ApoE4 weakens the blood-brain barrier in the hippocampus, and in this way takes its toll on cognition, according to a study published April 29 in Nature. Researchers led by Berislav Zlokovic of the University of Southern California in Los Angeles reported that ApoE4 carriers sprang more leaks in their hippocampal blood-brain barriers (BBBs) than did noncarriers, and that this permeability was even worse if carriers were cognitively impaired. Surprisingly, the cognitive impairment occurred regardless of Aβ or tau pathology—the hallmarks of Alzheimer’s disease. Markers of BBB damage soared in the cerebrospinal fluid of E4 carriers and predicted subsequent slippage on cognitive tests. In all, the findings suggest that apart from stoking Alzheimer’s protein pathology, ApoE4 also harms the brain by eroding its border.

  • Blood-brain barrier appears disrupted in hippocampi of ApoE4 carriers.
  • Barrier breakdown correlates with cognitive decline independently of Aβ or tau pathology.
  • Pericyte protein in CSF potential marker of BBB damage.

“These observations cast new light on APOE4 that runs contrary to the widely held idea that this gene variant contributes to Alzheimer’s disease solely by promoting Aβ and tau accumulation,” wrote Makoto Ishii and Costantino Iadecola of Weill Cornell Medical College in New York in an editorial. “Instead, it seems that BBB dysfunction might explain why APOE4 carriers are susceptible to Alzheimer’s disease.” They write that Aβ and tau pathology come later as the disease progresses, exacerbating cognitive decline, and that ApoE contributes to their accumulation also.

Zlokovic’s group had previously reported that the hippocampal BBB grows leakier with age, and might even trigger cognitive decline (Jan 2015 webinar; Jan 2019 news). Other groups have reported that the integrity of this cellular barrier wanes in the early stages of AD, though this claim is not universally accepted (Nov 2015 news; van de Haar et al., 2016; van de Haar et al., 2016). At the same time, the strongest genetic risk factor for AD, ApoE4, has been tied to a damaged cerebrovasculature, including injury to the pericytes that form part of the blood-brain barrier (Halliday et al., 2016; May 2012 news). However, it was unclear if ApoE4’s toxic relationship with the brain vasculature itself causes cognitive decline.

Lead authors Axel Montagne, Daniel Nation, Abhay Sagre, Giuseppe Barisano, and Melanie Sweeney investigated. They used dynamic-contrast-enhanced magnetic resonance imaging (DCE-MRI) to gauge the permeability of the BBB in a cohort of 245 participants averaging 69 years of age, including 101 who carried at least one copy of ApoE4, and 144 noncarriers who had two copies of ApoE3. DCE-MRI uses a contrast agent to measure leakage of a tracer through blood vessels in the brain. Based on their clinical dementia rating scale scores of 0 and 0.5, volunteers were deemed cognitively normal or mildly impaired. Compared with noncarriers, ApoE4 carriers had a disrupted BBB in their hippocampi and parahippocampal gyri (see image below).

BBB Breakdown. Maps of BBB permeability generated by DCE-MRI (left) revealed leakage in the hippocampi of ApoE4 carriers (right), and even more so in those with cognitive impairment. Permeability increases from blue to red. [Courtesy of Montagne et al., Nature, 2020.]

This BBB breakdown was worse in ApoE4 carriers who were cognitively impaired. In noncarriers, cognitive impairment came with a compromised barrier, as well, but the effect was smaller than in carriers.

Did the presence of plaques or tangles influence the relationships among ApoE4, the BBB, and cognition? Not according to CSF or PET imaging. The researchers report that among ApoE4 carriers, BBB disruption came with cognitive impairment regardless of CSF Aβ42 or p-tau. In a subset of 74 and 96 cognitively normal participants who underwent PET scans for brain amyloid or tau, respectively, plaques and tangles accumulated in different regions than those whose BBB looked disrupted. While ApoE4 carriers had more cortical plaques, with the orbital frontal cortex bearing the brunt, the cortex had but sparse signs of a disrupted BBB. Tangles were detected in the inferior temporal gyri of some participants, which is common during normal aging, but levels were not influenced by ApoE genotype.

Together, the findings suggested that while ApoE4 did bring a greater amyloid burden, it did so independently of docking BBB integrity in the medial temporal lobe. Furthermore, this erosion of the barrier associated with poorer cognition, regardless of AD pathology.

Next, Montagne and colleagues checked if CSF markers of BBB damage track with ApoE genotype and cognitive decline by measuring soluble platelet-derived growth factor receptor-β. sPDGFRβ sheds from stressed pericytes and has been tied to BBB damage and cognitive decline (Miners et al., 2019; Nation et al., 2019). The group recently developed an assay to detect the biomarker in CSF (Sweeney et al., 2020), and found that among 350 participants, those with higher baseline CSF sPDGFRβ slipped faster on cognitive tests over two to four years of follow-up—but only if they carried ApoE4. This remained significant regardless of amyloid or tau biomarkers. CSF sPDGFRβ correlated with BBB permeability in the medial temporal lobe detected by DCE-MRI.

Montagne found that ApoE4 carriers who were cognitively impaired also had more cyclophilin A (CypA) and metalloproteinase 9 (MMP9) in their CSF than did normal carriers. Previous studies reported that stressed pericytes pump out CypA, which prompts endothelial cells to crank expression of MMP9, a protease that digests the BBB (Jin et al., 2004; Bell et al., 2012). Cognitively impaired E4 carriers also had elevated CSF concentrations of neuron-specific enolase (NSE), a marker of neuronal stress. The researchers did not measure neurofilament light (NfL), a marker of neurodegeneration that rises early in AD.

Finally, the researchers report that pericytes derived from induced pluripotent stem cells from ApoE4 homozygotes expressed much more CypA and MMP9 than pericytes derived from ApoE3 homozygotes. This suggested that, in agreement with animal data, ApoE4 may ramp up the CypA-MMP9 pathway in human pericytes (Bell et al., 2012). 

All told, Zlokovic believes the findings link ApoE4 to BBB breakdown and implicate the pro-inflammatory CypA-MMP9 pathway as part of the mechanism. How ApoE4 trips off this pathway, or why the BBB damage happens only in the medial temporal lobe, is unclear.

Also unexplained is whether the cognitive impairment associated with this barrier breakdown represents AD. Zlokovic said it will be interesting to monitor the ApoE4 carriers who tested negative for AD biomarkers and were cognitively impaired, to see if they develop AD. Overall, Zlokovic argues that vascular dysfunction and Aβ accumulation are two distinct pathways, both of which are exacerbated by ApoE4. While some studies place vascular dysfunction among the earliest pathological changes on the AD trajectory, others have reported that vascular damage puts people at risk for dementia, though not necessarily for Alzheimer’s (Jul 2016 news; Nov 2019 news; Jan 2020 news). 

In their editorial, Ishii and Iadecola called it striking that the drivers of cognitive impairment would differ between ApoE4 and ApoE3 carriers. “Montagne and colleagues’ findings indicate that activation of the CypA pathway and pericyte damage might not be involved in cognitive impairment in people who carry the most common APOE variant, APOE3,” they wrote. The possibility remains that a leaky BBB caused by other factors—such as Aβ-inflicted damage to endothelial cells—might contribute to cognitive impairment in ApoE3 carriers, they added.

“This work greatly builds the literature of amyloid-independent effects of APOE4 in AD pathogenesis,” wrote Bill Rebeck of Georgetown University in Washington, D.C. “These exciting in vivo findings provide important support for the hypothesis that APOE4 increases risk of AD at least partially by increasing neuroinflammation,” he wrote, interpreting CypA, sPDGFRβ, and MMP9 as inflammatory markers.

“This is a very important paper because it provides yet more evidence for the crucial role that vascular dysfunction seems to play in Alzheimer’s disease, and also provides a plausible mechanism by which the vascular dysfunction damages the hippocampus,” commented Joanna Wardlaw of the University of Edinburgh.

Roy Weller and Roxana Carare of the University of Southampton in England called out the observation that PDGFRβ predicts cognitive impairment in ApoE4 carriers. Together with the linkage of the CypA-MMP9 pathway to BBB damage, the study implicates pericytes in cognition, they noted. “Inhibiting CypA may be an exciting therapeutic option to improve the function of pericytes in the aging brain,” they wrote.—Jessica Shugart


  1. In this paper, the authors used dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in 245 humans to demonstrate that there was breakdown of the blood-brain barrier (B4) in the hippocampi and parahippocampal gyri in cognitively normal APOE4 individuals and the degree of B4 increased with the decline in cognition. The breakdown was not observed in other areas of the brain. These findings could be relevant to the pathogenesis of Amyloid-Related-Imaging Abnormalities-E (ARIA-E), where there is evidence of sulcal hyperintensities that were presumed to be vasogenic edema, especially in APOE4 individuals. The present study suggests that the areas susceptible for B4 are different from those observed in ARIA and consistent with a hypothesis that impaired intramural periarterial drainage (IPAD) is a key pathogenic factor (Weller et al., 2005; Sperling et al., 2012). 

    Another major strength of the paper is the observation that PDGFRβ is a reliable predictor of cognitive impairment in APOE4 carriers, suggesting a key role for pericytes in cognition. Previous studies from the Zlokovic lab using APOE3/4 knock-in mice showed that APOE4 activates CypA via low-density lipoprotein receptor-related protein-1, followed by activation of MMP9 (Halliday et al., 2015) in pericytes. Inhibiting CypA may be an exciting therapeutic option to improve the function of pericytes in the aging brain, which may improve cerebral blood flow as well as facilitate IPAD and reduce expression of MMP9 (Roth et al., 2019; Carare et al., 2020). 


    . White matter changes in dementia: role of impaired drainage of interstitial fluid. Brain Pathol. 2015 Jan;25(1):63-78. PubMed.

    . Amyloid-related imaging abnormalities in patients with Alzheimer's disease treated with bapineuzumab: a retrospective analysis. Lancet Neurol. 2012 Mar;11(3):241-9. PubMed.

    . Accelerated pericyte degeneration and blood-brain barrier breakdown in apolipoprotein E4 carriers with Alzheimer's disease. J Cereb Blood Flow Metab. 2015 Mar 11; PubMed.

    . Regulator of G-protein signaling 5 regulates the shift from perivascular to parenchymal pericytes in the chronic phase after stroke. FASEB J. 2019 Aug;33(8):8990-8998. Epub 2019 May 2 PubMed.

    . Vasomotion Drives Periarterial Drainage of Aβ from the Brain. Neuron. 2020 Feb 5;105(3):400-401. PubMed.

  2. This work greatly builds the literature of amyloid-independent effects of APOE4 in AD pathogenesis. Using diverse imaging and CSF analyses, Montagne et al. show that blood-brain barrier (BBB) leakage in humans is more pronounced in APOE4 individuals with and without AD, and does not correlate with in vivo markers of Aβ or tau. Indeed, over all individuals, BBB breakdown was not linked to the accumulation of amyloid (although their measures would not preclude some amyloid angiopathy).

    These data suggest that BBB breakdown may be linked to other AD risk processes, such as inflammation. Indeed, CSF signs of neuroinflammation (CypA, sPDGFRβ, and MMP9) were elevated in the CSF of APOE4 individuals at the early stages of AD. BBB breakdown in APOE3 individuals with AD did not correlate with elevation of these biomarkers, suggesting that neuroinflammation is more pronounced in APOE4 individuals. These exciting in vivo findings provide important support for the hypothesis that APOE4 increases risk of AD at least partially by increasing neuroinflammation.

  3. This is a very important paper because it provides yet more evidence for the crucial role that vascular dysfunction seems to play in Alzheimer’s disease, and also provides a plausible mechanism by which the vascular dysfunction damages the hippocampus. These findings help explain why Alzheimer’s disease shares so many risk factors with stroke and they should encourage researchers to pay more attention to the vasculature in their experiments, and encourage clinicians to identify and manage vascular risk factors such as hypertension and diabetes when treating patients with or at risk of Alzheimer’s disease and encourage them to reduce dietary salt intake and stop smoking.

  4. Here, Montagne et al. present an elegant study that brings together elements from much of their previous work, ultimately culminating in a strong case for an APOE4-specific effect on blood-brain barrier (BBB) integrity. In this paper, they reported significant BBB breakdown in the hippocampi and parahippocampal gyri in cognitively normal APOE4 carriers compared to APOE3 homozygotes. The BBB breakdown was further increased with cognitive decline and was independent of Aβ and p-tau status. They then went on to investigate pericyte loss as a potential mechanism of this disparity and found that elevated levels of CSF sPDGFRβ predicted cognitive decline in APOE4 carriers only, and that the CypA-MMP9 pathway may be the underlying mechanism.

    This work highlights the potential impact of personalized medicine in the treatment of dementia and raises many broad questions for the field of AD research. Given that the APOE4-specific phenotypes presented here occur prior to AD pathology, could it be that we are considering AD as we once considered cancer—with a blanket diagnosis covering multiple distinct etiologies? Where do APOE3 homozygotes who develop AD fit into this proposed paradigm? Are they completely unaffected by this pathway or is there some sort of continuum? For instance, do APOE3 homozygotes with high barrier permeability function similarly to APOE4 carriers and do they also have higher CSF sPDGFRβ levels? The potential for a protective effect of APOE2 on this pathway also remains to be seen, although the rarity of the allele makes it difficult to power human studies to answer APOE2-related questions. Needless to say, there is still much work to do to define the relationship between the APOE gene and AD.

    Given the links between vascular risk factors and dementia, APOE4 and cardiovascular disease risk, and the protective effect of peripheral ApoE on coronary heart disease (Qi et al., 2018), we found it surprising that this study found no significant contributions of vascular risk factors to BBB breakdown. The effect of vascular risk factors was evaluated by comparing BBB breakdown in subjects with at least two risk factors to subjects with fewer than two risk factors. Although the authors provided a rationale for this strategy, it may have been more informative to parse apart individual vascular risk factors and include others, such as HDL-C levels, in their analysis.

    Other questions raised by this study include how ApoE from different cell types may contribute to this mechanism, as there is evidence for cell-specific effects of APOE4 on other cellular functions (Lin et al., 2018). Further, many questions remain to be answered by longitudinal studies of BBB breakdown in APOE4 carriers. For example, does BBB breakdown predict future Aβ and p-tau positivity? Do all of those with BBB breakdown go on to develop cognitive impairment? And what other factors explain why some APOE4 carriers develop BBB breakdown and cognitive impairment while others do not? Is CypA-MMP9-mediated BBB breakdown reversible and could thus be a targetable pathway for cognitive decline?

    Overall, this paper was a very thought-provoking and enticing read. It undoubtedly deepens our understanding of the APOE4 risk allele and opens up many new avenues for investigation.


    . Apolipoprotein E-containing high-density lipoprotein (HDL) modifies the impact of cholesterol-overloaded HDL on incident coronary heart disease risk: A community-based cohort study. J Clin Lipidol. 2018 Jan - Feb;12(1):89-98.e2. Epub 2017 Nov 14 PubMed.

    . APOE4 Causes Widespread Molecular and Cellular Alterations Associated with Alzheimer's Disease Phenotypes in Human iPSC-Derived Brain Cell Types. Neuron. 2018 Jun 27;98(6):1294. PubMed.

  5. In this elegant study, the authors used dynamic-contrast-enhanced magnetic resonance imaging (DCE-MRI) in 245 people to provide evidence that apolipoprotein E4 contributes to BBB breakdown in specific brain regions, specifically the hippocampi and para-hippocampal gyri, in older individuals without cognitive impairments. Furthermore, BBB breakdown is worsened in older individuals with cognitive impairment, independent of pathological hallmarks of AD. In E4 carriers, platelet-derived growth factor receptor beta (PDGFRβ) was a reliable predictive biomarker of cognitive impairment. These results raise the following important questions:

    1. At what age does BBB breakdown begin in E4 carriers?
    2. Are there sex differences involved in these E4 effects?
    3. How might E4 affect the interactions of plasma factors with the BBB?
    4. Does the brain “sense” when the BBB is disrupted and “shunt” blood flow away from this breach until it can be repaired? If this is the case, this might explain the decrease in brain volume (i.e., nutrient loss leading to atrophy).

    Our studies in ApoE targeted replacement (TR) mice generated by Patrick Sullivan suggest that the impact of E4, compared to E3, on the BBB begins early on in young adulthood (at 2 to 4 months of age) (Rhea et al., 2020; Rhea et al., 2020). At this age, insulin interactions with the BBB are already altered due to ApoE isoform and sex.

    In the frontal cortex, E4 females have higher levels of vascular insulin binding than E3 females. In the hypothalami of E3 mice, males have higher levels of vascular insulin binding than females. There are no significant effects in the hippocampus. These findings occurred in the absence of human Aβ or human tau in the model.

    The results from the current human study, combined with our mouse studies in young mice, also bring up the question of whether, in E4 individuals, the BBB alterations seen in young adulthood contribute to neurological conditions other than AD later in life. For a review on E4 being a risk factor for other neurological conditions, such as stroke, vascular dementia, multiple sclerosis, and Parkinson’s disease, please see Verghese et al., Lancet Neurology 2011, and Giau et al, Neuropsychiatr Dis 2015, and even in long-term effects following exposure to SARS-CoV-2.

    Clearly, increased efforts are warranted to address these timely questions.


    . Insulin BBB pharmacokinetics in young apoE male and female transgenic mice. PLoS One. 2020;15(1):e0228455. Epub 2020 Jan 31 PubMed.

    . ApoE and cerebral insulin: Trafficking, receptors, and resistance. Neurobiol Dis. 2020 Apr;137:104755. Epub 2020 Jan 21 PubMed.

    . Apolipoprotein E in Alzheimer's disease and other neurological disorders. Lancet Neurol. 2011 Mar;10(3):241-52. PubMed.

    . Role of apolipoprotein E in neurodegenerative diseases. Neuropsychiatr Dis Treat. 2015;11:1723-37. Epub 2015 Jul 16 PubMed.

Make a Comment

To make a comment you must login or register.


Webinar Citations

  1. Leaky Blood-Brain Barrier a Harbinger of Alzheimer's?

News Citations

  1. Absent Aβ, Blood-Brain Barrier Breakdown Predicts Cognitive Impairment
  2. Does the Blood-Brain Barrier Stand Up to Alzheimer’s? Study Finds No Breach
  3. ApoE4 Makes Blood Vessels Leak, Could Kick Off Brain Damage
  4. LOAD of Data Place Vascular Malfunction as Earliest Event in Alzheimer’s
  5. Already in Mid-30s, Poor Vascular Health Means Small Brain at 70
  6. Vascular Dysfunction Taxes Cognition, but Not Via Amyloid, AD

Paper Citations

  1. . Neurovascular unit impairment in early Alzheimer's disease measured with magnetic resonance imaging. Neurobiol Aging. 2016 Sep;45:190-6. Epub 2016 Jun 17 PubMed.
  2. . Blood-Brain Barrier Leakage in Patients with Early Alzheimer Disease. Radiology. 2016 Nov;281(2):527-535. Epub 2016 May 31 PubMed.
  3. . Accelerated pericyte degeneration and blood-brain barrier breakdown in apolipoprotein E4 carriers with Alzheimer's disease. J Cereb Blood Flow Metab. 2015 Mar 11; PubMed.
  4. . CSF evidence of pericyte damage in Alzheimer's disease is associated with markers of blood-brain barrier dysfunction and disease pathology. Alzheimers Res Ther. 2019 Sep 14;11(1):81. PubMed.
  5. . Blood-brain barrier breakdown is an early biomarker of human cognitive dysfunction. Nat Med. 2019 Feb;25(2):270-276. Epub 2019 Jan 14 PubMed.
  6. . A novel sensitive assay for detection of a biomarker of pericyte injury in cerebrospinal fluid. Alzheimers Dement. 2020 Jun;16(6):821-830. Epub 2020 Apr 16 PubMed.
  7. . Cyclophilin A is a proinflammatory cytokine that activates endothelial cells. Arterioscler Thromb Vasc Biol. 2004 Jul;24(7):1186-91. PubMed.
  8. . Apolipoprotein E controls cerebrovascular integrity via cyclophilin A. Nature. 2012 May 24;485(7399):512-6. PubMed. Correction.

Further Reading

No Available Further Reading

Primary Papers

  1. . APOE4 leads to blood-brain barrier dysfunction predicting cognitive decline. Nature. 2020 May;581(7806):71-76. Epub 2020 Apr 29 PubMed.
  2. . Risk factor for Alzheimer's disease breaks the blood-brain barrier. Nature. 2020 May;581(7806):31-32. PubMed.