Scientists know that neurons can generate a dribble of ApoE, but it is considered a drop in the bucket relative to the massive amounts of it churned out by their glial neighbors. Nevertheless, a study published May 6 in Nature Neuroscience makes the case that neuronal ApoE not only signals distress, but also may have a hand in killing the neurons. Using single-cell transcriptomics, researchers led by Yadong Huang at the University of California, San Francisco, report that, in mice, certain neurons ratchet up ApoE expression as the animals age. They do so earlier, and with more gusto, in mice expressing ApoE4. Similarly, in people, particular neurons rev up ApoE in the early stages of Alzheimer’s disease. In later stages, those neurons were more likely to have perished, hinting that stepped-up ApoE expression could spell doom for neurons. Those same neurons also start expressing immune-response genes such as that for the antigen-presenting, major histocompatibility complex I. The researchers tied MHC-I to tau pathology within individual neurons.
- In mice, neurons that express ApoE also express antigen-presenting genes.
- This rises with age and, in ApoE4 knock-ins, starts earlier.
- Deleting ApoE from neurons wards off age-related death.
- In AD, neuronal ApoE tracks with progression, perhaps with cell loss.
“This fascinating paper describes a novel role for neuronal ApoE expression in Alzheimer’s disease,” wrote Gwyneth Welch and Li-Huei Tsai of the Massachusetts Institute of Technology (full comment below).
“The results leave little doubt that neurons can express substantial levels of ApoE transcripts in humanized mice and human postmortem brains, and that this expression is linked with pathology,” commented Tony Wyss-Coray of Stanford University in Palo Alto, California. “The study also opens a glimpse at what is likely to emerge from single-cell molecular profiling of brain cells in the years to come, and how it may finally provide answers to the perennial question of ‘selective cellular vulnerability.’”
Astrocytes make the lion’s share of ApoE in the brain, and in recent years, microglia as well have been found to ramp up ApoE expression when activated (Sep 2017 news). While most research labs have focused on glial sources of the apolipoprotein, Huang has studied its expression in neurons for many years. He initially used a GFP reporter mouse to show neurons boost ApoE expression in response to injury, and later linked neuronal ApoE to a slew of detrimental phenotypes, including Aβ and tau aggregation (Xu et al., 2006; timeline; Jun 2018 news). Huang also noticed that only some neurons, even within the same region of the brain, or of the same subtype, expressed ApoE.
Could those neurons be more susceptible to neurodegeneration? To test this idea, first author Kelly Zalocusky and colleagues asked how gene expression profiles defined different subsets of neurons in the hippocampus, and how those profiles changed with ApoE isoform and with age. They sequenced the nuclear transcriptomes of cells in ApoE3 knock-in and ApoE4-KI mice at 5, 10, 15, and 20 months of age. These animals carry human ApoE genes in place of their endogenous one. In all, the researchers measured transcripts of 21,204 genes from 123,489 mouse nuclei. Their gene expression profiles fell into 27 clusters of cells, including 16 distinct neuronal clusters. Lo and behold, more than any other transcript, expression of ApoE—regardless of ApoE genotype—best explained the variance among neuronal clusters. For example, using a principle component analysis, the researchers found that ApoE expression levels explained 73 percent of the variance among granule cells in the dentate gyrus.
Next, the researchers asked how ApoE expression relates to the rest of a neuron's transcriptome. Using pathway analysis, Zalocusky found that ApoE levels correlated with expression of genes involved in metabolism, neurodegeneration, cellular senescence, and apoptosis, all of which had previously been tied to neuronal ApoE expression. They also uncovered some new connections. For example, neuronal ApoE was linked to changes in DNA damage and repair, the unfolded protein response, and immune response genes. This last one came as a surprise, Huang said, because immune genes are typically considered the business of glial cells, not neurons.
To address how age affects ApoE expression in the context of transcriptome changes overall, the scientists calculated the proportion of neurons expressing the largest amount of ApoE within each of the 16 neuronal subtypes they had identified. They found that the proportion of these ApoE super-expressers rose with age, then fell again. What dictated this? Apparently, ApoE genotype. For example, in ApoE4-KI mice, the proportion of dentate gyrus granule cells and CA1 neurons that expressed abnormally high levels of ApoE peaked at 10 months of age, whereas for ApoE3-KI mice, the peak came at 15 months.
Is neuronal ApoE merely a marker of shifts in transcription, or does it drive damage and even death of the cell in which it happens? To find out, the researchers investigated how neuronal ApoE expression influenced age-related neuronal loss in ApoE-KI mice. They found that ApoE4-KI mice lost more neurons in the hippocampus with age than did ApoE3-KIs, and that the remaining neurons had fewer synapses. However, knocking out ApoE only from neurons prevented this age-related synaptic loss and neuronal death. Even ApoE3-KI mice benefited from getting of rid of neuronal ApoE, and their hippocampi better maintained their volume with age. This suggested that ApoE expression renders neurons vulnerable to degeneration.
No ApoE, Please. Older ApoE4-KI mice had fewer hippocampal synapses, as judged by PSD-95, than ApoE3-KIs. With ApoE removed from neurons (right two columns) the synapses stayed. [Courtesy of Zalocusky et al., Nature Neuroscience, 2021.]
Might immune response genes expressed in neurons replete with ApoE play a part in their vulnerability? Of all of the immune genes that tracked with ApoE, MHC-I genes correlated most strongly. The researchers attempted to pick apart the causal nature of these associations in primary neuron cultures, reporting that ApoE drove the uptick in MHC-I. Curiously, they also found that heightened ApoE expression made neurofibrillary tau appear in the neuronal soma. This was largely mediated by expression of functional MHC-I molecules on the cell surface.
Moreover, MHC-I appeared to exacerbate tau pathology in the mouse brain. The scientists triggered tau pathology in mice by injecting adeno-associated virus expressing human pathological tau into their hippocampi. Six weeks later, wild-type mice had more aggregated, phosphorylated human tau in the brains than did mice lacking β2-microglobulin, a protein required to stabilize MHC-I on the cell surface.
Does any of this hold up in the human brain? The researchers looked at single-nucleus transcriptomic data from 48 prefrontal cortex samples in the Religious Orders Study and Memory and Aging Project (ROSMAP). Astrocyte ApoE expression dwarfed that in neurons; still, some of the neurons did express more ApoE than their neighbors. The authors detected nine clusters of excitatory and four clusters of inhibitory neurons. The proportion of high ApoE expressors varied both by cluster and disease stage. At 28 percent, it was highest in one cluster of interneurons in people with mild cognitive impairment.
Jibing with the mouse data, the researchers reported that expression of genes in several biological pathways—most notably, immune response genes—correlated with ApoE transcripts in individual human neurons. They found the same associations in other datasets that included people without AD neuropathology. This suggested to them that, regardless of disease state, ApoE expression in a neuron was tied to that of immune genes. They did not find this relationship in non-neuronal cell types, including astrocytes.
In the ROSMAP cohort, which included people who had died at different stages of AD, neuronal ApoE expression tracked with disease stage. Among seven of the 13 neuronal subtypes, the proportion of cells that expressed much ApoE was highest among people with MCI, and lower in cognitively normal people and those with more advanced AD. Though cross-sectional, the findings look as if neuronal ApoE expression rises early in disease, then wanes in later stages.
Perhaps neurons expressing the most ApoE are those destined to die, and that’s why neuronal ApoE expression is lower with advanced disease? The ROSMAP data hint that this could be the case. Indeed, those neuronal subsets that had the highest proportion of ApoE super-expressors in people with MCI were sparser by the later stages of AD. In toto, the data suggest neurons might ramp up ApoE early in disease, and those that do are likely to succumb. Finally, in the ROSMAP part of this study, the researchers once again tied neuronal expression of several MHC-I genes with tau pathology.
“Altogether, these data indicate that the upregulation of ApoE–MHC-I pathways in individual neurons is a cellular response that eventually causes neuronal death, even during healthy ageing,” wrote Jessica Wagner and Jonas Neher of the German Center for Neurodegenerative Diseases in Tübingen in an editorial.
Huang and several commentators noted that the mechanisms connecting neuronal ApoE to MHC-I and to tangles are unclear; more study is needed to fill in the gaps. Many other cell pathways correlated with ApoE expression as well, so they could be involved, too. Earlier this year, researchers led by Martin Kampmann of the University of California, San Francisco, reported that excitatory neurons expressing the transcription factor RORB are particularly vulnerable to tau toxicity (Jan 2021 news).
Stress to ApoE to MHC. In this proposed model of AD, normal neurons (green) increase expression of ApoE with age or stress (purple). This drives MHC-I expression, which, in turn, could trigger tau phosphorylation and attract CD8+ T cells or reactive microglia, which might prune synapses. [Courtesy of Zalocusky et al., Nature Neuroscience, 2021.]
“Selective vulnerability of specific neuronal subtypes is a hallmark of neurodegenerative diseases,” wrote Kampmann. “A systematic understanding of the underlying mechanisms would not only give us insights into disease mechanisms, but could also uncover therapeutic strategies to turn vulnerable neurons into resilient neurons.” His group plans to study the effect of ApoE genotype on RORB-linked vulnerability.
To Huang, the findings suggest that some neurons crank up ApoE under conditions of stress, including stress from aging and neuropathology. ApoE then somehow instigates a rise in immune response genes, including MHC-I, and this somehow instigates tau pathology—in neuronal culture. In the brain, a cadre of other players are involved. Huang hypothesized that MHC-I in neurons might serve as an “eat me” inviting immune cells to feast on synapses, reactivating a developmental pathway.
Michal Schwartz of the Weizmann Institute of Science in Tel-Aviv pointed out that neuronal expression of MHC-I has also been implicated in axonal regeneration following injury (Sabha et al., 2008; Joseph et al., 2011). “Expression of MHC-I by neurons is not all or none, good or bad, but rather, is dependent on context, timing, and persistence,” Schwartz wrote.
Conventionally speaking, MHC-I binds to T cell receptors on CD8+ T cells. Curiously, scientists led by Wyss-Coray recently spotted swarms of these cytotoxic cells in the brains of people with AD and PD (Jan 2020 news). “The studies open the possibility that ApoE may promote neuronal susceptibility to potential T cell killing by upregulating the cognate receptor (i.e., MHC-I-b2M-peptide) for virus-specific T cells, or T cells of yet unknown specificity for neuronal antigens,” Wyss-Coray noted.
If neuronal stress is what triggers ApoE expression, then whence the stress, and why does it happen sooner in ApoE4-expressing mice? Guojun Bu of the Mayo Clinic in Jacksonville, Florida, said aging brings on myriad stresses in the brain. They run the gamut from oxidative to erosion of the vasculature to sluggish clearance of harmful metabolites. ApoE4 may render the brain more vulnerable to these age-related stressors, he said.
Sam Gandy and Michelle Ehrlich of Mount Sinai School of Medicine in New York wondered how to square Zalocusky’s results on neuronal ApoE4 with those by David Holtzman’s group, which reported that selectively removing ApoE4 from astrocytes protected neurons in a mouse model of tauopathy (Apr 2021 news). No matter where it is expressed, ApoE4 seems to harm the brain. “As a result, one wonders whether whole-body knockdown of APOE expression might be the most parsimonious solution to what has turned out to be a huge challenge of ‘APOE-e4-whack-a-mole,’” Gandy and Ehrlich wrote.
To Ole Isacson of Harvard Medical School, the data suggest that neurons are compensating for the shortcomings of their glial support staff. Among a plethora of functions, astrocytes normally supply neurons with lipids, delivered in an ApoE package, which neurons use to replenish their extensive membranes. “If coupling between astrocyte and neuron in terms of lipid handling is faulty, the neuron is going to go under,” Isacson said. Perhaps neurons cope by churning out their own ApoE. In this context, Isacson said, neuronal ApoE might benefit the neuron, at least for awhile. Because ApoE4 is worse at carrying lipids than is ApoE3, neurons in ApoE4-KI mice may start to compensate earlier.
Isacson emphasized that beyond lipid handling, ApoE is also known to promote the resolution of inflammation, through binding to the complement protein C1q (Feb 2019 news). He therefore cautioned against the idea of targeting the apolipoprotein as a therapeutic strategy.—Jessica Shugart
- ApoE and Trem2 Flip a Microglial Switch in Neurodegenerative Disease
- In Human Neurons, ApoE4 Promotes Aβ Production and Tau Phosphorylation
- Selective Vulnerability News: RORB Neurons Are First Victims of Tangles
- Attack of the Clones? Memory CD8+ T Cells Stalk the AD, PD Brain
- Squelching ApoE in Astrocytes of Tau-Ravaged Mice Dampens Degeneration
- ApoE Binds Complement Protein, Keeps Inflammatory Cascade in Check
- Xu Q, Bernardo A, Walker D, Kanegawa T, Mahley RW, Huang Y. Profile and regulation of apolipoprotein E (ApoE) expression in the CNS in mice with targeting of green fluorescent protein gene to the ApoE locus. J Neurosci. 2006 May 10;26(19):4985-94. PubMed.
- Sabha M Jr, Emirandetti A, Cullheim S, De Oliveira AL. MHC I expression and synaptic plasticity in different mice strains after axotomy. Synapse. 2008 Feb;62(2):137-48. PubMed.
- Joseph MS, Bilousova T, Zdunowski S, Wu ZP, Middleton B, Boudzinskaia M, Wong B, Ali N, Zhong H, Yong J, Washburn L, Escande-Beillard N, Dang H, Edgerton VR, Tillakaratne NJ, Kaufman DL. Transgenic mice with enhanced neuronal major histocompatibility complex class I expression recover locomotor function better after spinal cord injury. J Neurosci Res. 2011 Mar;89(3):365-72. Epub 2010 Dec 22 PubMed.
No Available Further Reading
- Zalocusky KA, Najm R, Taubes AL, Hao Y, Yoon SY, Koutsodendris N, Nelson MR, Rao A, Bennett DA, Bant J, Amornkul DJ, Xu Q, An A, Cisne-Thomson O, Huang Y. Neuronal ApoE upregulates MHC-I expression to drive selective neurodegeneration in Alzheimer's disease. Nat Neurosci. 2021 Jun;24(6):786-798. Epub 2021 May 6 PubMed.