Alzheimer’s Disease (AD) remains the leading cause of dementia in the elderly, one for which there is still no cure. Efforts in AD drug development have largely focused on treating neuronal pathologies that appear relatively late in the disease. Most of this research is carried out in neuron-focused rodent models that overexpress penetrant human genes known to cause familial AD (FAD), such as APP, PSEN1, although it constitutes only about 5% of total AD cases, as opposed to sporadic AD which makes up the remaining cases. The consistent ineffectiveness of existing therapeutic approaches which mostly aim at reducing the level of Aβ in the brain has challenged the Aβ-dominated as well as the neurocentric way of viewing the disease.
Meanwhile, due to their complex interplay with neurons, there is greater appreciation for the role of astrocytes in neurodegeneration. The strongest genetic risk factor for sporadic AD is the E4 allele of the APOE gene (Corder et al., 1993). The APOE protein has 3 isoforms: APOE2 (protective), APOE3 (wild type) and APOE4 (pathogenic) and is primarily produced by astrocytes. Most APOE4-related dysfunctions in the AD brain are described in the context of the Aβ pathway which accumulates with aging, or by the abnormal expression of APOE in neurons, which occurs due to stress conditions possibly brought on with aging or other factors. However, there is a dearth of research on the cell-autonomous effects of APOE4 in astrocytes. Additionally, in order to catch the disease at its earliest, there needs to be greater focus on studying dysfunctions that are known to arise early in susceptible individuals, such as metabolic dysfunctions (Reiman et al., 2004) and endocytic abnormalities (Cataldo et al., 2000). Also, developing better human-focused disease models is the need of the hour, to avoid artefacts arising as a result of species-specific differences. Therefore, the goal of this study was to develop an isogenic human patient-derived neural cellular model to examine APOE4-isoform specific effects in neurons and astrocytes.
The first step towards that goal was the generation of a series of iPS cells derived from human APOE4 carriers and non-carriers. iPSCs were characterized for pluripotency and chromosomal integrity, by karyotyping. Next. In order to modify the APOE4/4 genotype of a patient-derived iPSC line to APOE3/3, genome editing protocols for precise homology directed repair (HDR) were optimized. The iPSC field suffers from a scarcity of protocols for CRISPR/Cas9 ribonucleoprotein-mediated HDR which is a rare event compared to non-homologous end joining which leads to frameshift mutations. Therefore, protocols were developed for CRISPR delivery, FACS and clonal isolation that would yield a high transfection efficiency, high clonal survival and high rate of HDR. Using these protocols, an isogenic APOE3/3 iPSC line was developed form an APOE4/4 AD patient line. The use of isogenic controls that differ only at a single locus is important when studying a polygenic disease like AD to ensure that subtle phenotypes are not lost due to individual-to-individual genomic background variation. The pair of isogenic lines were then differentiated to cortical neurons. Characterisation of the neurons using cortical markers revealed that cells of both genotypes gave rise to similar neuronal populations. Typical AD-related phenotypes were tested with these neurons. Compared to APOE3/3, APOE4/4 neurons exhibited somatodendritic mislocalisation of phosphorylated tau and increased cell death in response to oxidative stress. There was no significant difference in Aβ42/40 ratio or mean area of endosomes and lysosomes between the genotypes. However, treatment of APOE4/4 neurons with APOE4/4 astrocyte-conditioned medium lead to an increase in Aβ42/40 ratio as well as endosome and lysosome enlargement. Next, to study the contribution of astrocytes to APOE4-associated AD pathology, the isogenic lines were differentiated to astrocytes. Astrocytes were characterized for common astrocytic markers and functional properties. Astrocytes expressed and secreted APOE, were able to take up glutamate, elicit calcium transients in response to glutamate and ATP, and increased chemokine secretion upon exposure to TNFα. In comparison to APOE3/3 astrocytes, APOE4/4 astrocytes exhibit endosome and lysosome enlargement as well as reductions in glycolytic capacity, basal respiration and ATP production. This was accompanied by a decrease in mitochondrial complex expression and an increase in mitochondrial reactive oxygen species (ROS) production.
In summary, this study describes the generation and characterization of isogenic physiologically relevant iPSC-derived neurons and astrocytes that can be used for sporadic AD modelling studies. The neural cells were utilized to highlight early pathologies in AD astrocytes that are understudied, such as metabolic and endolysosomal deficits. The protocols developed for genome editing and differentiation can be used to develop models for other diseases with a genetic component.