion molecule gene CSMD1 (Fig. 3a,b), which is upregulated by ARID1B, CIC, MED13 and PHF3 perturbations and downregulated by LEO1 and KMT2C perturbations.”

In the paper

“We then analysed the fate transition probabilities of ventral progenitors and found that ARID1B-perturbed v-RGCs have significantly higher transition probabilities towards early OPCs than neuronal fates (Fig. 4e).”

In the paper

“Furthermore, we analysed specific neuronal subpopulations and found that L2/3 excitatory neurons were more impacted and depleted in the majority of perturbations (Fig. 2c).”

In the paper

“We found significant enrichment for a set of 40 TFs (adjusted Fisher test P < 0.01, more than twofold enrichment; for example, EOMES, OLIG1, DLX2) (Extended Data Fig. 14j), among which 14 TFs were encoded by ASD risk genes (for example, NFIA, BCL11A, MEF2C) (Fig. 3e).”

In the paper

“We ranked DEGs by detection frequency and discovered that many genes were differentially expressed in multiple perturbations (Fig. 3b).”

In the paper

“Among the 36 genes, we found that ARID1B perturbation caused one of the most significant enrichments of v-RGCs (Fig. 2c).”

In the paper

“In addition, LGE PAX6+ interneurons were depleted by PHF3 perturbation.”

In the paper

“All TF regulatory modules enriched in SFARI genes together form an ASD-associated sub-GRN (Supplementary Data 6).”

In the paper

“For 24 perturbations, we detected significant changes in the ratio of dorsal to ventral cells (Fig. 2c).”

In the paper

“Beyond neuronal cell clusters, we found that astrocytes were significantly enriched in CIC perturbation.”

In the paper

“Additionally, we observed an enrichment of v-RGCs and/or ccv-RGCs in 10 perturbations (for example, ARID1B, BCL11A and DEAF1) and an enrichment of INPs in 10 perturbations (for example, ILF2, MED13 and TCF20).”

In the paper

“We found significant enrichment for a set of 40 TFs (adjusted Fisher test P < 0.01, more than twofold enrichment; for example, EOMES, OLIG1, DLX2) (Extended Data Fig. 14j), among which 14 TFs were encoded by ASD risk genes (for example, NFIA, BCL11A, MEF2C) (Fig. 3e).”

In the paper

“Notably, the OLIG1 regulatory module was also enriched in DEGs caused by ARID1B perturbation (Extended Data Fig. 15a).”

In the paper

“Among the progenitors, we identified IPC depletion as a strong, convergent effect in 12 perturbations (Fig. 2c) (for example, CHD2, KAT2B and KMT2C).”

In the paper

“We visualized this GRN on the level of TFs using a uniform manifold approximation and projection (UMAP) embedding42, which revealed distinct TF groups active in neural progenitor cells (PAX6, GLI3), inhibitory neurons (DLX2, DLX6) and excitatory neurons (NEUROD2, NFIB, SATB2) as well as regulatory interactions between the TFs (Fig. 3e).”

In the paper

“Notably, most perturbations (21 of 24) lowered the dorsal-to-ventral ratio.”

In the paper

“Analysing the size of each clone, we found a mean average cell number per clone of 4.4 (Extended Data Fig. 4e,f).”

In the paper

“Perturbation of ARID1B increases the transition of v-RGCs to early OPCs.”

• Also stated in: § Discussion · plain

  “We discovered that loss of ARID1B leads to increased transition of ventral progenitors to early OPCs.”

In the paper

“When combining the TOP-DEGs from all perturbations, we found that they were significantly enriched in risk genes associated with ASD (SFARI database, 1,031 genes; approximately twofold enrichment, Fisher exact test P < 10−5 against genes expressed in 5% of cells as background) (Fig. 3d and Supplementary Data 2).”

In the paper

“We found that progenitors, including IPCs and INPs, as well as L2/3 excitatory neuronss, were among the most affected cell types.”

In the paper

“To explore these potential gene regulatory ‘hubs’, we first generated single-cell multiome data, including single-cell transcriptome and chromatin accessibility modalities, from 4-month-old cerebral organoids (Extended Data Fig. 14a–g).”

In the paper

“In line with our previous findings, we observed considerably increased OLIG2 +, DLX2 + and DLX2 + OLIG2 + cells in 40-day-old organoids from both patients compared with control organoids (Fig. 4f,g).”

In the paper

“Among the ventral telencephalon cells, we have identified strong enrichment of v-RGCs, ccv-RGCs and INPs, which are progenitors that differentiate into interneurons and oligodendrocytes44,51.”

In the paper

“IPCs, transit-amplifying dorsal progenitors that generate neurons for all cortical layers and contribute to the evolutionary expansion of the human cortex49,50, appeared to be particularly susceptible to ASD genetic perturbations.”

In the paper

“The OLIG1 module is particularly interesting as many of its downstream targets are ASD risk genes.”

In the paper

“Importantly, perturbations of three BAF complex members (ARID1B, BCL11A and SMARRC2) all lead to enrichment of v-RGCs”

In the paper

“In our screen, we assessed 19 ASD genes known to be involved in epigenetic regulation. Despite their broad role in cell differentiation, perturbations of these genes impacted specific cell types and biological processes during brain development.”

In the paper

“Furthermore, our findings indicate that L2/3 excitatory neurons are more vulnerable than other neuronal populations to ASD perturbations.”

In the paper

“By employing a pooled CRISPR screening system in conjunction with validation, our study provides a developmental and cell type-specific phenotypic database for ASD gene loss-of-function research.”

In the paper

“DEG and gene ontology enrichment analyses revealed both common and distinct molecular processes impaired in different perturbations”

In the paper

“To ensure that lentiviral integration frequency was limited to one per cell, weused a low infection rate of 2.5% (ref. 18)”

What the sources say

• ?Not checked — no full text of the sourcecites [18] Doench, J. G. Am I ready for CRISPR? A user’s guide to genetic screens. Nat. Rev. Genet. 19, 67–80 (2018). source

In the paper

“It is likely that this could limit the mutant cell–cell interactions within the mosaic tissues.”

In the paper

“In summary, our scRNA-seq dataset of 4-month-old cerebral organoids recapitulates diverse telencephalic cell types that are present in the developing human brain.”

In the paper

“Immunohistochemistry analysis of several perturbations in organoids further confirmed the loss of protein products for the majority of selected gRNAs (Extended Data Fig. 2).”

In the paper

“Altogether, we established a highly efficient and controlled pooled screening system with high clonal complexity in the organoid.”

In the paper

“Using our reporter assay, we selected efficient sgRNA pairs for 36 ASD genes (Extended Data Fig. 1d,e and Supplementary Table 1).”

In the paper

“Notably, some of the ASD-associated TF modules were among the most strongly enriched in CHOOSE DEGs, supporting their role in ASD-associated gene dysregulation (Fig. 3f).”

In the paper

“These data suggest that cells captured in the CHOOSE screen came from diverse and relatively small clones, which are crucial for mitigating dominant clonal effects.”

In the paper

“Furthermore, our data indicate that ASD pathology could emerge as early as the neural progenitor stage.”

In the paper

“suggesting that certain biological processes and regulatory programs might be specifically relevant to ASD-associated gene perturbations.”

In the paper

“These data indicate an interneuron subtype-specific response to ASD genetic perturbations.”

In the paper

“In conclusion, using a pooled, high-complexity barcoding screening system, we successfully identified ASD risk genes that play essential roles in cell proliferation and survival.”

In the paper

“Collectively, the CHOOSE system allowed us to simultaneously investigate the effects of multiple ASD genes on cell fate determination.”

In the paper

“The enrichment of OLIG1 and EOMES regulomes suggests potentially vulnerable cell fate specification-related regulatory networks upon ASD genetic perturbations.”

In the paper

“We observed that most effects detected at the single-cell level were also largely consistent across organoids grown from different batches, further supporting the robustness and reproducibility of our system.”

In the paper

“This module was previously identified as crucial for oligodendrocyte differentiation in the developing human cortex22, highlighting the involvement of the oligodendrocytes in ASD pathophysiology.”

What the sources say

• ?Not checked — no full text of the sourcecites [22] Trevino, A. E. et al. Chromatin and gene-regulatory dynamics of the developing human cerebral cortex at single-cell resolution. Cell 184, 5053–5069.e23 (2021). source

In the paper

“indicating the critical role of the BAF complex in regulating ventral telencephalon cell fate specification.”

In the paper

“The enlarged volume of GE in the patient with an ARID1B mutation is consistent with these observations.”

In the paper

“highlighting the involvement of the oligodendrocytes in ASD pathophysiology.”

In the paper

“The combination of high-content perturbation screening and validation in a patient-specific context exemplifies the effectiveness of employing organoid systems to study NDDs.”