Known gene mutations explain but a fraction of early onset Alzheimer’s disease. A majority of people who become symptomatic before age 65 have neither an autosomal-dominant inheritance pattern nor mutation. So what is causing their dementia? In the May 28 JAMA Neurology, researchers led by Thomas Wingo, Emory University School of Medicine, Atlanta, proffer a new candidate risk gene: apolipoprotein B. In a case-control analysis, they found that people with EOAD were more likely to have high levels of low-density lipoprotein cholesterol in their plasma. Rare coding variants in APOB, which encodes the major LDL protein, were also more common in these patients. The findings suggest a direct link between APOB variants and EOAD risk, Wingo and colleagues concluded.

  • Early onset AD linked to high plasma LDL cholesterol.
  • ApoB, the major component of LDL, was also elevated.
  • Rare variants in the APOB gene correlated with EOAD.

“Overall, this is an important study that provides the first evidence that rare genetic coding variants of APOB are strongly associated with EOAD,” wrote Makoto Ishii, Weill Cornell Medicine, New York, in a JAMA Neurology editorial. Eric Larson, Kaiser Permanente Washington Health Research Institute, Seattle, congratulated the authors for assembling so many EOAD cases, but cautioned that interpretation may be tricky. “APOB variants likely have associations with survival and, of course, a lot of people are being treated with lipid-lowering drugs as they age, which would complicate that relationship,” Larson noted, adding, “There may be a lot more to be learned about this association” (see full comment below).

EOAD differs from late-onset AD (LOAD) in manifesting clinically before age 65. Though it accounts for only about 5 percent of all AD cases, EOAD led to the discovery of the three known autosomal-dominant AD genes, which code for amyloid precursor protein, presenilin 1, and presenilin 2. Even though variants in these genes explain fewer than 10 percent of early onset cases, investigations into other genetic causes of EOAD languished while geneticists focused on common variants that cause LOAD (Cacace et al., 2016). To identify causes of EOAD, Wingo and colleagues focused on the influence of circulating cholesterol, a known risk factor for dementia.

The researchers studied 170 cases and 211 controls at Emory’s Alzheimer’s Disease Research Center as their discovery data set, and 94 cases and 87 controls in a replication set from the ADRC at the University of California, San Francisco, and 390 cases and 1,173 controls in a second replication set amalgamated from 27 other ADRCs. Wingo first eliminated from consideration 23 people with known pathogenic mutations in APP, PS1, or PS2, and discounted any effect from other rare coding mutations in any of these three AD genes. Then he looked for associations with plasma lipids. He found that at an average of 203 mg/dL versus 178 mg/dL, LDL-cholesterol was higher in cases than in controls. So was plasma ApoB, which averaged 82 mg/dL in cases and 69 mg/dL in controls. The case-control disparity held after adjusting for sex, educational attainment and ApoE4, a strong risk factor for LOAD. EOAD did not associate with HDL-cholesterol, and only weakly with triglyceride levels.

Could variants in APOB, which are well established to affect LDL levels, explain these case-control differences? The researchers deep-sequenced the ApoB gene to find missense or nonsense mutations. Among 381 people in the Emory data set, ApoB variants were enriched in EOAD. This held true when the researchers analyzed data from 1,548 samples from the 27 other ADRCs, but not in the 173 samples from UCSF, which the authors attributed to small sample size. Meta-analysis of samples from all three data sets indicated a strong link between APOB variants and EOAD. The scientists found 33 missense or nonsense variants in 31 of 654 EOAD patients, two of whom carried two different variants each. No variant popped up in more than one person; the same went for the 24 variants found in 24 of 1,471 controls.

“Given that both ApoE and ApoB strongly influence LDL levels, the study shines a light on cholesterol metabolism in EOAD, and particularly at LDL as warranting further investigation,” senior author Allan Levey from Emory told Alzforum. Here, the authors were unable to attribute risk with specific gene and protein variants. “With very rare variants, you typically don’t have enough statistical power to tease out such associations,” said Levey. To do that, Wingo is planning a Mendelian randomization approach in a much larger sample. “Hopefully then we’ll be able to correlate risk with ApoB variants,” Levey said. Some of the variants may be protective, he noted.

The story is likely to go beyond ApoB. In their analysis the researchers also found a strong association between EOAD and cholesterol in people whose ApoB variants were normal, suggesting other factors contribute. ApoE is one such factor, accounting for 10 percent of risk for EOAD, as has been seen previously (Bellenguez et al., 2017). Mutations in APP, PS1, or PS2 were found in only 23 of the cases. Two were found in APP, three in PS2, and 18 in PS1. Statistically, this meant only PS1 variants associated with EOAD in this data set. “This illustrates how rare these mutations are and how unlikely their association would reach genome-wide significance without much larger sample sizes,” said Levey. “At the same time, variants in ApoE and ApoB do achieve significance in explaining early onset AD.”

Levey thinks these findings are informative for LOAD, which has been linked to elevated plasma cholesterol levels. However, Ishii noted that a prior Finnish study found no association between LOAD and plasma ApoB (Tynkkynen et al., 2016). Still, he thinks this new study highlights how EOAD studies can uncover variants that underlie early and late forms of AD. “Instead of excluding or ignoring EOAD owing to its low prevalence or perceived differences from LOAD, there should be a concerted effort to conduct large-scale comprehensive studies on EOAD,” he wrote.

The Longitudinal Early Onset Alzheimer’s Disease Study, lead by Liana Apostolova at Indiana University, Indianapolis, will follow EOAD patients over two years, testing for cognition, genetics, and imaging and fluid biomarkers, and will compare with the same data from LOAD studies.—Tom Fagan


  1. It makes a lot of sense to look at early onset cases for potential genetic associations because early onset is clearly more strongly “genetic.” If we better understood how these variants work mechanistically, we might be able to design prevention or treatment.

    It’s good that Wingo and colleagues were able to assemble so many EOAD cases, however, I have some concerns about their choice of controls and how different the controls were from the cases they assembled. ApoE and ApoB variants likely have associations with survival and, of course, a lot of people as they age are being treated with lipid-lowering drugs which would affect that relationship. There may be a lot more to be learned about this association in my judgment.

Make a Comment

To make a comment you must login or register.


Paper Citations

  1. . Molecular genetics of early-onset Alzheimer's disease revisited. Alzheimers Dement. 2016 Jun;12(6):733-48. Epub 2016 Mar 24 PubMed.
  2. . Contribution to Alzheimer's disease risk of rare variants in TREM2, SORL1, and ABCA7 in 1779 cases and 1273 controls. Neurobiol Aging. 2017 Nov;59:220.e1-220.e9. Epub 2017 Jul 14 PubMed.
  3. . Apolipoproteins and HDL cholesterol do not associate with the risk of future dementia and Alzheimer's disease: the National Finnish population study (FINRISK). Age (Dordr). 2016 Dec;38(5-6):465-473. Epub 2016 Sep 23 PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Association of Early-Onset Alzheimer Disease With Elevated Low-Density Lipoprotein Cholesterol Levels and Rare Genetic Coding Variants of APOB. JAMA Neurol. 2019 Jul 1;76(7):809-817. PubMed.
  2. . Apolipoprotein B as a New Link Between Cholesterol and Alzheimer Disease. JAMA Neurol. 2019 Jul 1;76(7):751-753. PubMed.