Circadian Control of Heparan Sulfate Levels Times Phagocytosis of Amyloid Beta Aggregates
The following datasets are from the manuscript entitled “Circadian Control of Heparan Sulfate Levels Times Phagocytosis of Amyloid Beta Aggregates” published in PLOS Genetics. This paper investigates the link between Alzheimer’s disease and circadian disruption by looking at naïve macrophages and the phagocytosis of a neurotoxic protein involved in Alzheimer’s Disease, amyloid-beta (1-42). We found that there is a circadian oscillation in the phagocytosis of amyloid-beta (1-42) with the peak of phagocytosis at Post Shock time (PS) 32 and the trough at PS 16. To confirm the observed rhythm, we repeated our phagocytosis experiment in PER1-/-/PER2-/- knockout cells, which lack a circadian rhythm, and observed a change in the rhythm, confirming circadian influence. We then probed a previously published macrophage proteomic dataset (Collins et al., Genome Research, 2021) and found genes involved in the heparan sulfate and chondroitin sulfate biosynthesis pathways to be circadianly controlled. To demonstrate the circadian control of cells surface heparan sulfate and chondroitin sulfate proteoglycans, we performed liquid chromatography tandem mass spectrometry on our macrophages and found that there is a circadian rhythm to both heparan sulfate and chondroitin sulfate proteoglycan expression. We found that the rhythm of heparan sulfate expression was anti-phase to the oscillation of amyloid-beta phagocytosis, indicating an inhibitory relationship. Therefore, we performed phagocytosis experiments in the presence of heparan sulfate cleaving enzymes (heparinases I, II, and III) and found an ablation of the previously observed rhythm of amyloid-beta phagocytosis. Furthermore, we investigated if this oscillation was unique to amyloid-beta (1-42) by repeating our phagocytosis experiments with other homologs of amyloid-beta, amyloid-beta (1-40) and mouse amyloid-beta (1-42). These homologs were of interest as they do not rapidly aggregate like amyloid-beta (1-42) and mouse amyloid-beta contains mutations in the known heparan sulfate proteoglycan binding region. We found that the amyloid-beta homologs did not have an oscillation in phagocytosis, demonstrating that aggregation and heparan sulfate proteoglycan binding are essential to the observed rhythm. The attached datasets are from all the above-mentioned experiments. For the phagocytosis experiments, the fluorescent microscopy images as well as the determined pixel intensity data are included. For the liquid chromatography tandem mass spectrometry data, the normalization method as well as the raw and normalized data are available.
Steps to reproduce
Amyloid-beta (1-42) phagocytosis data was collected using murine PER2::Luciferase cells (or PER1-/-/PER2-/- cells for the knockout experiment) that were grown to confluence and serum shocked to synchronize their circadian rhythms. The cells were induced with human amyloid-beta every 4 hours for 24 hours starting at Post Shock time (PS) 16 (time points in dataset are PS16, PS20, PS24, PS28, PS32, PS36, and PS40). The cells were then incubated for 2 hours, then harvested and fixed with formalin. They were imaged using a Zeiss LSM 510 Laser scanning confocal microscope using the 40x objective and an argon laser at 488nm excitation wavelength. Macrophages were viewed in two channels, the brightfield channel and the fluorescent channel. Images were taken until 100+ macrophages were sampled. To analyze these images, we used a custom MATLAB script. This script uses defined cell size ratios and edge detection to pick out single cells and objects (up to 4 cells clumped together) and measures the average pixel intensity, max pixel intensity, and area of each object. See related links to access this script on GitHub. Using the average cellular pixel intensities, we first log transformed the data and then removed outliers using the inter-quartile range (IQR). The resultant data was used for all downstream analyses and conclusions. Mouse Amyloid-beta (1-42) phagocytosis data was collected and analyzed the same way as described above but with the addition of mouse amyloid-beta at each timepoint. Amyloid-beta (1-42) phagocytosis with heparinases I, II, and III data was collected and analyzed the same way as above with the exception that heparinases and amyloid-beta were added at each timepoint. Amyloid-beta (1-42) phagocytosis with heparinase I and Amyloid-beta (1-42) phagocytosis with heparinase III data was collected the same as described above but only at timepoints PS20 and PS32. Amyloid-beta (1-40) phagocytosis data was collected and analyzed as described above with the exception of sampling only happening at PS20 and at PS32. For the liquid chromatography tandem mass spectrometry (LC-MS/MS) experiment, the cells were prepared the same way as described above. Following synchronization starting at PS16 we collected and flash froze cell pellets, spent media, and extracellular matrix (ECM) samples for every 4 hours over 24 hours (timepoints: PS16, PS20, PS24, PS28, PS32, PS36, PS40). The pellets were sampled in quadruplicate (n=4 for each timepoint), the spent media was sampled in triplicate (n=3 for each timepoint), and the ECM was sampled in duplicate (n=2 for each timepoint). We additionally took control samples of unused media and of cell stripper, which is the reagent we used to remove the macrophages from the cell culture dish. The cell pellet data was normalized to nanograms (ng) per cell by using a CYQUANT assay to calculate the number of cells per pellet. Please see our datasheet for our cell calculations.