Alzheimer’s disease (AD) is a devastating neurodegenerative disorder, and its pathological hallmarks begin to accumulate in the brain decades before clinical symptoms manifest. This preclinical phase presents a critical window of opportunity to understand and potentially intervene in the disease process before irreversible brain damage occurs. Characterizing this early stage is crucial, and research is increasingly focusing on cerebrospinal fluid (CSF) proteins as vital indicators – essentially, biological “transporters” – of these subtle, early changes. AD is not simply driven by amyloid and tau, but involves a complex interplay of factors including immunity, lipid metabolism, vascular health, and endocytic pathways [1]. Understanding these multifaceted aspects in the preclinical stage is key to developing effective early diagnostics and treatments.
The CSF, which bathes the brain and spinal cord, acts as a mirror reflecting the biochemical alterations occurring within the brain. Analyzing the CSF proteome – the complete set of proteins present in the CSF – offers a unique opportunity to define the pathobiological fingerprint of AD in living individuals [2,3,4]. The CSF proteome is not static; protein levels fluctuate as AD progresses through its stages [2, 5]. Therefore, a detailed investigation of the CSF proteome in cognitively unimpaired individuals during the preclinical AD stage can reveal proteins and biological pathways that are particularly relevant to the development and progression of AD. These proteins can serve as crucial biomarkers, improving our ability to predict AD, identify potential therapeutic targets, and act as surrogate endpoints for clinical trials aimed at pre-dementia stages and targeting diverse disease mechanisms [6, 7].
Unveiling CSF Protein “Transporters” in Preclinical AD
In a recent study, researchers analyzed over 900 CSF proteins to understand the biological processes at play in the presymptomatic phase of AD. The study included 297 cognitively unimpaired individuals, categorized as amyloid negative (232) and amyloid positive (65), with a significant 72% undergoing clinical follow-up [8, 9]. Proximity extension assay (PEA) proteome data from a subset of 122 participants from the EMIF-AD preclinical cohort was used to ensure the robustness of findings and validate a biomarker panel [10].
The study design aimed to:
- Define the biological changes characterizing preclinical AD.
- Identify and validate a panel of CSF protein markers to detect individuals in the preclinical stage and predict progression to cognitive impairment.
- Model the levels of these markers in relation to CSF Aβ42 levels (a proxy for pathophysiological progression) to determine if biomarker changes occur before or after significant amyloid accumulation in the brain.
Caption: Study overview and key findings. a illustrates the study design. b Volcano plot of differentially regulated CSF proteins. c UpSet plot showing overlap of proteins dysregulated in preclinical and symptomatic AD. d ROC curves for the 12-CSF protein panel in discovery and validation cohorts. e ROC analysis for prediction of progression to MCI or dementia. f Protein panel levels modeled against CSF Aβ42.
Key CSF Proteins Dysregulated in Preclinical AD
The CSF proteome profiling identified 100 unique proteins that were differentially regulated in amyloid-positive cognitively unimpaired individuals compared to those who were amyloid-negative. This was after removing 43 proteins that showed inconsistent results in the independent EMIF-AD dataset. The top 5 proteins showing differential regulation are involved in critical biological functions:
- Immune Function (ITGB2, CCL11): These proteins highlight the role of the immune system in early AD pathology.
- Protein Glycosylation and Folding (ENTPD5): This suggests disruptions in fundamental cellular processes.
- Insulin Growth Factor Signaling Pathway (IGFBP-1): Implicating metabolic dysregulation in early AD.
- Protein Phosphorylation (ABL1): Indicating altered cellular signaling mechanisms.
Notably, CHIT1, a protein involved in reactive gliosis and known to be elevated in neurodegenerative dementias, showed the strongest correlation with amyloid pathology. Other significant proteins included ITGB2, IGFBP-1, PRCP, and LGMN, the latter two involved in lysosomal proteolytic function. These findings strongly support the concept of AD as a multifactorial disease, even in its earliest stages.
Interestingly, a large majority (79 out of 100) of these preclinical AD-related proteins were also found to be dysregulated in symptomatic AD stages [2,3,4, 11, 12]. This overlap reinforces their relevance to AD pathology. The fact that only ITGB2 and APOL1 were previously identified in dementia stage AD underscores the importance of studying preclinical phases to pinpoint relevant proteins at these very early stages. Functional enrichment analysis revealed that these 100 CSF markers are primarily linked to proteolysis and immune response, pathways known to be involved in AD pathophysiology and the development of amyloid and tau misfolding [1, 2]. This is consistent with findings from targeted CSF proteomic studies in autosomal AD, which showed immune-related protein dysregulation as early as 6 years before disease onset [5].
A 12-Protein CSF Panel for Preclinical AD and Progression Prediction
To create practical biomarker tools, researchers condensed the proteomic data into minimal protein signatures using generalized linear modeling. This resulted in a 12-CSF protein panel capable of detecting individuals in the preclinical phase of AD with a high degree of accuracy (AUC of 0.93). This finding was robustly validated in the EMIF-AD cohort (AUC of 0.89).
The 12-protein panel includes proteins primarily associated with:
- Immune Function (ITGB2, CXCL13, CLEC5A, CCL11, MCFD2, CRTAM, IL7): Further emphasizing the critical role of immune processes in early AD.
- Dopamine Biosynthesis (DDC): Suggesting neurotransmitter system involvement early in the disease.
- Lysosome Activity (GLB1): Highlighting the importance of cellular waste disposal mechanisms.
- Protease Inhibition (CST3/Cystatin-C): Indicating imbalances in protein breakdown and regulation.
- Lipoprotein Metabolism and Lipid Transport (IGFBP-1, PLTP): Reinforcing the role of lipid dysregulation.
Intriguingly, CLEC5A and ITGB2, components of this panel, can regulate TYROBP/DAP12, a key microglia network regulator linked to sporadic late-onset AD [14]. This suggests their potential involvement in the earliest stages of AD. Many of these proteins (ITGB2, IGFBP-1, CLEC5A) have also been specifically associated with AD compared to non-AD dementias [2], and most correlated with CSF Aβ42 or (p)Tau levels, further supporting their association with AD pathophysiology.
This 12-CSF protein panel also demonstrated an 84% accuracy in predicting progression to mild cognitive impairment (MCI) or dementia in a subset of clinically followed cases. This predictive performance increased to 90% when focusing on progression to AD dementia (amyloid-positive progressors). Panel positivity at baseline was linked to a significantly increased risk of clinical dementia (Hazard ratio = 8.37). Furthermore, panel positivity was associated with a steeper cognitive decline over time, as measured by MMSE, highlighting the clinical relevance of this biomarker panel.
Temporal Dynamics of “Transporter” Proteins in AD Pathogenesis
Researchers further modeled biomarker levels along the AD pathology continuum using CSF Aβ42 as an early marker of pathological change. They aimed to determine when these novel biomarkers begin to change in relation to amyloid accumulation. All 12 panel proteins exhibited a single structural break in their trajectory along CSF Aβ42 levels. Notably, all proteins except GLB1 and MCFD2 showed this structural break before CSF Aβ42 positivity, suggesting that changes in these proteins occur very early in the disease process, even before significant amyloidosis is detectable in the CSF. This provides further evidence for the involvement of these proteins in the earliest, etiologically relevant stages of AD. It indicates that processes related to immune function, energy metabolism, neurotrophic support, and endolysosomal function begin to shift before significant amyloid accumulation in the brain, potentially playing a role in the very origins of the disease.
Limitations and Future Directions
While these findings are promising, certain limitations should be acknowledged. The number of cases progressing to MCI or AD dementia was limited, necessitating validation of the prognostic capabilities of the panel in larger, independent cohorts with longer follow-up periods. The cross-sectional nature of the study also limits the understanding of the temporal evolution of these protein changes in individual patients. Future longitudinal studies spanning the entire AD continuum are needed to provide a more detailed picture, accounting for inter-individual variability. Additionally, while PEA technology is robust, further validation using single immunoassays may be beneficial to confirm effect sizes.
Despite these limitations, this study provides compelling evidence that CSF proteome profiling can identify key proteins involved in immune function, proteolysis, and lipoprotein metabolism that are altered even before clinical signs of AD appear. The 12-protein panel developed demonstrates strong potential for identifying preclinical AD and predicting progression to AD dementia. The fact that some of these proteins change before amyloid detection suggests they could be crucial targets for preventing amyloidosis and subsequent clinical symptoms. This panel can be translated into accessible custom assays, particularly focusing on proteins with the strongest effect sizes like ITGB2 and CCL11. Ultimately, this research advances our understanding of the multifactorial nature of AD in its earliest stages, offering new avenues for therapeutic development and biomarker-based clinical tools.
References
[1] Jansen IE, et al. Genome-wide meta-analysis identifies new loci and functional pathways influencing Alzheimer’s disease risk. Nat Genet. 2019;12:43.
[2] del Campo M, et al. CSF proteome profiling across the Alzheimer’s disease spectrum reflects the multifactorial nature of the disease and identifies specific biomarker panels. Nat Aging. 2022;2:1040–53. https://doi.org/10.1038/s43587-022-00300-1 .
[3] Higginbotham, L, et al. Integrated Proteomics Reveals Brain-Based Cerebrospinal Fluid Biomarkers in Asymptomatic and Symptomatic Alzheimer’s Disease. Sci Adv. 2020;6:eaaz9360, 806752. https://doi.org/10.1126/sciadv.aaz9360 .
[4] Bader JM, et al. Proteome profiling in cerebrospinal fluid reveals novel biomarkers of Alzheimer ’ s disease. Mol Syst Biol. 2020;16(6):e9356 1–17. https://doi.org/10.15252/msb.20199356 .
[5] Johnson ECB, et al. Cerebrospinal fluid proteomics define the natural history of autosomal dominant Alzheimer’s disease. Nat Med. 2023;29:1979–88.
[6] Jack CR. Advances in Alzheimer’s disease research over the past two decades. Lancet Neurol. 2022;21:866–9.
[7] van Bokhoven P, et al. The Alzheimer’s disease drug development landscape. Alzheimers Res Ther. 2021;13(1):186. https://doi.org/10.1186/s13195-021-00927-z .
[8] Slot RER, et al. Subjective Cognitive Impairment Cohort (SCIENCe): study design and first results. Alzheimers Res Ther. 2018;10:76.
[9] Verberk IMW, et al. Plasma Amyloid as Prescreener for the Earliest Alzheimer Pathological Changes. Ann Neurol. 2018;84:648–58.
[10] Konijnenberg E, et al. The EMIF-AD PreclinAD study: study design and baseline cohort overview. Alzheimers Res Ther. 2018;10:75.
[11] Yanaika S. Hok-A-Hin et al. Thimet Oligopeptidase as a potential CSF biomarker for Alzheimer’s Disease. A cross-platform validation study. Alzheimers Dement Diagn Assess Dis Monit Press. 2023;15(3):e12456. https://doi.org/10.1002/dad2.12456 .
[12] Hok-A-Hin YS, et al. Neuroinflammatory CSF biomarkers MIF, sTREM1, and sTREM2 show dynamic expression profiles in Alzheimer’s disease. J Neuroinflamm. 2023;20(1):107. https://doi.org/10.1186/s12974-023-02796-9 .
[13] del Campo M, et al. CSF proteome profiling reveals biomarkers to discriminate dementia with Lewy bodies from Alzheimer´s disease. Nat Commun. 2023;14:5635.
[14] Zhang B, et al. Integrated Systems Approach Identifies Genetic Nodes and Networks in Late-Onset Alzheimer’s Disease. Cell. 2013;153:707–20.
