Study detects biomarker changes nearly 30 years before symptom onset of Alzheimer’s disease

By Dr. Liji Thomas, MDReviewed by Sophia CoveneyAug 10 2023

A new paper, recently published in Nature Medicine, has shown how Alzheimer’s disease (AD) pathology comes into being over a period of almost three decades.

AD is among the top causes of morbidity and loss of independence among the elderly. The disease process begins long before cognitive symptoms appear. Drawing upon changes in the cerebrospinal fluid (CSF) that bathes the brain and spinal cord,

 Introduction

As the global population becomes older, millions of people are at risk for dementia, a catch-all term used to denote cognitive incompetence and loss of independence caused by neurodegeneration. At present, its pathological hallmark is the high levels of beta-amyloid (Aβ) peptide and neurofibrillary tangles (NFTs) of tau protein in the brain.

These are measured by various methods, including positron emission tomography (PET) imaging and the assessment of molecular proteins in the CSF or blood. However, it is necessary to understand the changes that occur over time in the demented brain, which are not easily traceable using such techniques.

Autopsy studies have shown how brain tissue changes its attributes, at all levels, but biomarkers remain essential to track the course of such alterations over the lifespan, especially while the individual remains young and healthy, and in those who never develop symptoms of dementia at all.

For this reason, the current study focuses on those people who have autosomal dominant AD mutations (“ADAD”) affecting one of three genes – the amyloid precursor protein (APP), presenilin 1 (PSEN1), or presenilin 2 (PSEN2) gene.

All of these boost the production of the Aβ42 peptide at all stages and trigger the early formation of Aβ plaques. These mutations produce symptoms in almost every individual affected and can predict the age of dementia onset in each case, based on the family history and the type of mutation present.

Such data, obtained from the Dominantly Inherited Alzheimer Network (DIAN) international observational study, can help understand how AD biomarkers evolve over time. The researchers used a panel of 59 proteins along with other protein coexpression modules already developed, to help track such changes in ADAD, in combination with other biomarkers, imaging techniques, and cognitive tests.  

What does the study show?

The researchers found that changes in secreted and excreted proteins in the CSF reflect the earliest steps in AD pathology.

Using a specialized targeted quantitative mass spectrometry (MS) method called selected reaction monitoring mass spectrometry (SRM-MS), they found significant differences in the levels of 33/59 proteins between those who carried the ADAD mutation, and noncarriers, with most changes predating the onset of symptoms.

Among the most important was an early rise in the level of SMOC1 (SPARC-related modular calcium-binding protein 1). SMOC1 is a protein found in association with Aβ plaques and is among the brain cortex proteins that show the greatest rise in asymptomatic AD. Similarly, a decreased Aβ42/40 ratio is associated with Aβ plaque development.

Among the five categories of changed biomarkers, the M42 matrisome module is a set of proteins found in connection with the extracellular matrix. It includes amyloid precursor protein (APP), which closely reflects total Aβ levels, along with many other proteins that are associated with Aβ plaques.

These proteins, including Aβ, associate with each other via their common ability to bind heparin, which facilitates plaque formation. The Christchurch APOE mutation knocks down this heparin-binding capacity and is associated with markedly lower susceptibility to ADAD. So is the APOE ε2 allele that protects against late-onset AD (LOAD) and binds heparin at reduced rates.

M42 matrisome protein expression is driven mainly by SMOC1, and the latter was found to be raised as long as ~30 years before symptom onset and to rise progressively over time.

This rise was observed to precede a drop in Aβ42 and the Aβ42/40 ratio in ADAD carriers and to occur before phosphorylated tau pTau181 and pTau217 levels rise. Both of these are markers of early Aβ deposition.

Another protein in the matrisome, SPON1 (spondin1), also rose early but did not persist at an elevated level throughout the course of ADAD. These proteins play varied roles in brain metabolism, such as activating the MAPK pathway that is linked to impaired cognition, and SPON1, required for neurite development.

Successive changes were then observable in other categories of biomarkers, beginning 26 years before the onset of symptoms and involving the 14-3-3 family of proteins that are expressed at neuronal synapses.

Elevations in the 14-3-3 proteins accelerated 8 years prior to symptom onset, at which point neurofilament light chain (NEFL) levels were also found to rise. NEFL is a neurodegeneration marker.

There were also short-lived rises in associated proteins that are involved in glycolysis. Alongside, other possibly protective proteins also rose in concentration. Simultaneously, individuals with ADAD mutations showed a period of better cognitive function than noncarriers, indicating that the latter might be a neuroprotective response.

At 19 years prior to symptom onset, total and pTau205 levels rose, along with other proteins associated with markers of microglial activation like c-sTREM2, that reflect inflammation.

About 10 years later, this was associated with a rise in NEFL, indicating the slow progression of changes in brain axons and white matter metabolism. This is the third category of changing markers.

The fourth category involves changes beginning at 6 years before symptom onset, with inflammatory changes heralded by rises in proteins like osteopontin (SPP1), chitinase-3-like protein 1 (CHI3L1), as well as a more marked rise in c-sTREM2. This is the point at which brain metabolism begins to decline and cognitive functions are first observably impaired.

The final category includes reduced levels of neuronal and neurosecretory proteins, indicating that brain atrophy has set in, with the loss of neurons and synapses. Glycolysis markers rise again during this period. This glycolytic response could indicate an astroglia-driven reduction in neuronal metabolic support, or alternatively a glial response to the death of neurons.

Both SMOC1 and a composite panel of biomarkers from the 33 elevated proteins were found to correctly and strongly distinguish ADAD mutation carriers from noncarriers. They performed better than either amyloid or tau biomarkers, especially during very early disease.

What are the implications?

SMOC1 may be a useful biomarker for amyloid deposition in the brain at very early stages of AD and to assess the response to anti-Aβ immunotherapies. The M42 matrisome may also contain both potential AD biomarkers as well as new targets for the treatment of AD.

While ADAD and LOAD share a common pathophysiology, other differences could be present that will alter the biomarker alteration sequence described here. For instance, vascular disease and Lewy bodies are more common in LOAD, whereas Aβ plaque and NFT levels are higher in ADAD. The former is characterized by Aβ42 overproduction but the latter by reduced clearance.

Further studies will be required to elucidate such differences over a lifetime. However, the study points to three intervention targets, namely, the initiation of Aβ plaque formation 30 years before symptoms appear, the beginning of axonal and white matter breakdown 19 years before symptom onset, and the marked inflammation that immediately precedes cognitive impairment and brain atrophy, at six years prior to symptoms.

Targeting pathological changes for therapeutic intervention will likely be most successful before, at or near the onset of such changes. Once an individual develops symptoms, a multitarget therapeutic approach will likely be required.

Johnson et al. (2023)