The consensus view among researchers holds that, in general, late-onset Alzheimer’s disease begins when clearance of the amyloid-β peptide fails to keep up with normal production, allowing the peptide to accumulate over time. Now, a paper in the March 7 Nature Communications suggests that, in some people, a different force may tip the scales. Researchers led by Niklas Mattsson and Oskar Hansson at Lund University, Malmö, Sweden, report that in the early stage of sporadic AD, some people produce more Aβ in the brain than do controls. This was most noticeable in people who tested negative for ApoE4, a major genetic risk factor for the disease. “We hypothesize that between-person variation in APP processing may be associated with brain amyloid deposition,” Mattsson told Alzforum. “We believe that this variation explains a minor part of the overall amyloid deposition, and it may be easiest to detect it in ApoE4-negative people.”

Randall Bateman, who studies the kinetics of APP processing at Washington University, St. Louis, said that he was a little surprised at the result. “No one had ever linked AD to increased CSF Aβ peptides,” he said. However, he told Alzforum that he has since corroborated the findings in a WashU cohort.

How does this new data jibe with current assumptions about Aβ biomarkers? Overwhelming evidence indicates that CSF levels of Aβ42 are lower in AD patients than in controls (see Alzbiomarker meta analysis). Researchers believe this reflects accumulation of this peptide in plaques as the disease progresses. A similar, but much weaker relationship holds for Aβ40

However, that biomarker data compares AD patients with controls. Mattsson emphasized that he looked for correlation with plaques in healthy controls and in people with preclinical AD. The researchers theorized that any uptick in APP processing in sporadic AD that led to plaques would cause an increase in levels of Aβ40 in the brain and hence in the CSF. They looked to see if CSF Aβ40 correlated with the presence of amyloid in the brain measured by PET imaging. They reasoned no such relationship would exist for Aβ42, because this sticky peptide gets trapped in brain tissue before it reaches the CSF. They also figured that ApoE4 would modulate CSF Aβ40 because it restricts clearance of Aβ from the parenchyma.

The researchers analyzed data from 120 cognitively normal controls, 102 people with subjective cognitive decline, and 108 people with mild cognitive impairment. All took part in the Swedish BioFINDER biomarker study. Though there was considerable variation in the data, regression analysis predicted that higher CSF Aβ40 equated with more amyloid in the brain as measured by flutemetamol PET. The relationship was stronger in those who had no ApoE4 allele. The authors found the same relationship between CSF Aβ38 and plaques. “We are not aware of any previous study of this size that has tested the association between CSF Aβ40 and amyloid PET imaging in early stage AD, while also adjusting for APOE e4,” noted Mattson.

New data from Bateman’s group supports the analysis. He and WashU colleague Bruce Patterson re-analyzed data from 50 healthy controls and 50 AD patients. “We have replicated some of their findings of an association of Aβ40 concentrations and amyloidosis in our data set,” Bateman wrote to Alzforum. He noted that the effect was so small that it was only apparent using multivariate analyses.

What does this new data say about the etiology of AD? Previously, using isotopic labelling techniques to measure the production and half-life of Aβ species in the brain, Bateman and researcher colleagues reported that clearance of Aβ slows down dramatically in sporadic AD (see July 2010 conference news). The new data suggest that, in some people at least, overproduction of Aβ may also be a driving force. “We know that a 30 percent increase in Aβ production in familial AD takes decades to manifest as plaques, so it would fit that a much smaller increase would take even longer in sporadic cases,” Bateman said. Mattsson has since analyzed data from another 100 people and he said the results still hold up in that extended cohort.

What would cause the Aβ uptick in sporadic AD? Mattsson said that was unclear. “We don’t know the exact mechanism behind these findings, however, we found similar effects for CSF Aβ38 and Aβ40, which would be consistent with increased BACE activity,” he said. Researchers have long known that that the amount and activity of this protease, which catalyzes the rate-limiting step in clipping Aβ peptides from their APP precursor, increases in AD, but no one had linked that to Aβ production (see Sep 2002 news). Mattsson cautioned that measuring Aβ40 is an indirect way to gauge APP metabolism. There is no way to tell, for example, if CSF Aβ40 reflects APP processing or some other event, such as the release of Aβ from neurons due to neural activity. “In the future it would be interesting to test other CSF Aβ species and soluble APP fragments that may give more information about specific alterations in APP metabolism,” he said.—Tom Fagan


  1. Interestingly, last year, in Maia et al., we have shown that in three different APP transgenic mice, Aβ deposition seems to be ruled distinctively. In the initial stage of amyloid build-up, increased processing of APP seems more relevant, and in the second stage of amyloid progression sequestration outbalances production. This is true for both Aβ40 and 42.

    Our results are particularly curious since we were able to depict an increase in both CSF Aβ peptides, coincident with the start of Aβ deposition and before the massive amyloid deposition in all three mouse models studied. It would be interesting to assess if in this set of patients a similar biphasic profile occurs, as this would constitute a major finding with clinical implications: the earliest biomarker change in pre-clinical AD potentially impacting patient stratification and patient selection for clinical trials in pre-clinical AD. 


    . Increased CSF Aβ during the very early phase of cerebral Aβ deposition in mouse models. EMBO Mol Med. 2015 May 15;7(7):895-903. PubMed.

Make a Comment

To make a comment you must login or register.


Biomarker Meta Analysis Citations

  1. Alzheimer's Disease vs Control: Aβ42 (CSF)
  2. Alzheimer's Disease vs Control: Aβ40 (CSF)

News Citations

  1. Honolulu: Wake-Up Call—Aβ Clearance, Not Production, Awry in AD
  2. BACE Above Base in Alzheimer’s Patients

Further Reading

Primary Papers

  1. . Increased amyloidogenic APP processing in APOE ɛ4-negative individuals with cerebral β-amyloidosis. Nat Commun. 2016 Mar 7;7:10918. PubMed.