High-throughput single-cell CRISPRi screens stratify neurodevelopmental functions of schizophrenia-associated genes

Why we highlight this preprint

This work makes excellent use of precise edits in single cells to capture the big picture of a complex psychiatric disease affecting millions, effectively bridging technological advances with new insights into schizophrenia. We appreciate the robustness of the results and the way the paper is presented as a well-written, engaging narrative. Moreover, the study reinforces the hypothesis that schizophrenia has a neurodevelopmental origin, while also providing compelling evidence for the convergence of diverse genetic traits on shared phenotypic outcomes. Importantly, the authors uncover new molecular mechanisms that are not only relevant to schizophrenia but may also inform strategies to refine in vitro protocols, enabling the acceleration or delay of neural differentiation through targeted genetic manipulation.

Introduction

Schizophrenia (SCZ) is a severe mental illness affecting approximately 24 million people worldwide, or about 1 in 300 individuals. It typically manifests in young adulthood with episodes of psychosis, hallucinations, and profound cognitive impairments that last throughout life. Although SCZ has a strong heritable component (~80%), the molecular mechanisms through which genetic variations confer risk remain largely unresolved.

Schizophrenia has neurodevelopmental roots which are hard to study in affected adult patients. Indeed, neurodevelopment is a highly complex process, involving the coordinated progression of multiple cell types across extended periods of pre- and post-natal brain development. Gene mutations can exert cell type–specific effects, leading to abnormal phenotypes. Luckily, thanks to the use of pluripotent stem cells, scientists can now reenact human prenatal developmental processes in a dish, unravelling molecular and cellular processes impossible to study otherwise.

In this study, the authors present a compelling use case of single-cell CRISPR screening in stem cell–derived neural models. This enables the identification of transcriptional programs underlying defects in neural differentiation and maturation, which may contribute to SCZ pathogenesis. Furthermore, they identified phenotypic convergence of distinct genetic mutations, with different perturbations yielding similar transcriptional and cellular outcomes. These findings reinforce the neurodevelopmental basis of SCZ and highlight molecular convergence as a unifying feature of diverse genetic risk factors in neurodevelopmental disorders.

Main results

This story begins with a systematic literature mining effort that identified 65 genes whose loss-of-function mutations are linked to schizophrenia. These hits were validated by looking at post-mortem gene expression data derived from post-mortem brain tissue of SCZ patients. To assess the impact of reduced expression of these genes during brain development, Yildiz and colleagues designed a pooled CRISPR knockdown screen in pluripotent stem cells, directing their differentiation in parallel toward either neural progenitors or excitatory neurons. Although this workflow may appear complex, its logic is straightforward: scientists can generate “villages” of cells, each carrying a perturbation in one of the 65 schizophrenia-associated genes. In a single experimental procedure, this approach enables simultaneous interrogation of multiple genetic perturbations at single-cell resolution, uncovering how genetic predisposition shapes differentiation trajectories (Fig. 1).

Figure 1 – Graphical representation of the pooled CRISPR interference (CRISPRi) knockdown screen. A library of sgRNAs is introduced into human induced pluripotent stem cells expressing dCas9, followed by differentiation into NPCs or neurons whose transcriptome, along with the sgRNA, is profiled at the single cell level. Figure generated with a licensed version of BioRender.

After rigorous quality control and validation, the authors sequenced RNA and chromatin accessibility of individual neural progenitors and excitatory neurons, pairing each readout with its corresponding genetic perturbation. This allowed them to directly connect gene knockdown with transcriptional and epigenomic changes.

The results revealed that knockdown of a set of 15 genes consistently delayed neuronal differentiation, highlighting a striking convergence of independent genetic traits on shared phenotypic outcomes. This aligns with mounting evidence that patients with the same neurodevelopmental disorder often carry distinct genetic risk factors that converge on common pathways.

Interestingly, however, neural progenitors with reduced expression of MCRS1 exhibited the opposite behavior: they differentiated faster and more closely resembled neurons than all their companions. This unexpected finding prompted further investigation of MCRS1’s role in early development. Through a series of complementary experiments, Yildiz and colleagues discovered two complementary transcription factors implicated in MCRS1 signaling; ZEB1 and TCF4, which respectively accelerate and delay stem cell differentiation towards the neural lineage (Fig.2). This molecular mechanism could form the basis for explaining developmental abnormalities in a subset of patients with MCRS1 mutations who eventually develop the disease.

Figure 2 – Schematic summary of how loss-of-function schizophrenia (SZ) risk genes alter neural stem/progenitor cell fate, causing either neurodevelopmental delay (e.g., TCF4) or accelerated neurodevelopment (e.g., ZEB1). Preprint figure 6h made available under a CC-BY-NC-ND 4.0 International license.

Questions and future directions

• Have you considered examining the expression or accessibility levels of MCRS1, TCF4, and ZEB1 in patient datasets, such as PsychENCODE, particularly in individuals with MCRS1 loss-of-function mutations?
• What if this work would have been carried out in organoids? Which aspects do you think would be recapitulated, and what might differ? Have you considered this approach?
• Do you think MCRS1 knockdown could be used to induce stem cell differentiation in a way similar to NGN2 overexpression, or would its broad cellular effects make this approach less feasible?
• We’re curious about how you organized your experiments and ideas while developing this paper. Were all your validation experiments included in your original plan, or did you design some later to strengthen your initial results? In your view, what are the best practices for conducting validation experiments?

Tags: crispri, ipsc, ngn2, schizophrenia

doi: https://doi.org/10.1242/prelights.41851

Read preprint (1 votes)