A fluorescent perilipin 2 knock-in mouse model reveals a high abundance of lipid droplets in the developing and adult brain

Lipid droplets (LDs) are dynamic lipid storage organelles. They are tightly linked to metabolism and can exert protective functions, making them important players in health and disease. Most LD studies in vivo rely on staining methods, providing only a snapshot. We therefore developed a LD-reporter mouse by labelling the endogenous LD coat protein perilipin 2 (PLIN2) with tdTomato, enabling staining-free fluorescent LD visualisation in living and fixed tissues and cells. Here we validate this model under standard and high-fat diet conditions and demonstrate that LDs are highly abundant in various cell types in the healthy brain, including neurons, astrocytes, ependymal cells, neural stem/progenitor cells and microglia. Furthermore, we also show that LDs are abundant during brain development and can be visualized using live imaging of embryonic slices. Taken together, our tdTom-Plin2 mouse serves as a novel tool to study LDs and their dynamics under both physiological and diseased conditions in all tissues expressing Plin2.

Supplementary Western blot analysis shows that PLIN3 is not altered in NSPCs from the tdTom-Plin2 mice compared to Ctrl NSPCs (n=3 samples per condition, Blots repeated 3 times with similar outcome, mean +/-SEM, uncropped blots in Suppl.Figure8).b and c) Proteomics analysis of tdTom-Plin2 NSPCs and Ctrl NSPCs shows that 99.52% of the proteins are unchanged.The changed proteins are not related to lipid metabolism (n=4 samples per condition).d) Only PLIN2 is sufficiently present to be detected by proteomics, underlining its importance for NSPCs.PLIN2 levels are the same in Ctrl and tdTom-Plin2 NSCPs.Shown is a heatmap of the median values of the log2 quantity (n=4 samples per condition).e) Schematic representation of the histological evaluation (n=3 mice per genotype), performed by a certified pathologist.f) Overview of the organs and tissues analyzed."No difference" indicates that tissue from the two genotypes could not be distinguished.g and h) Representative images of hematoxylin and eosin (H&E) stained sections of organs known to contain LDs, such as adipose tissue, intestine, heart, muscle, and liver.The histological evaluation did not reveal any differences between Ctrl and tdTom-Plin2 mice.i) Overview and high magnification images of Oil red O (ORO) stained liver sections.Periportal (PP) and centrilobular (CL) regions are shown.
and c) Serum analysis of Ctrl and tdTom-Plin2 mice on SD or HFD show a significant increase in cholesterol, low-density lipoprotein (LDL) and high-density lipoprotein (HDL) after HFD.(n=5 mice per group, mean +/-SEM).d) Free fatty acids in the serum were not significantly changed.e and f) Blood levels of alanine aminotransferase (ALAT) and aspartate aminotransferase (ASAT), which are used as indicators of liver disease, were not significantly changed with HFD, even though tdTom-Plin2 mice on HFD had a slight elevation in both.g and h) Western blot analysis of tdTom-Plin2 liver homogenates from SD and HFD fed mice.The increase in PLIN2 protein can be seen in both the tagged and untagged version, revealed by an antibody against PLIN2, or by detecting only the tagged version using an anti tdTomato (RFP) antibody (n=4 mice per condition, mean +SEM).Asterisks indicate the following pvalues: *< 0.05.**< 0.01 ns= non-significant.
Analysis of mRNA expression by RT-qPCR show no significant difference in the lipases Atgl, Hsl and Mgl in the brain of Ctrl and tdTom-Plin2 mice.(n=4 Ctrl and 5 tdTom-Plin2 male mice, measured in duplicates, fold change +/-SEM).b) There are no significant differences in brain TAGs, DAGs and MAGs, measured by lipidomics analysis, between Ctrl and tdTom-Plin2 mice.(n=5 mice per group, mean +/-SEM).c and d) FACS gating strategy for Thy1GFP positive neurons and Aldh1l1GFP positive astrocytes.All samples were first selected on viability based on DAPI and RedDOT staining, followed by sorting based on size and granularity, exclusion of doublets to have a population of viable single cells.These were then sorted based on GFP (cell marker) and tdTomato (tdTom-Plin2) to quantify how many cells have LDs in the different cell populations of the brain.

a
Feature plots of commonly used microglia marker genes c Comparison with summarised disease signatures (Chen and Colonna, J. Exp.Med.2021) d Comparison with LDAM-signature (Marschallinger et al.Nat.Neuro.2021)

a
Feature plots of commonly used microglia marker genes shows the microglial nature of the cells analysed.b) Quality control of the scRNA seq data, showing number of Unique Molecular Identifiers (UMIs), detected gene number, and mitochondrial gene percentage, indicative of good data quality for cells from adult mice.c) Comparison of the gene signature of the tdTom-Plin2 enriched cluster 2 with the summarised disease signatures published by Chen and Colonna (Chen and Colonna, 2021).d) Comparison of the gene signature of the tdTom-Plin2 enriched cluster 2 with the LDAM signature published by Marschallinger and colleagues (Marschallinger et al., 2020).and b) tdTom-PLIN2 reveals that LDs are abundant the DG and SVZ of both 1 week and 3 week old mice.Representative images of non-stained sections showing NesGFP positive NSPCs and tdTom-PLIN2 positive LDs in the DG and SVZ.(maximum intensity projections, 20 µm stacks).c and d) FACS gating strategy for NesGFP positive cells in the hippocampus and SVZ.All samples were first selected on viability based in DAPI and RedDOT staining, followed by sorting based on size and granularity, exclusion of doublets to have a population of viable single cells.These were then sorted based on GFP (cell marker) and tdTomato (tdTom-PLIN2) to quantify how many cells have LDs.a b