Author ORCID Identifier

Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation





First Advisor

Dr. Steven Estus


Microglia are the resident immune cells of the brain, undertaking many critical tissue maintenance functions such as immune surveillance and phagocytosis. Microglial dysfunction has recently been identified as a multi-stage signature of many neurodegenerative diseases, including late-onset Alzheimer’s Disease (LOAD). Genome-wide association studies (GWAS) have identified single nucleotide polymorphisms (SNPs) in over thirty genes that modulate risk of developing LOAD. In the central nervous system, roughly half of these LOAD-associated genes are primarily expressed in microglia. The proteins encoded by these genes include cell surface receptors that contain either immunomodulatory tyrosine-phosphorylated activating motifs (ITAMs) or inhibitory motifs (ITIMs), including TREM2, CD33, and SIGLEC14. Here, I studied the molecular genetics underlying these three genes and their respective contributions to LOAD risk.

First, I found that TREM2 undergoes extensive alternative splicing in multiple tissues, including brain. Total TREM2 expression is not different as a function of LOAD diagnosis (p = 0.1268), but TREM2 expression is increased by 34% in tissues with higher National Institute on Aging/Reagan Institute (NIARI) scores (p = 0.0033). I also found that a novel TREM2 isoform lacking exon 2, D2-TREM2, accounts for 11% of the total TREM2 mRNA in human brain, and that this splicing efficiency is not altered as a function of AD status (p = 0.4909) or brain pathology (p = 0.9502). I also found that the D2-TREM2 protein has similar subcellular localization to its parent TREM2 protein, as both are primarily retained in the Golgi apparatus.

Next, I studied the exon 2-lacking CD33 isoform, D2-CD33. I developed an in vitro model to study the function of the D2-CD33 using a CRISPR-Cas9 approach in the U937 human monocyte cell line. After validating this model with sequencing, qPCR, and flow cytometry, I found that a nearby pseudogene, SIGLEC22P, was used as a repair template in approximately 10% of edited cells. This finding also provided the highest resolution to date of the clinically relevant anti-CD33 P67.6 antibody clone, gemtuzumab.

Finally, I combined a recent LOAD GWAS with a protein quantitative trait loci (pQTL) study to uncover SIGLEC14 as a potentially overlooked LOAD risk factor. I found that a previously described SIGLEC14 genetic deletion occurs within a 692 bp crossover region. I also found additional copy number variation not previously described using both qPCR-based and in silico assays, with copy numbers identified ranging from zero to four. While SIGLEC14 deletion does correlate well with a proxy single nucleotide polymorphism (SNP), rs1106476, additional SIGLEC14 genomic copies do not correlate with this SNP. Further, the SIGLEC14 genomic deletions correlate stepwise with decreased SIGLEC14 expression (p = 0.0002), and also correlate significantly with decreased SIGLEC5 expression (p = 0.0389).

In conclusion, microglial cell surface receptors are heavily implicated in the risk of developing LOAD, and these studies advance the field by adding to the molecular mechanisms which underlie their risk contribution. Further studies will be needed to address whether these findings can be translated clinically to either potential druggable targets or biomarkers.

Digital Object Identifier (DOI)

Funding Information

This study was supported by National Institute of Aging (RF1AG059717 from 2018 to 2022 and R21AG068370 from 2020 to 2022) to Dr. Steven Estus, the National Institute on General Medical Sciences (T32GM118292 from 2017 to 2018) to Benjamin Shaw, and National Institute on Neurological Disorders and Stroke (F99NS120365 from 2021 to 2022) to Benjamin Shaw