Single-Nucleus RNA Sequencing of Cortical Tissue in Microglial Bin1-Deficient Mice under Neuroinflammation
Description
Bridging Integrator 1 (BIN1) is a major genetic risk factor for Alzheimer's disease (AD) and is highly expressed in microglia. However, its exact role in regulating microglial activation and neuroinflammation remains fully elucidated. We hypothesized that microglial-specific deficiency of BIN1 alters the baseline transcriptomic landscape of microglia and exacerbates or aberrantly modulates the cortical inflammatory response during acute neuroinflammation. To test this, we used a conditional knockout mouse model (Bin1cKO) targeting Bin1 specifically in microglia, subjected to lipopolysaccharide (LPS)-induced systemic inflammation to model acute neuroinflammation. This dataset consists of single-nucleus RNA sequencing (snRNA-seq) data derived from the cerebral cortex of adult mice. Experimental Groups: Control (WT) + Vehicle; Control (WT) + LPS; Microglial Bin1 Conditional Knockout (Bin1cKO) + Vehicle; Microglial Bin1 Conditional Knockout (Bin1cKO) + LPS Tissue Processing & Sequencing: Cortical tissues were harvested from adult mice [Insert time, e.g., 24 hours] following intraperitoneal administration of LPS [Insert dosage, e.g., 1 mg/kg] or saline vehicle. Nuclei were isolated from frozen mouse cortex, multiplexed using antibody hashtags, and FACS-sorted into neuronal (NeuN+) and non-neuronal (NeuN-) fractions to enrich for glial cells. These fractions were remixed at a 1:4 ratio, pooled across samples, and processed using 10x Genomics v3 chemistry for sequencing on an Illumina NovaSeq 6000. Data Contents: This repository includes processed count and metadata table. The snRNA-seq data provides a high-resolution, cell-type-specific transcriptomic map of the cortex under Bin1 deficiency and inflammatory challenge. Notable findings from this dataset show: 1) Successful identification and clustering of major cortical cell populations, including microglia, astrocytes, oligodendrocytes, oligodendrocyte precursor cells (OPCs), excitatory neurons, and inhibitory neurons. 2) Marked transcriptomic alterations specifically within the microglial clusters following Bin1cKO, both at baseline (vehicle) and aggravated phenotypes upon LPS administration. 3) Robust upregulation of pro-inflammatory gene networks in Bin1cKO microglia exposed to LPS compared to controls. 4) Secondary transcriptomic shifts in neighboring non-myeloid cortical cells (e.g., astrocytes and neurons), demonstrating how altered microglial state cross-talks with the broader cortical microenvironment. These data can be interpreted as genetic evidence that microglial BIN1 acts as a critical negative regulator of the cortical inflammatory response. Loss of microglial BIN1 "primes" or hyper-sensitizes microglia, leading to an exaggerated, hyper-inflammatory state when confronted with an inflammatory trigger (LPS).
Files
Steps to reproduce
To identify cell clusters, the Seurat FindClusters function was applied to Harmony outputs (resolution: 0.5). Major cell types were annotated using specific markers, microglial states were classified via CellID, and gene ontology analysis was performed using gprofiler. To evaluate cell type-specific transcriptomic changes and see if Bin1 deficiency alters the microglial response to lipopolysaccharide (LPS), pseudo-bulk differential expression analysis was performed via DESeq2. The dataset was stratified by major cell types from the Seurat object, and raw transcript count matrices were extracted from the RNA assay layers. To ensure statistical power, genes were filtered to retain those with a count > 1 in at least 5 cells within that cell type; cell types with fewer than 50 passing genes were excluded. Statistical evaluation involved pairwise comparisons and a specialized interaction analysis. Pairwise Wald tests evaluated four primary contrasts: baseline LPS response (Bin1fl/fl + LPS vs. Bin1fl/fl), homeostatic Bin1 deficiency (Bin1cKO vs. Bin1fl/fl), LPS response under Bin1 deficiency (Bin1cKO + LPS vs. Bin1cKO), and the effect of Bin1 deficiency on the LPS response (Bin1cKO + LPS vs. Bin1fl/fl + LPS). Concurrently, to evaluate synergistic, non-additive transcriptomic effects between genotype and inflammation in microglia, a fully factorial generalized linear model (GLM) was assessed. Metadata entries were decoupled into Genotype (Control vs. Bin1cKO) and Treatment (PBS vs. LPS) and modeled using a negative binomial distribution: Design = ~ Genotype + Treatment + Genotype x Treatment. The Treatment effect captures the baseline LPS response in controls; the Genotype effect captures baseline differences from Bin1 loss under PBS; and the interaction term tests the "difference of differences" to identify genes whose LPS response is altered by microglial BIN1 loss. Size factors were estimated using the poscounts estimator to handle sparse matrices, dispersions were modeled with a local fit type, and hypothesis testing utilized the Wald framework. To ensure biological robustness and mitigate artificial log-fold-change inflation from low-count genes or high single-nucleus variance, log2 fold-changes and standard errors for all effects, contrasts, and the interaction term were regularized post-hoc using the Adaptive Shrinkage (ashr) framework. Significant differentially expressed genes (DEGs) were defined by a Benjamini-Hochberg adjusted p-value (p_adj < 0.05) and an absolute shrunken log2 fold-change > 0.25.