06 Sep 2024

More than any other biofluid, the cerebrospinal fluid is in intimate contact with the brain. What can the thousands of proteins churning within this brain bath tell us about how Alzheimer’s develops, and how to counteract this disease? A study published June 26 in Science Translational Medicine adds to the growing pile of proteomics data implicating cellular processes such as autophagy, ubiquitination, endocytosis, and glycolysis as central players in the disease process. By combining two proteomic techniques, researchers led by Erik Johnson of Emory University in Atlanta took stock of more than 4,000 proteins in the CSF, tying some 34 modules of co-expressed protein culprits to disease.

They also zeroed in on which pathways ApoE4 influences, implicating oxidative stress, mitochondrial function, and neddylation—a protein modification similar to ubiquitination. In serum from cognitively normal people, these protein networks predicted future AD. Finally, the scientists found that in a clinical trial, the norepinephrine reuptake inhibitor atomoxetine squelched some of these AD-associated pathways, particularly those involved in glucose metabolism.

“This paper represents yet another proteomics tour de force from the Emory group,” wrote Russell Swerdlow of Kansas University Medical Center. “As someone with an intense interest in the role of mitochondria in AD, I am consistently excited to see these proteomics studies time and time again end up emphasizing prominent changes in their mitochondria and glycolysis modules.” Swerdlow noted that the findings also underscore a tight interconnectedness across functional domains—such as metabolism, inflammation, and proteostasis—in brain health and disease.

The data complement a plasma proteomics study co-authored by Johnson, which analyzed thousands of plasma samples from an Icelandic cohort to dissect the influences of incipient AD and ApoE genotype on the proteome. For more on that paper, see previous story.

Johnson and his Emory colleagues have been cranking out proteomics studies for years, each expanding in size, scope, and technical prowess (for example, Feb 2022 news; Aug 2023 news on Johnson et al., 2023; Sep 2023 news). For this latest exploration, they not only went hunting for proteins tied to AD, but also wanted to parse what ApoE genotype, and an experimental treatment, does to the CSF proteome.

First author Eric Dammer and colleagues analyzed the CSF proteomes of 300 people, 140 controls and 160 with AD as determined by their CSF ratio of total tau to Aβ42. They used two complementary techniques, each capturing different aspects of the proteome, Johnson told Alzforum. One is tandem mass tag-mass spectrometry (TMT-MS), the other SomaScan. The latter uses aptamers—single-stranded DNA molecules that bind specific proteins—to capture and quantify thousands of proteins in a sample. Integrating the spoils of both methods enabled measurement of more than 4,500 proteins in each sample.

How did the scientists make sense of this data? First, they grouped the proteins into 34 co-expression modules representing distinct biological functions. Many of these modules tracked with at least one AD-related biomarker and/or with cognitive impairment (image below). For example, a group of glycolysis and redox homeostasis proteins had the strongest ties to both CSF Aβ42 and cognition. Modules containing autophagy, endocytosis, ubiquitination, and synaptic function proteins associated with CSF total tau.

Wheel of Misfortune. Grouped into 34 co-expression modules (outer ring), thousands of proteins in CSF were correlated with AD endophenotypes (labeled in cross-section of circle, top-center). Modules that tracked with AD attributes are bolded in red; with ApoE4, in black. [Courtesy of Dammer et al., Science Translational Medicine, 2024.]

Zeroing in on three groups of proteins that correlated with the number of ApoE4 alleles a person carried, Dammer found an oxidant detoxification/MAPK signaling module, a mitochondria module and, curiously, proteins involved in neddylation. This post-translational modification adds the ubiquitin-like protein NEDD8 to a protein, relegating it to destruction. The oxidant/MAPK and neddylation proteins were reduced among both ApoE4 carriers and people with AD; the mitochondria module tracked higher.

Would these signatures show up in the blood? To find out, the researchers hunted for the CSF protein modules within serum samples from more than 5,000 participants of the AGES cohort. This longitudinal Icelandic study tracks emergent AD diagnoses among adults who were cognitively normal at baseline. Among the more than 4,000 proteins measured by SomaScan in AGES, most CSF modules overlapped with at least one serum module, though only half of the serum modules overlapped with those identified in CSF. Notably, the groups of proteins that had associated with ApoE4 in the CSF also did for similar serum modules. Specifically, oxidant detoxification/MAPK signaling and mitochondrial function proteins were tied to ApoE4 and to a future AD diagnosis.

The same was true in the Atherosclerosis Risk in Communities cohort. In ARIC, plasma levels of all three ApoE4-linked modules were associated with development of dementia up to 21 years later.

Might these groups of proteins respond to a potentially disease-modifying treatment? The researchers measured the CSF proteomes of participants in a small clinical trial at Emory of atomoxetine, a norepinephrine reuptake inhibitor approved for ADHD. Norepinephrine tone flags in the locus coeruleus early in AD, and scientists propose that restoring levels of this neurotransmitter could curb AD pathology development in the region. In the trial’s cross-over design, 39 participants with MCI due to AD received either atomoxetine or placebo for six months, then switched for another six months. Published results from this biomarker-focused study indicate that while the drug did not improve cognition, it boosted CSF norepinephrine, reduced CSF total tau and p-tau181, and stoked glucose metabolism in the brain as gauged by FDG-PET (Levey et al., 2021).

Comparing participants’ CSF proteomes before and after treatment surfaced four protein modules that shifted in response to ATX: glycolysis/redox homeostasis, extracellular matrix/vasculature, translation, plus there was an ambiguous module. Regarding the first two, the scientists reported that the 65- protein glycolysis/redox module, was elevated in AD and reduced in response to atomoxetine. The extracellular matrix module was also ramped up in AD relative to controls, but it increased further upon treatment. It’s unclear what these changes mean, Johnson told Alzforum. If having more glycolysis proteins is detrimental, then lowering them with atomoxetine could help. If instead the boost represents a compensatory response to AD, then lowering this module could hurt.

The researchers also flipped their analyses, using unbiased proteomes to group people. Proteome characteristics split members of the original CSF cohort into 10 groups. Perhaps surprisingly, members of only one—group 10—were uniform in their biomarker levels of Aβ and tau. All other groups included people who were positive and others who were negative for AD biomarkers.

Take group 6. It had a mix of controls and people with AD. Controls had proteomic features that strongly resembled an AD-associated proteome, except they had higher CSF Aβ, lower p-tau181 and total tau, and a dampened glycolysis module. This could reflect a resilient phenotype, or an early disease stage in which other changes precede changes to AD biomarkers, the authors wrote. “[The findings] illustrate the heterogeneity of CSF proteomic changes that do not necessarily correlate with Aβ and tau,” they wrote.

A person’s proteomic profile also failed to track with the extent of their subsequent response to atomoxetine—as gauged by how much the drug would lower expression of the glycolysis module.

To Johnson, the findings underscore the multifaceted complexity of AD. “We’re defining the disease based on Aβ and tau, but so many other processes are contributing to the cognitive decline that people are dealing with,” he said. “We are trying to peel back all those layers to understand the process.”—Jessica Shugart

Comments

No Available Comments

Make a Comment

To make a comment you must login or register.

References

News Citations

1. Among Icelanders, Blood Proteome Predicts Dementia, Parses Role of ApoE4 6 Sep 2024
2. Proteomics Highlight Alzheimer’s Changes in Matrisome, MAPK Signaling 3 Feb 2022
3. Proteins in Biofluids Foreshadow Dementia by 30 Years 29 Aug 2023
4. CSF Proteomic Panel Better Predicts Decline Than Do Classic AD Biomarkers 14 Sep 2023

Therapeutics Citations

1. Atomoxetine

Paper Citations

1. Johnson EC, Bian S, Haque RU, Carter EK, Watson CM, Gordon BA, Ping L, Duong DM, Epstein MP, McDade E, Barthélemy NR, Karch CM, Xiong C, Cruchaga C, Perrin RJ, Wingo AP, Wingo TS, Chhatwal JP, Day GS, Noble JM, Berman SB, Martins R, Graff-Radford NR, Schofield PR, Ikeuchi T, Mori H, Levin J, Farlow M, Lah JJ, Haass C, Jucker M, Morris JC, Benzinger TL, Roberts BR, Bateman RJ, Fagan AM, Seyfried NT, Levey AI, Dominantly Inherited Alzheimer Network. Cerebrospinal fluid proteomics define the natural history of autosomal dominant Alzheimer's disease. Nat Med. 2023 Aug;29(8):1979-1988. Epub 2023 Aug 7 PubMed.
2. Levey AI, Qiu D, Zhao L, Hu WT, Duong DM, Higginbotham L, Dammer EB, Seyfried NT, Wingo TS, Hales CM, Gámez Tansey M, Goldstein DS, Abrol A, Calhoun VD, Goldstein FC, Hajjar I, Fagan AM, Galasko D, Edland SD, Hanfelt J, Lah JJ, Weinshenker D. A phase II study repurposing atomoxetine for neuroprotection in mild cognitive impairment. Brain. 2021 Dec 17; PubMed.

Further Reading

News

1. NULISA—A New Proteomic Method to Revamp Biomarker Analysis 23 Aug 2024