Development of methods for enriching microglia nuclei from frozen brain samples to study the effect of CHIP on Alzheimer's disease

Abstract Background: Clonal hematopoiesis of indeterminate potential (CHIP) associates with increased risk of age-related diseases such as blood cancers and cardiometabolic diseases, the latter likely driven by altered functions of mutant myeloid cells. Recent work revealed a surprising protective association between CHIP and Alzheimer's disease (AD). CHIP-associated mutations have been detected in 30–90% of brain microglia, key mediators of AD pathogenesis, suggesting mutant microglia may directly modulate disease risk. Most archival human brain tissue available for analysis is flash-frozen in liquid nitrogen, but methods to isolate and study microglia from such tissue are not well established. In this project, we developed a strategy to sort highly pure microglia, enabling detailed characterization of their mutational drivers and genomic properties to uncover mechanisms underlying CHIP-associated neuroprotection. Methods: We procured frontal lobe samples from 22 post-mortem donors, 11 of whom were confirmed CHIP carriers from blood DNA, and 6 of whom had dementia. Tissue processing was done with an established protocol to isolate high quality nuclei, followed by a novel flow cytometry panel using antibodies to four transcription factors (NeuN, Olig2, cMaf, and PU.1). Unfractionated nuclei from these samples underwent single-nucleus ATAC-seq (snATAC-seq) to identify cell types and chromatin state, and the cell type proportions obtained from snATAC-seq were compared to flow cytometry results. Additionally, for three samples, we sorted cMaf+ PU.1+ NeuN- Olig2- microglia and performed snATAC-seq to assess purity. We also assessed the variant allele fraction (VAF) of CH variants from whole brain DNA and sorted fractions from each CHIP carrier (neurons [NeuN+], oligodendrocytes [Olig2+], microglia [cMaf+ PU.1+], and a triple-negative astrocyte rich compartment). Results: From the 22 brain samples, we recovered 221,982 high-quality nuclei, which were then integrated with an additional 28,131 nuclei derived from previously published snATAC-seq datasets of circulating immune cells. This combined dataset provided a comprehensive view of both hematopoietic and brain cell lineages. Unsupervised clustering revealed multiple distinct cell groups, including a single microglial cluster marked by the expression of the canonical microglial genes SALL1, P2RY12, TMEM119, and APOE. All the brain samples contained very low numbers of non-microglial immune cells. We found that snATAC-seq underestimated microglial proportions compared to flow cytometry. This was due to selective microglial loss after size-based sorting of intact brain nuclei after digest, as unsorted nuclei showed a substantially higher number of microglia by snATAC-seq. Sorting for cMAF+ PU1+ nuclei was highly effective for isolating brain-derived microglial fractions, yielding 90%-99% pure microglia in three samples. Analysis of four-fraction sorted brain samples showed that the same CHIP mutations seen in blood were detected in the microglia gate in 8 out of 9 evaluated samples. A small number of variant reads could be found in the astrocyte-enriched gate, likely due to contaminating microglia in this population, but mutant reads were very rare in neuron and oligodendrocyte gates. There was a very strong correlation between VAF of CHIP variants in sorted microglia and whole blood VAF, regardless of driver gene, suggesting that engraftment of brain by circulating precursors is not a rare event. Finally, we asked whether any genes showed differential accessibility of chromatin in the microglia of CHIP carriers versus controls. One of the most up-regulated genes in CHIP carriers was ITGA4, which was previously shown to be a marker of bone-marrow derived microglia (Bowman, Cell Reports 2016). This further validates the presence of bone-marrow derived microglia in CHIP carriers, and we hypothesize that ITGA4 may be involved in CHIP-derived immune cell migration into brain. Conclusions: These findings support the novel concept that CHIP-bearing, bone marrow–derived cells can engraft in the aging human brain, differentiate into microglia, and potentially confer protection against AD. The methods developed in this work sets the stage for mechanistic studies into how somatic mutations shape microglial function, ultimately opening new therapeutic avenues that harness the beneficial immune properties of CHIP in neurodegenerative diseases.