UC Irvine

Alzheimer’s Disease (AD) is the leading cause of age-related dementia, effecting an estimated 5.8 million people in the United States alone. Despite considerable progress in our understanding of the two hallmark pathologies that are thought to drive this disease; amyloid plaques and neurofibrillary tangles, no effective disease-modifying therapies have yet been developed. Since the initial reports by Alois Alzheimer in 1907, we have known that, in addition to plaques and tangles, the brains of patients exhibit considerable reactive gliosis1. More recently, genetic studies have provided compelling new evidence that microglia in particular play a critical role in the development and progression of AD.Although the field has known for over a century that microglia migrate to and surround amyloid plaques, the first genetic indication that microglia might play a more prominent role in AD came with the discovering of coding mutations in the microglial-enriched gene TREM2 (triggering receptor expressed on myeloid cells 2). Soon thereafter, Genome Wide Associative Studies (GWAS) begun to identify many additional AD risk genes that were highly or even specifically expressed by microglia. Yet, despite the identification of these new AD risk genes, much remains to be discovered about the roles of microglia in either the development or progression of this disease.

A major challenge to our understanding of microglia, both in general and in AD, has arisen from the difficulty of studying and manipulating human microglia. As highly plastic cells, microglia are extremely sensitive to their environment and thus examining these cells in disease- and brain-relevant ways has proven to be a challenging and at times misleading process. For example, microglia cultured in vitro rapidly change their transcriptional programs from those observed within the brain. Furthermore, studies of microglia in rodent models of AD can suffer from the fact that many AD risk genes lack appropriate homology to their equivalent human genes. Given that microglia are the primary immune cell of the brain and intimately involved in many aspects of both health and disease, it is critical to develop better ways of understanding their response to a changing brain environment. The focus of this dissertation is to develop and characterize a first of its kind chimeric microglia model to better capture the morphology, behavior, and transcriptional patterns of human microglia in vivo, and use to apply the model to study how Alzheimer’s Pathology effects human microglial genetics.

In this model, human HPCs are directly transplanted into the ventricles and overlaying cortex of postnatal day (P1) immunodeficient mice that are capable of supporting human microglial engraftment (MITRG and hCSF1 mice). These cells display robust long-term engraftment, migrating and populating much of the forebrain and differentiating into mature microglia within 2 months. Transplanted human cells exhibit morphology and markers typical of microglia, including complex ramifications and establish distinct niches that tile neatly together, making up approximately 80% of the total microglia across several forebrain regions and multiple HPC preparations. Following in vivo maturation, human microglia can be isolated from the brain by FACS or MACS sorting and RNA sequencing performed. Bulk RNA-seq demonstrates that xenotransplanted human microglia (xMGs) recover key in vivo microglial signature genes and much of the transcriptomic profile which were recently shown by Chris Glass’ lab to be lost during in vitro culturing of human microglia. Furthermore, when wild type (WT) P1 transplanted animals are compared to those given vehicle (saline) cerebroventricular injections, no deficits in either Morris Water Maze (MWM) or Elevated Plus Maze (EPM) are detected, indicating that this xenotransplantation model could be used to examine how disease-associated mutations in microglial genes might influence memory or anxiety-like behavior. (Chapter 1).

This paradigm can also be used to study the interactions between human microglia and AD-associated amyloid pathology, by using immunodeficient MITRG mice that are backcrossed with 5XFAD mice to produce a colony of xenotransplantation-compatible 5X-MITRG mice which develop amyloid plaque pathology. Following transplantation into this model, iPSC-derived HPCs differentiate into microglia, migrate towards A-plaques, and exhibit several disease-associated microglial (DAM) phenotypes including downregulation of the homeostatic marker P2RY12 and adoption of a more amoeboid morphology. Next, wildtype eGFP-expressing iHPCs were transplanted into P1 5x-MITRG and wildtype littermates and allowed to age to 10 months to ensure robust accumulation of plaques. Half-brains were fixed and used for histological analysis, whereas the other half was used isolate human microglia for single-cell RNA sequencing (scRNA seq). The resulting data was then compared to equivalent existing murine microglial datasets. We found that xMGs display a uniquely human transcriptomic response to amyloidopathy. To further demonstrate the utility of this chimeric AD model, we performed transplantations with isogenic WT and TREM2 R47H HPCs and found that this disease relevant mutation could lead to similar migration deficits as those observed in human postmortem samples. (Chapter 1).

Although tau pathology plays a critical role in AD and microglia have been implicated in the propagation of tau, very little research has yet been conducted to understand the responses and interactions of human microglia with neurofibrillary tangles. Given the significant links of microglia with both amyloid plaques and neurofibrillary tangles separately and the co-occurrence of these two hallmark pathologies in human AD, this gap in our understanding becomes even more surprising. In order to address these significant deficits, we have applied our xenotransplantation paradigm to a newly generated mouse in our lab, a cross between the hCSF1-5X mouse and the Tau P301S/PS19 mouse model in order to examine human microglial interactions with both amyloid plaques and neurofibrillary tangles Male mice were genotyped and transplanted on P1, allowed to age, and euthanized at 6 months old to allow for accumulation of both amyloid and tau pathologies. xMGs isolated from animals exhibiting amyloid and tau pathologies demonstrate expanded DAM, INF and IL-1β clusters as well as develop a dramatic Rod phenotype within CA1 that is correlated with a Type 1 Interferon Response. These results provide a wealth of transcriptomic data from mice exhibiting amyloidopathy and/or tauopathy as well as evidence for a fascinating and little understood interferon-induced microglial Rod phenotype present in AD. (Chapter 2)

Taken as a whole, the development of these novel chimeric microglia models has allowed us to ask deep questions about human microglia dynamics in health and disease. We are able to fully recapitulate the transcriptional profiles of ex vivo microglia isolated from human patients, transplanted cells respond to both acute and peripheral immunological stimuli, migrate toward amyloid beta plaques, adopt activation profiles reported in AD patient microglia, and morphologically adapt to tau pathology and the changing brain environment. The recapitulated human microglial phenotypes observed in these models coupled with the power of induced pluripotent stem cell modeling allows researchers to alter genes of interest via CRISPR modification and following transplantation gain a greater understanding of how key Alzheimer’s risk genes affect human microglial responses to AD pathologies.