The concentration of Aβ in the cerebrospinal fluid climbs temporarily just as the first plaques appear in the brain. Only after scaling this little peak does it gradually begin its well-known downward slope as plaques hog more Aβ and clog up the interstitial spaces in the brain. This blip upward is the new finding in a mouse study in the May 15 EMBO Molecular Medicine by Mathias Jucker, University of Tubingen, Germany, and colleagues. It dovetails with recent observations in people with familial AD mutations, and could indicate that this CSF Aβ spike represents the earliest stage of preclinical AD. Exploiting this biphasic pattern for diagnosis will be tricky, researchers agreed, as levels will be the same on the way up as they are on the way back down. Jucker thinks there may be other markers associated with either the ebb or flow of Aβ that might prove more useful than the wave itself.

Accumulation of amyloid in the brain precedes the onset of dementia by about two decades. To catch people just as amyloid settles in, researchers are looking for biomarkers that may herald the beginning of the build-up. It is established that as people inch nearer to dementia, CSF levels of Aβ drop while tau goes up (see Jul 2012 news and Shaw et al., 2009). Previous work revealed that in transgenic mice, too, CSF Aβ dropped and tau rose as Aβ deposition rose in the brain (see Jul 2013 conference newsKawarabayashi et al., 2001Liu et al., 2004).

However, recent evidence from two studies that track biomarkers in asymptomatic people with familial AD mutations—the Dominantly Inherited Alzheimer’s Network (DIAN) and the Alzheimer Prevention Initiative (API)—have found that just prior to the characteristic drop in CSF Aβ40/42, something different happens: Aβ40 goes up (see Reiman et al., 2012, and Mar 2014 news). Might this blip also occur in mice? If so, researchers could study it more carefully.

Jucker’s previous mouse study had correlated CSF Aβ levels with rising brain amyloid. In this one, first author Luis Maia measured CSF Aβ before plaques appeared. The researchers compared three transgenic mouse models—APP23, APP24, and APP51—in which amyloid crops up at different ages. They found that in all three models, the CSF concentration of Aβ40 and Aβ42 peaked just prior to the first appearance of amyloid plaques, then dropped, creating an inverted U-shaped curve.

Bump in the Road to AD.

An uptick in CSF Aβ40/42 (yellow/red) precedes the rise in amyloid deposition in the brain (green line) in APP23 mice. [Image courtesy of Maia et al., 2015.]

In APP23 mice, the CSF biomarkers reached their zenith at 8 months—peaking at 22 percent above levels found in 3-month-old mice. Immunoassays of brain extracts showed a rise in Aβ40 and Aβ42 at this time. Immunohistochemistry on brain slices indicated that the first plaques appeared by 6 months of age, but did not exceed one plaque per sagittal brain section until 8 months. By 12 months, CSF Aβ40 and Aβ42 had dropped back to the levels of 3-month-old mice, and then continued to decline. Meanwhile, plaques took off.

Similar trends occurred in the APP24 and APP51 mice. CSF Aβ40 and Aβ42 peaked in APP24 mice at 3 to 4 months of age—the same time when at least one plaque per sagittal section popped up. In APP51 mice, the CSF peak came later, at 15 months, a time when both immunohistochemistry and immunoassay of brain extracts first revealed amyloid accumulation in the brain. In all three mouse models, CSF Aβ42 dropped more steeply than Aβ40 as plaques accumulated in the brain.

Maia proposes a two-stage process. “We think the first stage, when CSF Aβ climbs, is governed by a rise in APP processing; the second stage, when CSF Aβ falls, is driven by sequestering of Aβ in plaques,” he told Alzforum. It makes sense that CSF Aβ42 dropped faster than Aβ40, Maia said, as it is predominantly sequestered in plaques.

“A primary endpoint of most preclinical Alzheimer mouse studies is postmortem neuropathology, which is arguably not very useful for translational purposes,” Jucker wrote to Alzforum. “I am pleased to see that CSF measurement (and hopefully blood will follow) in mice begins to generate meaningful results for clinical studies.” 

The rise in CSF Aβ could signify a time when elevated levels of Aβ in the brain are just starting to take a toll on neurons, speculated Henrik Zetterberg of the University of Gothenberg in Sweden, who was not involved in the study. In collaboration with Frances Edwards at University College London, Zetterberg found that synaptic damage occurred in mice as Aβ levels rose, but before plaques emerged (see Cummings et al., 2015). “One interesting hypothesis is that increasing Aβ levels might affect synapses, which might simulate further Aβ production in an almost auto-catalytic cascade, which is then curbed by frank plaque deposition,” he wrote to Alzforum. He added that future studies could include non-amyloidogenic Aβ species, such as Aβ38, in CSF measurements to truly assess changes in Aβ production.

Takaomi Saido of the RIKEN Brain Science Institute in Wako, Japan, offered another explanation. Work in his lab indicates that autophagy mediates the secretion of Aβ from neurons (see Nilsson et al., 2013Nilsson et al., 2015). “The increased CSF Aβ levels from younger APP Tg mice could at least partially be due to increased autophagy,” he wrote to Alzforum. He said it would be interesting to test this hypothesis, adding that autophagy is known to rise in the early stages of AD in people.

How about humans? Emerging evidence from the DIAN study indicates elevated CSF Aβ concentrations in mutation carriers prior to plaque deposition; however, data published thus far are cross-sectional, not longitudinal. Hints of the trend have even surfaced in people who do not carry causal mutations. At the AD/PD conference in Nice, Mony de Leon of New York University presented findings from 99 people whose CSF Aβ42 levels were measured more than three times for up to a decade. In 11 of them, de Leon observed an increase in CSF Aβ42 followed by a decrease some years later, which correlated with cognitive decline. He told Alzforum that the numbers and sampled time windows were still too small to draw meaningful conclusions, but added that both the uptick in CSF Aβ42 and the decrease were correlated with an increased risk for future cognitive decline and imaging evidence for progressive brain damage.

Will the Aβ bump be useful to clinicians searching for signs of early AD? Only with sufficient longitudinal data, Maia said. Because CSF Aβ40 and Aβ42 concentrations will be the same on the way up and back down, researchers cannot know with a one-time measurement which side of the hump a person is on.

Maia said that if the Aβ boost occurs in people, it would be best to find other biomarkers that coincide with the rise and fall of CSF Aβ. “We need additional biomarkers at these two time points,” he said. Jucker’s lab is searching in mice. If successful, the researchers hope their candidates can be checked in familial AD and sporadic cohorts. This would yield a much earlier tool for stratifying patients as they enter prevention trials.—Jessica Shugart


  1. Previously, we found that Aβ secretion from neurons is mediated by autophagy. In addition, it has been shown previously that autophagosomes contain both Aβ and the Aβ-generating enzymes (Nilsson et al., 2013; Nilsson et al., 2015). The increased levels of Aβ in CSF from younger APP Tg mice which the authors have found in this study could, at least partially, possibly be due to increased autophagy in the early stage of the APP mice—that is, if we assume CSF to be an indicator of brain ISF.

    Data from humans has indeed shown an upregulation of autophagy in early AD, for example as a response to an imbalance in proteostasis or oxidative stress. Therefore it would be interesting to know if the initial increase in Aβ secretion is paralleled by an increase of autophagy, which could be investigated biochemically in the APP Tg mice (Nilsson and Saido, 2014).


    . Aβ secretion and plaque formation depend on autophagy. Cell Rep. 2013 Oct 17;5(1):61-9. PubMed.

    . Autophagy-related protein 7 deficiency in amyloid β (Aβ) precursor protein transgenic mice decreases Aβ in the multivesicular bodies and induces Aβ accumulation in the Golgi. Am J Pathol. 2015 Feb;185(2):305-13. Epub 2014 Nov 26 PubMed.

    . Dual roles for autophagy: degradation and secretion of Alzheimer's disease Aβ peptide. Bioessays. 2014 Jun;36(6):570-8. Epub 2014 Apr 8 PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Paper Alert: DIAN Biomarker Data Show Changes Decades Before AD
  2. Cerebrospinal Fluid Tau Climbs in Aβ Mouse Models
  3. DIAN Longitudinal Data Surprises With Late Drop in Tau

Research Models Citations

  1. APP23

Paper Citations

  1. . Cerebrospinal fluid biomarker signature in Alzheimer's disease neuroimaging initiative subjects. Ann Neurol. 2009 Apr;65(4):403-13. PubMed.
  2. . Age-dependent changes in brain, CSF, and plasma amyloid (beta) protein in the Tg2576 transgenic mouse model of Alzheimer's disease. J Neurosci. 2001 Jan 15;21(2):372-81. PubMed.
  3. . Longitudinal observation on CSF Abeta42 levels in young to middle-aged amyloid precursor protein/presenilin-1 doubly transgenic mice. Neurobiol Dis. 2004 Dec;17(3):516-23. PubMed.
  4. . Brain imaging and fluid biomarker analysis in young adults at genetic risk for autosomal dominant Alzheimer's disease in the presenilin 1 E280A kindred: a case-control study. Lancet Neurol. 2012 Dec;11(12):1048-56. PubMed.
  5. . First effects of rising amyloid-β in transgenic mouse brain: synaptic transmission and gene expression. Brain. 2015 Jul;138(Pt 7):1992-2004. Epub 2015 May 16 PubMed.
  6. . Aβ secretion and plaque formation depend on autophagy. Cell Rep. 2013 Oct 17;5(1):61-9. PubMed.
  7. . Autophagy-related protein 7 deficiency in amyloid β (Aβ) precursor protein transgenic mice decreases Aβ in the multivesicular bodies and induces Aβ accumulation in the Golgi. Am J Pathol. 2015 Feb;185(2):305-13. Epub 2014 Nov 26 PubMed.

Further Reading


  1. . Tracking pathophysiological processes in Alzheimer's disease: an updated hypothetical model of dynamic biomarkers. Lancet Neurol. 2013 Feb;12(2):207-16. PubMed.

Primary Papers

  1. . 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.