TL;DR
This study reveals that clonally related excitatory neurons in the developing mouse neocortex, originating from the same progenitor cells, are not uniformly connected. Instead, they preferentially form vertical connections across cortical layers, while lateral connections within the same layer are not enhanced. This suggests a specific developmental program for circuit assembly, integrating intra-clonal vertical input with inter-clonal lateral input.
What Problem Does This Paper Address?
The neocortex, responsible for complex cognitive functions, is wired by billions of neurons. A long-standing hypothesis suggests that neurons born from the same progenitor (clonally related) form "ontogenetic columns" that act as fundamental computational units. However, the precise connectivity patterns and cell type diversity within these clonal units, especially in the developing brain, remained unclear. Previous studies hinted at increased connectivity among related neurons, but a detailed understanding of how this lineage-based connectivity is organized across cortical layers and within them was lacking. This paper aims to characterize the cell type composition and circuit organization of these clonally related excitatory neurons to test the "ontogenetic column" hypothesis and understand the developmental basis of cortical circuits.
Methods
The researchers employed a sophisticated lineage tracing strategy using a tamoxifen-inducible Cre-lox system driven by the Nestin promoter. This allowed for sparse labeling of progenitor cells at embryonic day 10.5 (E10.5), leading to the formation of "translaminar clones" spanning multiple cortical layers.
Key techniques included:
- Tamoxifen-inducible Cre-lox system: To label progenitor cells and their progeny.
- Immunohistochemistry: To identify progenitor types (Pax6, Tbr2) and visualize tdTomato-labeled clones.
- Patch-seq: A combined electrophysiological recording and single-cell RNA sequencing method. This allowed for simultaneous characterization of neuronal properties and gene expression profiles.
- Multi-cell patch clamp recordings: To directly measure synaptic connections between pairs of neurons, both clonally related and unrelated.
- Computational analysis: Including t-distributed stochastic neighbor embedding (t-SNE) for dimensionality reduction of transcriptomic data, generalized linear models (GLMs) for predicting cell type and region, and statistical modeling to estimate connectivity probabilities and input fractions.
Datasets
- Single-cell RNA-sequencing data: Generated from 206 neurons (both clonally related and unrelated) from juvenile mouse neocortex (P15-P20) using the Patch-seq method. This data was deposited in GEO under accession code GSE140946.
- Reference transcriptomic dataset: A previously published cell type atlas of adult mouse cortex (Tasic et al., 2018) was used for mapping the transcriptomic profiles of the newly generated single-cell data.
Key Insights and Ideas
- Clonal diversity is the norm: Contrary to a strictly deterministic view, progenitors labeled early (E10.5) give rise to a diverse range of excitatory neuron types, not a restricted subset. This challenges the idea of rigid fate restriction at this stage.
- Vertical connections are lineage-biased, lateral ones are not: This is the central finding. While related neurons are more likely to connect vertically across layers, they don't show a preference for connecting laterally within the same layer. This suggests a specific developmental wiring program.
- Integration of intra-clonal and inter-clonal inputs: The findings propose a model where vertical inputs from clonally related neurons are integrated with lateral inputs from unrelated neurons, forming a fundamental circuit motif.
- Nestin-CreER labels early progenitors: The study confirms that Nestin-CreER labeling at E10.5 primarily labels individual radial glia, which are the progenitors of excitatory neurons.
Key Results
- Clone composition: Translaminar clones labeled at E10.5 were derived from an average of two radial glia and spanned cortical layers 2-6. They contained a diverse ensemble of transcriptomic cell types, similar to unlabeled neurons.
- Connectivity:
- Overall, there was no significant difference in connection probability between related and unrelated pairs.
- However, vertical connections (across layers) were significantly more likely between related neurons (6.0% vs. 2.7% for unrelated, p=0.0056). This was particularly pronounced for connections from L4 to L5 and L2/3 to L5.
- Lateral connections (within the same layer) showed no increase in probability between related neurons (5.6% vs. 8.1% for unrelated, p=0.25). In fact, there was a trend towards lower lateral connectivity among related neurons in L2/3 and L5.
- Cell type mapping: Neurons from the labeled clones mapped to nearly all excitatory cell types found in the adult mouse cortex atlas, with only a slight underrepresentation of one specific L2/3 IT subtype.
Limitations and Future Work
- Progenitor pool bias: The study acknowledges that tamoxifen induction might preferentially label progenitors with higher Nestin promoter activity, potentially biasing the sampled clones.
- Temporal window: The study focused on juvenile mice (P15-P20). Transient increases in lateral connectivity observed in other studies at earlier developmental stages (P13-P16) might have been missed.
- Layer 6 exclusion: Neurons in layer 6 were not included due to technical challenges and their distinct migratory behavior, limiting the complete picture of translaminar clone connectivity.
- Regional specificity: The study focused on primary sensory areas (V1 and SS1). Generalizability to other cortical regions, like primary motor cortex, needs further investigation.
- Future directions: The authors suggest exploring the functional role of lineage-associated synaptic connectivity in vivo, investigating the molecular mechanisms underlying selective vertical connectivity (e.g., gap junction coupling, cell adhesion molecules), and examining if lateral connectivity is actually decreased between related neurons. Further studies using different lineage tracing methods, adult animals, and other brain regions are needed to assess the generalizability of these findings.
Summary updated