The molecular landscape of neural differentiation in the developing Drosophila brain revealed by targeted scRNA-seq and a multi-informatic analysis paradigm

Nigel S. Michki, Ye Li, Kayvon Sanjasaz, Yimeng Zhao, Fred Shen, Logan A. Walker, Cheng‐Yu Lee, Dawen CaibioRxiv 2020 · DOI

TL;DR

This study leverages targeted single-cell RNA sequencing (scRNA-seq) and a novel bioinformatics tool (MiCV) to map the molecular diversity of neural differentiation in Drosophila type-II neuroblast lineages. By integrating computational analysis with experimental validation, the researchers identify new molecular markers that define distinct neural subtypes and propose a model for how these factors orchestrate complex neurogenesis.

What Problem Does This Paper Address?

Understanding how a diverse array of neurons and glia are generated from a limited number of stem cells is a fundamental challenge in developmental biology. While the Drosophila type-II neuroblast (NB) system offers a simplified yet evolutionarily conserved model for neurogenesis (similar to the human outer subventricular zone), dissecting the molecular mechanisms that drive this process has been difficult. Specifically, untangling the interplay of intrinsic cell properties (lineage identity, division history) and extrinsic cues (spatial information) that lead to distinct neuronal fates requires sophisticated analytical approaches. Traditional methods often struggle to capture the full spectrum of molecular diversity, especially when dealing with complex, multi-dimensional differentiation processes.

Methods

The researchers employed a multi-informatic approach combining experimental techniques and computational analysis.

  • Targeted scRNA-seq: They used a fluorescent reporter (UAS-hH2B::2xmNG) to specifically label type-II NB progeny, allowing for their isolation via fluorescence-activated cell sorting (FACS) from a mixed population of brain cells. This targeted approach minimizes noise from non-relevant cell types.
  • MiCV (Multi-informatic Cellular Visualization): A novel web-based tool was developed to integrate results from various bioinformatics analyses. MiCV allows for simultaneous visualization of co-expression patterns of multiple genes within a 2D UMAP projection, facilitating the exploration of gene function and ortholog annotations.
  • Pseudotime Analysis (Palantir): This algorithm was used to order cells along a continuous differentiation axis, effectively reconstructing developmental trajectories from progenitor to mature neuron states.
  • Clustering Analysis (Leiden algorithm): Standard clustering techniques were applied to identify distinct cell populations within the scRNA-seq data.
  • In Situ Hybridization (HCRv3): This technique was used to validate the spatial expression patterns of identified marker genes within the Drosophila brain, providing crucial spatial context often missing in scRNA-seq data alone.
  • Logistic Regression-based Marker Gene Analysis: This method was employed to identify genes that distinguish between different cell populations.

Datasets

  • scRNA-seq data: Transcriptomes from approximately 3942 type-II neuroblast-derived cells were generated using 10X Chromium technology. This dataset represents a targeted capture of neural progenitors and their progeny.
  • Existing literature and databases: The study extensively references and integrates data from previous publications and databases for gene annotations, known marker genes, and functional information.

Key Insights and Ideas

  • Type-II NB lineages are molecularly diverse: Even within the relatively small number of type-II NB lineages, there is significant molecular heterogeneity, suggesting complex fate specification mechanisms.
  • A multi-axis model of differentiation: The study proposes that neural fate is determined by the interplay of three key axes: differentiation state, progenitor cell division number, and progenitor lineage identity.
  • Novel markers for neural subtypes: The research identified several new genes (e.g., Sp1, TfAP-2, Fas3, E(spl)m6-BFM, tap, ytr) that specifically label distinct subsets of INPs, GMCs, and neurons, offering new tools for studying neurogenesis.
  • Importance of spatial context: The integration of in situ hybridization with scRNA-seq data was crucial for understanding the spatial organization of gene expression and validating computational findings.
  • MiCV as a powerful visualization tool: The developed MiCV tool significantly aids in exploring complex scRNA-seq datasets, enabling researchers to visualize multi-gene co-expression and integrate various analytical outputs.

Key Results

  • The study successfully filtered and analyzed scRNA-seq data from approximately 3942 type-II neuroblast-derived cells.
  • Pseudotime analysis revealed distinct temporal expression patterns for known canonical markers (CycE, dap, Hey, nSyb) that align with differentiation stages.
  • Sub-clustering of INPs and GMCs identified three distinct subclusters for each, suggesting a finer granularity of cell states than initially apparent.
  • The transcription factors Sp1, TfAP-2, and Fas3 were identified as key markers for specific INP populations and their potential GMC progeny, correlating with different lineage identities and division histories.
  • In situ validation confirmed the expression patterns of these novel markers, demonstrating their utility in distinguishing neural subtypes and their spatial localization within specific lineages.
  • The study proposes that transcription factors and surface molecules, rather than genes involved in terminal neuronal functions, are the primary determinants of neural subtype identity at the late 3rd instar larval stage.

Limitations and Future Work

  • Single Time Point Analysis: The current study focuses on a single developmental time point (late 3rd instar larvae). The authors acknowledge that repeating scRNA-seq experiments at earlier and later stages would provide a more complete picture of neurogenesis over time.
  • Deconvoluting Axes: While the study identifies factors associated with lineage identity, division number, and differentiation state, fully disentangling the contributions of each axis remains a challenge.
  • Functional Validation: While novel markers were identified and their expression patterns characterized, their precise functional roles in neural differentiation require further investigation through gene manipulation studies.
  • Broader Lineage and Temporal Coverage: The study focused on type-II neuroblasts. Expanding to other neuroblast types and incorporating lineage tracing technologies (like Bitbow) would provide a more comprehensive understanding of the entire brain's neurogenesis.
  • MiCV Development: The authors suggest that further development of MiCV could enhance its utility for biologists by integrating more advanced analytical features and making it even more user-friendly.

Summary updated

Abstract

SUMMARY The Drosophila type-II neuroblast (NB) lineages present an attractive model to investigate the neural differentiation process. With only 16 stem cells, the type-II NB lineages generate many intermediate neural progenitors (INPs) to rapidly expand the neuron and glia pool, similar to those in the human outer subventricular zone (OSVZ). We performed targeted single-cell mRNA sequencing (scRNA-seq) in 3rd instar larval brains and created MiCV, an scRNA-seq data visualization web tool to integrate results from multiple bioinformatics analyses, display co-expression patterns of multiple genes simultaneously, and retrieve gene function and ortholog annotations. We identified novel markers that label distinct neural subsets using MiCV and subsequently in situ profiled them to recover the spatial information lacking in the scRNA-seq data. These new markers further enabled us to build novel neural developmental trajectories that lead to unique neuronal cell fates. Combining prior knowledge, in silico analyses, and in situ evidence, this multi-informatic investigation describes the molecular landscape of neural differentiation from a single developmental snapshot in Drosophila , and provides an experimental and analytical roadmap for navigating the differentiation process of more complex brains.
BibTeX
@misc{michki2020molecular,
  author = {Nigel S. Michki and Ye Li and Kayvon Sanjasaz and Yimeng Zhao and Fred Shen and Logan A. Walker and Cheng‐Yu Lee and Dawen Cai},
  title = {{The molecular landscape of neural differentiation in the developing <i>Drosophila</i> brain revealed by targeted scRNA-seq and a multi-informatic analysis paradigm}},
  year = {2020},
  doi = {10.1101/2020.07.02.184549},
  howpublished = {bioRxiv},
}
Figure