Compared to the vast majority of the population carrying the more common E3 and E4 isoforms of ApoE, the few who have an ApoE2 allele enjoy reduced risk of late-onset Alzheimer’s disease. Now, a new study sheds light on the underpinnings of this protection. Analyzing cognitively normal subjects enrolled in the Alzheimer’s Disease Neuroimaging Initiative (ADNI), researchers report that E2 carriers not only have biomarker signatures indicating less AD pathology, but also lower hippocampal atrophy rates, relative to E3 homozygotes. Cognitive decline did not differ between the two groups over the study’s 24-month timeframe, however. The work “provides evidence that the protective effect of the ApoE2 genetic polymorphism is detectable in vivo, before there is evidence of cognitive impairment,” the authors write. Gloria Chiang of the University of California, San Francisco, and colleagues published the findings online October 27 in the journal Neurology.

Only 5 percent of the population carries an E2 allele (see Huang, 2006). Those who do are less likely to develop late-onset AD (Corder et al., 1994) and seem to have slower rates of cognitive decline (Wilson et al., 2002). “We were trying to find a structural biological basis for these observations,” Chiang told ARF.

Of the 193 normal ADNI participants with reliable magnetic resonance imaging (MRI) data, 27 were E2 carriers (i.e., E2/E3 or E2/E2) and 111 were E3/3. The researchers compared hippocampal atrophy and cognitive decline in both groups, and also analyzed cerebrospinal fluid (CSF) biomarkers in those who agreed to spinal taps—about half the participants. This paper did not report on PiB-PET imaging in these volunteers.

The study found that, relative to E3 homozygotes, E2 carriers had baseline CSF profiles reflecting little or no AD pathology—that is, higher Aβ and lower phospho-tau levels. This confirms another recent study that found higher CSF levels of Aβ42 in normal E2-positive seniors (Morris et al., 2010), and seems to jibe with data linking E2 to decreased Aβ accumulation (Kim et al., 2009).

“What is new (in the present study) is that the rate of neurodegeneration, as marked by hippocampal atrophy, was lower in E2-positive people,” noted David Holtzman, Washington University School of Medicine, St. Louis, Missouri, in an e-mail to ARF. Atrophy rates averaged 0.5 percent per year among E2 carriers, compared with 1.3 percent for the E3/3 group. Taken together with the CSF data, Holtzman wrote, the findings suggest “that ‘preclinical’ AD pathology, which is less in E2-positive people, is associated with reduced downstream brain injury.”

On cognition, Chiang said she “expected to find some trend, maybe not in global measures, but at least in episodic memory,” since this was the cognitive domain that distinguished E2 carriers from E3/3 and E3/4 groups in a previous eight-year prospective study (Wilson et al., 2002). However, E2 carriers’ annual decline in overall cognition (measured by the ADAS-cog) and episodic memory (measured by delayed paragraph recall in the Wechsler Memory Scale-Revised) was statistically indistinguishable from the E3/3 group.

Chiang attributes this in part to the study’s short timeframe (two years), but also to the fact that the participants came in with normal brain function. “We wouldn’t expect their cognition to change substantially in a couple of years,” she said. The authors note, in addition, that the ADNI cohort was “more highly educated, and had fewer comorbidities than a community population at this age”—factors that may have further slowed cognitive decline in this study. Other researchers have also noted that ADNI volunteers may not be representative of the general population.

In an accompanying commentary, Richard Caselli and Amylou Dueck of Mayo Clinic Arizona, Scottsdale, highlight several lines of research that complicate analysis of the current findings. One is the recent work of Allen Roses, Duke University, Durham, North Carolina, and colleagues, which “clouds the interpretation of all ApoE-based analyses,” Caselli and Dueck write. Roses and colleagues have identified a gene, Tomm40, that is co-inherited with APOE, and appears to predict LOAD onset age (Roses et al., 2009; ARF related Las Vegas conference story; ARF related Honolulu conference story). ApoE/Tomm40 interactions in LOAD may stem from the role of these two proteins in mitochondrial dynamics (see Roses’ comment below), though this interpretation is controversial, as other scientists point out that an alternative explanation has less to do with Tomm40 function and more with variation in ApoE4 expression.

Another intriguing observation, from the 90+ Study of nonagenarians, seems to turn the tables on the apparent link between the E2 isoform and reduced AD pathology. Non-demented E2 carriers had higher amyloid load than did E3/3 individuals (Berlau et al., 2009). Furthermore, in studies with a new AD mouse strain made by coupling ApoE targeted replacement mice with Bob Vassar’s 5xFAD model, Mary Jo LaDu and colleagues at the University of Illinois, Chicago, have also found more extracellular Aβ deposition in ApoE2 TR/5xFAD mice relative to E3- and E4-TR counterparts. In contrast, intracellular Aβ, which may be more toxic, was most abundant in the E4 mice. Katherine Youmans, a graduate student in the LaDu lab, reported the findings earlier this summer at a one-day ApoE symposium in St. Louis, Missouri (see ARF related conference story and Youmans comment below).

On a broader level, Yadong Huang of the Gladstone Institute of Neurological Disease, San Francisco, notes that the ApoE2 allele is understudied. This is not surprising for research in humans, because the allele is so rare. Even animal studies tend to be few and far between, he said, possibly because E2’s impact on cognition is small and often hard to measure. “In our own experience, we rarely get clear cognitive differences between E2 and E3 animals. And there is no place to publish negative results,” Huang said. In light of these difficulties, the “current paper adds good evidence to support that E2 is a protective allele,” he said. “It’s a very important study.”

In a follow-up to the current study, Chiang and colleagues await further longitudinal data to see whether E2 carriers ultimately show slower rates of mental decline, compared to people with E3 or E4 alleles.—Esther Landhuis

Comments

  1. This new work from Chiang et al. is important because it emphasizes our lack of understanding regarding the neuroprotective roles of ApoE2. For years, researchers have tried to establish the link between E4 carriers or non-carriers and brain region-specific neuronal loss, particularly in a diseased state. Thus, understanding the contribution of ApoE genotype to brain atrophy in the presence and absence of neurodegenerative pathology is not novel. However, previous research has primarily centered on the E3 versus E4 alleles, with little attention paid to individuals harboring E2 (albeit largely due to the rarity of elderly E2 carriers). Chiang not only found a decreased rate of hippocampal atrophy in the presence of ApoE2, but an increase in CSF Aβ levels.

    Although these results are suggestive, their interpretation is limited because the distinction between Aβ42 versus Aβ40 or other species was not made. In addition, the human cohort analyzed by Chiang et al. spanned 35 years and contained only 17 E2 participants for the biomarker analyses. Although this is understandable, it is important for researchers to know the subdivision of ages as they relate to ApoE isoform. This would allow us to determine whether the results were skewed due to a predominance of younger (i.e., healthier) individuals with an E2 allele, as older individuals would reasonably be expected to show both increased rates of atrophy and decreased Aβ clearance. In our lab’s new mouse model of Aβ pathology, ApoE has no isoform-specific effects on brain Aβ40 levels but does alter Aβ42 levels and solubility. Importantly, mice homozygous for ApoE4 have increased levels of soluble Aβ42, while levels remain low with ApoE2. This suggests an increased clearance of soluble Aβ42 out of the brain parenchyma in the presence of ApoE2 that does not occur, or is impaired, with ApoE4. Chiang’s recent findings support our theory. Additional data from our lab indicate that both ApoE2 and ApoE4 correlate with increased Aβ42 accumulation, although ApoE2 may alter the nature of Aβ42 deposition toward diffuse versus compact amyloid, particularly in disease-vulnerable regions. Substantial Aβ pathology has also been observed in E2 carriers older than 90 years of age, further supporting our hypothesis and suggesting that the clearance of Aβ42 into the CSF may ultimately depend on the body’s ability to break up and remove various conformational species of the peptide, which could vary significantly with age and ApoE genotype. Hopefully, the data from our and Chiang’s labs regarding the pathological effects of ApoE2 will articulate to others the need for comprehensive analyses that include all three isoforms of ApoE, if we are to understand the mechanism(s) underlying the onset and progression of AD pathology.

    View all comments by Katherine Youmans
  2. It is not surprising at all that there is a difference in the "hippocampal atrophy rates and less AD pathology in CSF of ApoE2 carriers (relative to E3/E3 participants) from the ADNI cohort of cognitively normal individuals." The critical contributing factors to the progression of disease is the ApoE isoform/Tomm40 interaction rates. ApoE2 contains two sulfhydryl groups, while ApoE3 contains one, and none on ApoE4. Huang and colleagues have demonstrated exquisite evidence of the facts that were not believed in the 1990s—that ApoE is expressed in the neuronal cytoplasm, that the terminal 20 amino acids are cleaved, allowing ApoE(1-272) to bind to the outer mitochondrial membrane at the Tomm40 site (see Chang et al., 2005). The degree of binding is ApoE4>ApoE3, with no binding of ApoE2 due to a change of secondary protein structure with two cysteine groups replacing two arginines. Compared to ApoE3[1-272), the binding of ApoE4(1-272) causes decreased mitochondrial dynamics [(speed and distance moved). ApoE2 has no (273-299] cleavage and presumably has no effect on mitochondrial dynamics. Thus two APOE3 alleles is worse than one APOE3 and one APOE2 allele.

    From our current view of the pathophysiology of AD and its relationship to the TOMM40 Rs10524523 nucleotide polymorphism and APOE genotypes that are inherited together throughout evolution, the age of onset appears to be determined by the TOMM40 Rs10524523 variable polyT polymorphisms on each DNA strand. APOE2 is known to be derived from APOE3 strands. It is too early to say in Caucasians, but we have data to demonstrate that APOE2 is connected to the very long form (523 VL) of Tomm40 in both Caucasians and African populations. Studies are ongoing to assess the proportion of 523-VL, versus 523-Short, that predict an earlier age of onset compared to the age of onset curves for the APOE3/3 homozygotes: VL/VL and S/S. Preliminary evidence shows that when an APOE2/3 patient develops AD, the VL polymorphisms predominate. With many more APOE2/3 carriers we will be able to construct age of onset distributions for this generally later group of patients. To date, we have demonstrated that 523-VL/APOE2 is enriched in AD patients compared to 523-S/APOE2 carriers. This is now still a hypothesis but it is consistent with the viewpoint and data in hand. Damaged mitochondrial undergo an apoptotic process, releasing cytochrome c, which stimulates the caspase pathways. This can be viewed as turning on the "so-called" amyloid cascade with the proximal problem being slow toxicity of the mitochondria over years. How 523 affects expressed Tomm40 channel protein is currently also under study. In yeast and other genomes, intronic polyT variants can affect the expression levels, with an increase in number leading to progressive decreases in protein expression relative to other genes. It is also possible that VL polyTs can lead to alternative splicing, thus changing the sequence of the Tom40 channel as well.

    From this point of view, the report from ADNI is consistent, at least in this important location, with the current data available about the etiological process underlying AD.

    References:

    . Lipid- and receptor-binding regions of apolipoprotein E4 fragments act in concert to cause mitochondrial dysfunction and neurotoxicity. Proc Natl Acad Sci U S A. 2005 Dec 20;102(51):18694-9. PubMed.

    View all comments by Allen Roses

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Las Vegas: AD, Risk, ApoE—Tomm40 No Tomfoolery
  2. Honolulu: Tomm40 Reported to Track With Brain Atrophy, Cognition
  3. St. Louis: ApoE—Receptors, Theories and Therapies

Paper Citations

  1. . Apolipoprotein E and Alzheimer disease. Neurology. 2006 Jan 24;66(2 Suppl 1):S79-85. PubMed.
  2. . Protective effect of apolipoprotein E type 2 allele for late onset Alzheimer disease. Nat Genet. 1994 Jun;7(2):180-4. PubMed.
  3. . The apolipoprotein E epsilon 2 allele and decline in episodic memory. J Neurol Neurosurg Psychiatry. 2002 Dec;73(6):672-7. PubMed.
  4. . APOE predicts amyloid-beta but not tau Alzheimer pathology in cognitively normal aging. Ann Neurol. 2010 Jan;67(1):122-31. PubMed.
  5. . The role of apolipoprotein E in Alzheimer's disease. Neuron. 2009 Aug 13;63(3):287-303. PubMed.
  6. . A TOMM40 variable-length polymorphism predicts the age of late-onset Alzheimer's disease. Pharmacogenomics J. 2010 Oct;10(5):375-84. Epub 2009 Dec 22 PubMed.
  7. . APOE epsilon2 is associated with intact cognition but increased Alzheimer pathology in the oldest old. Neurology. 2009 Mar 3;72(9):829-34. PubMed.

External Citations

  1. Alzheimer’s Disease Neuroimaging Initiative
  2. 90+ Study
  3. 5xFAD

Further Reading

Papers

  1. . Apolipoprotein E and Alzheimer disease. Neurology. 2006 Jan 24;66(2 Suppl 1):S79-85. PubMed.
  2. . Protective effect of apolipoprotein E type 2 allele for late onset Alzheimer disease. Nat Genet. 1994 Jun;7(2):180-4. PubMed.
  3. . APOE predicts amyloid-beta but not tau Alzheimer pathology in cognitively normal aging. Ann Neurol. 2010 Jan;67(1):122-31. PubMed.
  4. . The apolipoprotein E epsilon 2 allele and decline in episodic memory. J Neurol Neurosurg Psychiatry. 2002 Dec;73(6):672-7. PubMed.
  5. . The role of apolipoprotein E in Alzheimer's disease. Neuron. 2009 Aug 13;63(3):287-303. PubMed.
  6. . A TOMM40 variable-length polymorphism predicts the age of late-onset Alzheimer's disease. Pharmacogenomics J. 2010 Oct;10(5):375-84. Epub 2009 Dec 22 PubMed.

Primary Papers

  1. . Hippocampal atrophy rates and CSF biomarkers in elderly APOE2 normal subjects. Neurology. 2010 Nov 30;75(22):1976-81. PubMed.
  2. . APOE varepsilon2 and presymptomatic stage Alzheimer disease: how much is not enough?. Neurology. 2010 Nov 30;75(22):1952-3. PubMed.