Fig 1.
Patch-seq of microglia: Method development and data quality analysis.
A Confocal Image of a brain slice generated from a MacGreen mouse as used in this study, showing IBA+ microglial cells (immunolabeled as described in Matyash et al 2017)). Investigated regions- cortex (Cx), hippocampus (HC) and corpus callosum (CC) are marked. B Experimental scheme. (Upper image) Microglial cells (MG) were identified by their EGFP flurorescence (small fluorescence image on the left) and approached with the patch-clamp pipette to establish the whole cell configuration (phase contrast image, patched). A series of de- and hyperpolarizing voltage steps ranging from -170 and +60 mV were recorded from each cell as shown on the right. (Middle images) The cytoplasm was harvested by suction and subsequently the pipette was withdrawn (Eppendorf tube created with BioRender.com). In some cases, the soma remained attached as shown in the left image. Extracellular controls (EC) were generated by moving a patch pipette 20 μm to 50 μm below the slice surface without approaching a cell. (Lower image) The content of the patch pipettes was expelled into a tube and Smart-seq protocol was immediately applied. Number of MG samples (137 patched, 32 only picked from cell-attached mode, 55 Extracellular controls) are indicated. C. Histograms depicting the number of genes (x axis) obtained in MG (top) and EC (bottom) samples (y axis). D Heatmap displaying the expression of selected microglial (MG), neuronal (N), astrocytic (A), oligodendrocytic (O) marker genes as well as some housekeeping genes (HK). Brain regions are indicated on the right: cortex (Cx), hippocampus (HC) and corpus callosum (CC). The samples collected for RNA assessment were either microglial cells which were patched (Microglia patched), just picked while the electrophysiological profile was not recorded (picked), or extracellular controls.
Fig 2.
Electrophysiological data obtained before harvesting the cytoplasm for sequencing.
A EGFP Fluorescence (top) and brightfield image (BF, bottom) of a microglial cell approached by a patch pipette. B (Inset) Samples of membrane currents from microglial cells clamped for 50 ms to a series of de- and hyperpolarizing voltages ranging from -170 mV to +60 mV from a holding potential of -70 mV. Cells were recorded in cortex, hippocampus and corpus callosum as indicated. (Main) Based on the recorded membrane currents, average current-voltage relationships for all patch-clamped cells in the notified region were constructed. C Summary of series resistances (Rs; left), capacities (Cm; middle left), membrane resistances (Rm; middle right) and reversal potentials (Vrev; right) of microglia from cortex (Cx), corpus callosum (CC) and hippocampus (HC). Significant differences are indicated by *** (p<0.001) and * (p = 0.029), non-significance is indicated by n.s. Significance was tested by Kruskal-Wallis followed by a Dunn´s test. P-values were as followed: Rs: Cx-CC:>0.9999, Cx-HC:>0.9999, CC-HC:>0.9999.; Cm: Cx-CC:>0.9999, Cx-HC:0.0402, CC-HC:<0.0001; Rm: Cx-CC:0.0654, Cx-HC:0.0088, CC-HC:>0.9999; Vrev: Cx-CC:0.2137, Cx-HC:>0.9999, CC-HC:0.0022.
Fig 3.
UMAP representation of MG and EC samples.
A Scheme depicting the filtering and computational pre-processing applied to the data set. B UMAP plot of MG and EC samples on top 80% variable genes. Clustering analysis revealed two clusters which are indicated by pink (Cluster 1) and purple (Cluster2) colors. As seen in the bar plot the majority of microglia samples are enriched in cluster 2, while the majority of EC samples in cluster 1. In the bar plot the y axis represents proportions of samples per cluster. Numbers on the plot indicate the number of samples in each cluster (EC (extracellular control, gray), MG (microglia) in Cx (cortex, green), CC (corpus callosum, pink), HC (hippocampus, blue).).C Expression levels of the microglia marker genes Ctss, Cx3cr1, Tmem119, P2ry12, Fcrls, Tyrobp, Siglech and C1qc. Note that all of these genes exert higher expression levels in Cluster 2.
Fig 4.
Comparison of single cell transcriptomes obtained by patch seq and after FACS sorting.
A Left, Comparison of microglia single cell data from the current patch-seq and a previously published data set after FACS isolation (Masuda et al., 2019). The latter data were obtained from microglia after brain dissociation and FACS. Note that clusters 1 and 4 appear in both data sets whereas clusters 2 and 3 are specific for FACS samples. Right, Gene expression analysis of the data sets shown in A for the microglia activation markers Cdh23, Myd88 as well as for the homeostatic genes Ctss, Cx3cr1 and Tmem119. B Gene ontology (GO) enrichment analysis of 275 top marker-genes of a FACS specific Cluster 2 reveals Cluster specific pathways. The bar chart represents the top 5 significantly enriched pathways in Cluster 2 in comparison to all other clusters. X-axis shows the fold enrichment of each pathway. C Detection of Ctss and Cdh23 in cortical brain slices from MacGreen mice (top) and FACS-isolated microglia from Bl6 mice (bottom) by RNAscope and confocal microscopy. Ctss (left) and Cdh23 (middle left) are indicated in red and cyan, respectively. Intrinsic EGFP signal of cell in slices is given in green (middle right); FACS isolated microglia in lower image show autofluorescence (auto). Merged images (right) show overlap of signals, indicated by arrows. The small square in each image is shown enlarged in the insets on the top left of each image. Scale bars: 50 μm. D Quantification of Ctss and Cdh23 in microglial cells in situ and after FACS isolation. Note that the homeostatic microglial marker Ctss (left) was expressed in both, in situ and in vitro microglial cells whereas Cdh23 (middle) was significantly more present after FACS sorting. Significane was tested by Mann-Whitney (p < 0.0001).