APOE4 to APOE2 allelic switching in mice improves Alzheimer’s disease-related metabolic signatures, neuropathology and cognition

In the paper

“E2 carriage has been linked to a higher risk of melanoma38, age-related macular degeneration39 and even some neurological disorders such as posttraumatic stress disorder40 and progressive supranuclear palsy41.”

What the sources say

• Disputed

  Citation 38 is attached to melanoma, not the other conditions in the sentence. The paper’s patient-genotype analysis states that APOE2 carriers were not enriched among melanoma patients and that APOE2 was not associated with increased melanoma incidence. It links APOE2 instead to poorer survival among patients with melanoma. (LLM, conf=0.90)

  Checked by a reasoning model — may be wrong.

  cites [38] Ostendorf, B. N. et al. Common germline variants of the human APOE gene modulate melanoma progression and survival. Nat. Med. 26, 1048–1053 (2020). sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites [39] Adams, M. K. M. et al. Apolipoprotein E gene associations in age-related macular degeneration: the Melbourne Collaborative Cohort Study. Am. J. Epidemiol. 175, 511–518 (2012). source
• Supported

  The Introduction states that E2 has been associated with increased risk and severity of PTSD. The Discussion reports more severe PTSD symptoms among E2/E3 combat veterans with PTSD, consistent with the cited sentence’s PTSD link. This citation is attached to PTSD, not to melanoma, macular degeneration, or progressive supranuclear palsy. (LLM, conf=0.91)

  cites [40] Johnson, L. A. et al. ApoE2 exaggerates PTSD-related behavioral, cognitive, and neuroendocrine alterations. Neuropsychopharmacology 40, 2443–2453 (2015). source

In the paper

“Because APOE4 is also a risk factor for cerebral amyloid angiopathy (CAA)31, we measured CAA before and after the inducible, astrocyte-specific transition to APOE2.”

What the sources say

• Misattributed

  The Results section reports APOE4-associated brain amyloid burden and plaque biomarkers, not CAA risk. In the References section (page unavailable), the only explicit CAA statement is the title of a separate study by Fryer et al. (2005), which says apoE4 promotes CAA in a transgenic mouse model. (LLM, conf=0.82) It matches “Human apolipoprotein E4 alters the amyloid-beta 40:42 ratio and promotes the formation of cerebral amyloid angiopathy in an amyloid precursor protein transgenic model” (2005) instead.

  Checked by a reasoning model — may be wrong.

  cites [31] Castellano, J. M. et al. Human apoE isoforms differentially regulate brain amyloid-β peptide clearance. Sci. Transl. Med. 3, 89ra57 (2011). sourceExperimental — accuracy not yet measured

In the paper

“Compared with E3, E4 confers up to a 13-fold increase in risk for AD in homozygous carriers1.”

What the sources say

• Supported

  The Results table in the Race and Ethnicity Analyses uses APOE*33 (E3/E3) as the reference and reports an odds ratio of 13.04 for APOE*44 (E4/E4) among White individuals. This supports the approximately 13-fold comparison, though the estimate varies across population groups. (LLM, conf=0.93)

  cites [1] Belloy, M. E. et al. APOE genotype and Alzheimer disease risk across age, sex, and population ancestry. JAMA Neurol. 80, 1284–1294 (2023). source

In the paper

“Previous work established that human APOE affects amyloid plaque deposition in the 5xFAD brain in a genotype-dependent fashion (E4 > E2)31.”

What the sources say

• Partly supported

  The Results section on old PDAPP mice reports that hippocampal amyloid plaque load varied by human apoE isoform, with E4 showing greater Aβ deposition than E2. The mice studied were PDAPP/TRE, not 5xFAD, so the source does not establish the claim specifically for the 5xFAD brain. (LLM, conf=0.90)

  cites [31] Castellano, J. M. et al. Human apoE isoforms differentially regulate brain amyloid-β peptide clearance. Sci. Transl. Med. 3, 89ra57 (2011). sourceExperimental — accuracy not yet measured

In the paper

“Conversely, E2 is strongly neuroprotective, reducing AD risk by up to 40%, slowing cognitive decline and increasing longevity2–4.”

What the sources say

• ?Not checked — no full text of the sourcecites [2] Corder, E. et al. Protective effect of apolipoproteinE type 2 allele for late onset Alzheimer disease. Nat. Genet. 7, 180–184 (1994). source
• Partly supported

  The Discussion (p. 7) reports lower amyloid-β oligomer levels in AD patients carrying ε2 and protection against cognitive decline in the oldest old. The paper also associates ε2 with exceptional longevity in its Italian and Japanese cohorts (abstract, p. 2), but it does not report a 40% reduction in AD risk or establish that ε2 is strongly neuroprotective. (LLM, conf=0.87)

  cites [3] Garatachea, N. et al. ApoE gene and exceptional longevity: insights from three independent cohorts. Exp. Gerontol. 53, 16–23 (2014). sourceExperimental — accuracy not yet measured
• Partly supported

  The Incident AD in apoE subgroups section reports a relative risk of 0.76 for E2 versus E3/3—an estimated 24% reduction, not a reported 40% reduction—and its 95% confidence interval (0.40–1.44) includes no reduction. The Discussion explicitly says the lower AD incidence was not statistically significant, so calling E2 strongly neuroprotective overstates this finding. The Discussion supports slower episodic-memory decline, but not a general slowing of cognitive decline; the cited passages do not establish increased longevity. (LLM, conf=0.88)

  cites [4] Wilson, R., Bienias, J., Berry-Kravis, E., Evans, D. & Bennett, D. The apolipoprotein E ε2 allele and decline in episodic memory. J. Neurol. Neurosurg. Psychiatry 73, 672–677 (2002). sourceExperimental — accuracy not yet measured

In the paper

“ApoE is a major component of plaques51.”

What the sources say

• Partly supported

  The Results report strong ApoE staining in a plaque core and extensive ApoE localization in plaque-like structures. These findings support ApoE’s presence in plaques, but the paper does not establish that it is a major component of plaques. (LLM, conf=0.82)

  cites [51] Gal, J. et al. Apolipoprotein E proteinopathy is a major dementia-associated pathologic biomarker in individuals with or without the APOE Epsilon 4 Allele. Am. J. Pathol. 192, 564–578 (2022). sourceExperimental — accuracy not yet measured

In the paper

“findings from a study by Mahan et al.36 suggested a role for astrocyte-derived ApoE in amyloid reduction and associated gliosis.”

What the sources say

• Supported

  The Conclusion reports that reducing astrocytic apoE3 or apoE4 decreased amyloid plaque deposition and microglial activation around fibrillar plaques. It also reports isoform-dependent effects on astrocyte activation, supporting the claim about associated gliosis in this mouse model. (LLM, conf=0.91)

  cites Mahan et al.36 Mahan, T. E. et al. Selective reduction of astrocyte apoE3 and apoE4 strongly reduces Aβ accumulation and plaque-related pathology in a mouse model of amyloidosis. Mol. Neurodegener. 17, 13 (2022). source

In the paper

“a process downregulated in AD brains34.”

What the sources say

• Partly supported

  The Discussion reports downregulated expression of UGCG, the enzyme that synthesizes glucosylceramide from ceramide, and cites lower UGCG activity in AD frontal cortex. But the Results report variable UGCG expression across the paper’s AD samples, including four with higher expression; the paper does not directly establish that glucosylceramide synthesis is generally downregulated in AD brains. (LLM, conf=0.84)

  cites [34] Filippov, V. et al. Increased ceramide in brains with Alzheimer’s and other neurodegenerative diseases. J. Alzheimers Dis. 29, 537–547 (2012). sourceExperimental — accuracy not yet measured

In the paper

“Additionally, microglia cluster around Aβ deposits in CAA49.”

What the sources say

• Partly supported

  The Results section ('Microglial Aβ binding via CD11b'; page unavailable) reports Aβ40 colocalized with microglial processes in AD/CAA brains. It does not report microglia clustering around Aβ deposits; the Discussion proposes that microglia transport Aβ to the vasculature, which is different from observing such clusters. (LLM, conf=0.84)

  cites [49] Zabel, M. et al. A shift in microglial β-amyloid binding in Alzheimer’s disease is associated with cerebral amyloid angiopathy. Brain Pathol. 23, 390–401 (2013). sourceExperimental — accuracy not yet measured

In the paper

“Using an inverse approach, Liu et al.46 showed that astrocyte-specific expression of E4 suppressed amyloid clearance, with particularly strong effects during the amyloid seeding phase.”

What the sources say

• Partly supported

  The Results section, “Expression of apoE4 in astrocytes suppresses Aβ clearance” (page unavailable), reports impaired Aβ clearance with astrocytic apoE4. A separate Results section finds stronger effects on Aβ levels and amyloid deposition when apoE4 was expressed during seeding than during rapid plaque growth. It does not establish that the suppression of clearance itself was particularly strong during seeding. (LLM, conf=0.84)

  cites Liu et al.46 Liu, C.-C.et al. ApoE4 accelerates early seeding of amyloid pathology. Neuron 96, 1024–1032 (2017). sourceExperimental — accuracy not yet measured

In the paper

“While that study did not report decreases in IBA1 as those seen in this study, they observed a similar decrease in reactive microglia using Clec7a, and a reduction in plaque-associated ApoE.”

What the sources say

• Supported

  In Results, “Microglial activation after the loss of astrocytic apoE” (page not provided), Clec7a around plaques decreased by approximately 50–70%, and apoE in plaque-associated microglia decreased by approximately 60–70% in Cre+ mice. The same passage reports that overall cortical Iba1 staining and Iba1+ microglial clustering around plaques were unchanged between Cre+ and Cre− groups. (LLM, conf=0.93)

  cites Mahan et al.36 Mahan, T. E. et al. Selective reduction of astrocyte apoE3 and apoE4 strongly reduces Aβ accumulation and plaque-related pathology in a mouse model of amyloidosis. Mol. Neurodegener. 17, 13 (2022). source

In the paper

“The literature suggests that there are isoform-dependent differences in ApoE–plaque associations but that several factors, such as ApoE lipidation state, compartment or cell type of origin, and Aβ conformational state, also have important roles in ApoE–Aβ interactions52–56.”

What the sources say

• ?Not checked — no full text of the sourcecites [52] Strittmatter, W. J. et al. Binding of human apolipoprotein E to synthetic amyloid beta peptide: isoform-specific effects and implications for late-onset Alzheimer disease. Proc. Natl Acad. Sci. USA 90, 8098–8102 (1993). source
• ?Not checked — no full text of the sourcecites [53] LaDu, M. J. et al. Purification of apolipoprotein E attenuates isoform-specific binding to β-amyloid. J. Biol. Chem. 270, 9039–9042 (1995). source
• ?Not checked — no full text of the sourcecites [54] LaDu, M. J. et al. Isoform-specific binding of apolipoprotein E to β-amyloid. J. Biol. Chem. 269, 23403–23406 (1994). source
• Supported

  Split into 2 atoms — 2 confirmed, 0 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  cites [55] Tai, L. M. et al. Levels of soluble apolipoprotein E/amyloid-β (Aβ) complex are reduced and oligomeric Aβ increased with APOE4 and Alzheimer disease in a transgenic mouse model and human samples. J. Biol. Chem. 288, 5914–5926 (2013). source
• Partly supported

  Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  cites [56] Hashimoto, T. et al. Apolipoprotein E, especially apolipoprotein E4, increases the oligomerization of amyloid β peptide. J. Neurosci. 32, 15181–15192 (2012). sourceExperimental — accuracy not yet measured

In the paper

“These results are consistent with studies implicating ApoE in plaque compaction and reflect a similar loss of function as seen in Apoe knockout mice, which also feature reductions in plaque-associated microgliosis57.”

What the sources say

• Supported

  The Introduction reports that plaques in Apoe-deficient mice were less compact and that plaque-associated microgliosis was significantly reduced. The plaque-morphology and microgliosis results sections report both findings in the knockout mouse models; page numbers were not supplied in our copy. (LLM, conf=0.94)

  cites [57] Ulrich, J. D. et al. ApoE facilitates the microglial response to amyloid plaque pathology. J. Exp. Med. 215, 1047–1058 (2018). source

In the paper

“Also in line with these results is a recent study, which showed that plaque-associated ApoE attracts microglia to the plaque and stimulates their activation via a vascular cell adhesion molecule 1-based mechanism58.”

What the sources say

• Supported

  The Results sections on ApoE chemoattraction and VCAM1–ApoE interaction report that plaque-associated ApoE directs VCAM1-dependent microglial migration toward Aβ plaques and is required for their subsequent transition to a phagocytic state. These effects were demonstrated after IL-33 treatment in APP/PS1 mice. (LLM, conf=0.91)

  cites [58] Lau, S.-F. et al. The VCAM1–ApoE pathway directs microglial chemotaxis and alleviates Alzheimer’s disease pathology. Nat. Aging 3, 1219–1236 (2023). source

In the paper

“Canonical CNS cell type markers were compiled from previous studies66–69.”

What the sources say

• Supported

  The Discussion (page unavailable) says the study assembled a collection of brain cell type-specific marker genes and that its most enriched genes include well-known markers. This supports citing it as a previous study from which CNS cell type markers could be compiled; it does not establish what the other cited studies contributed. (LLM, conf=0.88)

  cites [66] McKenzie, A. T. et al. Brain cell type specific gene expression and co-expression network architectures. Sci. Rep. 8, 8868 (2018). source
• Supported

  The Results section on vascular cells identifies canonical markers for a CNS-associated cell type and reports additional distinguishing marker genes. The Discussion states that major nervous-system cell classes are distinguished by class-specific genes. This makes the paper a relevant previous study from which CNS cell-type markers could be compiled; the citing sentence does not attribute the compilation itself to this paper. (LLM, conf=0.88)

  cites [67] Zeisel, A. et al. Molecular architecture of the mouse nervous system. Cell 174, 999–1014 (2018). source
• Supported

  The ‘Determination of cell type identity’ section says the authors used cell-specific markers previously described in the literature. The Fig. 1 legend lists representative CNS cell-type markers and cites earlier studies for them. Our copy does not establish which works the citing document numbers 66–69 refer to, but it supports the attributed practice. (LLM, conf=0.93)

  cites [68] Ximerakis, M. et al. Single-cell transcriptomic profiling of the aging mouse brain. Nat. Neurosci. 22, 1696–1708 (2019). source
• ?Not checked — no full text of the sourcecites [69] Han, X. et al. Mapping the mouse cell atlas by Microwell-seq. Cell 172, 1091–1107 (2018). source

In the paper

“A similar, albeit more modest, shift toward APOE2 expression was observed in whole-brain tissue, suggesting that most of total brain ApoE is derived from astrocytes, which were now synthesizing E2 (Fig. 4d).”

In the paper

“However, 1 month after TAM, APOE4s2G mice fed a Western diet displayed a substantial increase in very-low-density lipoprotein (VLDL) TGs and plasma ApoE levels, which rose to E2-TR levels (Fig. 2b–e and Extended Data Fig. 2d).”

In the paper

“While neurons synthesize ApoE under certain conditions, we observed negligible colocalization of ApoE with NeuN (Extended Data Fig. 2k–o).”

In the paper

“Dimensionality reduction identified 41 clusters, which were assigned to one of 12 unique cell types, with cell and cluster proportions similarly distributed across all groups (Fig. 3a and Extended Data Fig. 3).”

In the paper

“Colocalization of ApoE with cell-specific markers for astrocytes, microglia and neurons showed that like the E2-TR, E4-TR and APOE4s2 controls, ApoE primarily colocalized with astrocytes (glial fibrillary acidic protein (GFAP)+) in APOE4s2G mice, with measurable, but much lower, colocalization with the microglial marker IBA1 (Fig. 2g,h and Extended Data Fig. 2k–o).”

In the paper

“While the Morris water maze did not show striking differences in spatial learning or memory, 4s2A/FAD mice showed improvements in associative learning and more robust contextual and cued memory, an effect that was strongest in females (Fig. 5b–d and Extended Data Fig. 6a–d).”

In the paper

“Gene expression and proteomic analyses confirm that APOE4s2 mice synthesize E4 at baseline and E2 after tamoxifen administration.”

• Also stated in: § APOE‘switch’ mice efficiently transition from E4 to E2 · boosted

  “Together, these data demonstrate that APOE4s2G mice synthesize human ApoE4 at baseline, and that TAM induces a highly efficient transition to APOE2 and ApoE2 at the gene and protein level.”

In the paper

“We then performed a weighted gene coexpression network analysis (WGCNA) on the cerebral lipidome, where we identified five coexpression modules, including one network that was significantly correlated with the E4 to E2 switch (blue module) and consisted primarily of PC and phosphatidylethanolamine (PE) species (Fig. 2k,l).”

In the paper

“plaque proximity analyses showed more pronounced decreases in GFAP and IBA1 in 4s2A/FAD mice when close to the plaque center, with distribution of the GFAP+ and IBA1+ signal becoming more similar to controls as the distance from the center increased (Fig. 6j–m and Extended Data Fig. 8h–k).”

In the paper

“APOE4s2G mice, in which all cell types transition from APOE4 to APOE2, reliably phenocopy the humanized E4-TR mice before the switch and E2-TR mice within 1 month after the switch.”

In the paper

“We did note subtle sex differences in the expression of the postsynaptic density marker PSD95 in the cortex, although this was limited to plaque-distal regions and was not seen in the hippocampus (Extended Data Fig. 7i–m).”

In the paper

“Strikingly, we show that a 2-month-long astrocyte-specific E4 to E2 switch after pathology onset is sufficient to decrease plaque load, reduce markers of glial reactivity and improve associative, but not spatial, memory in the 5xFAD model.”

In the paper

“Several measures were significantly correlated but, paradoxically, positively correlated in 4s2/FAD mice yet negatively correlated in 4s2A/FAD mice (Fig. 7k).”

In the paper

“While higher amyloid plaque loads were observed in female compared to male 4s2/FAD Cre− controls, we did not detect any other amyloid-dependent sex differences, nor was this difference reflected in the soluble and insoluble Aβ42 and Aβ40 measures (Extended Data Fig. 7e–h).”

In the paper

“We characterize how a global E4 to E2 replacement drives peripheral metabolic changes and AD-relevant alterations in the cerebral transcriptome and lipidome.”

In the paper

“Forty-five clusters were identified.”

In the paper

“Leveraging the Ai9 cross, we then used cell sorting and qPCR to validate the high efficiency and cell specificity of the astrocyte-specific E4 to E2 switch (Fig. 4c).”

In the paper

“Considering the potential translational concerns associated with germline, whole-body APOE editing, we reasoned that a central nervous system (CNS)-selective APOE4 to APOE2 switch could afford the neuroprotection of E2 while minimizing off-target effects.”

In the paper

“Taken together, these results suggest that astrocyte-selective expression of ApoE2 leads to downstream reductions in microglial activation and lower levels of plaque-associated ApoE”

In the paper

“These findings underscore the dynamic and reversible nature of APOE’s contribution to AD pathology”

In the paper

“Together, these results suggest that a relatively late-stage switch to astrocytic expression of E2 is sufficient to substantially lower parenchymal plaque burden, but not CAA, and improve some, but not all, of the cognitive deficits associated with E4.”

In the paper

“Four mixed clusters with high hemoglobin were removed along with several clusters with pituitary contamination (high expression of Gh, Prl and Lhx3).”

In the paper

“Cell viability was checked using the Acridine Orange/Propidium Iodide viability kit (cat. no. LGBD10012, Logos Biosystems).”

• Also stated in: § Fluorescence-activated cell sorting · plain

  “Cell viability was checked using the Acridine Orange/Propidium Iodide viability kit.”

In the paper

“Feature counts were normalized using sctransform and the effects of percent.mito were regressed out.”

In the paper

“Cells were then filtered to reduce the potential of including low-quality cells with the following criteria: 200 < nGene < 2000; 200 < nCount; and percent.mito <20%.”

In the paper

“Final library quantification and quality checking were performed using a Bioanalyzer instrument (Agilent Technologies); 150-bp paired-end sequencing (Novogene) was performed on a NovaSeq 6000 S4 flow cell.”

In the paper

“This process resulted in stringent filtering of cells with ambiguous assignments, leaving a total of 85,701 cells in 41 carefully assigned clusters.”

In the paper

“Doublets were removed using the scDblFinder package.”

In the paper

“Ambient RNA contamination was removed using the SoupX R package.”

In the paper

“However, the location in which mice were dropped into the pool differed for each session between the four quadrants to avoid procedural bias in task learning.”