J Alzheimers Dis. 2017;55(4):1639-1657. doi: 10.3233/JAD-160090.

The Brain’s Structural Connectome Mediates the Relationship between Regional Neuroimaging Biomarkers in Alzheimer’s Disease.

Pandya S, Kuceyeski A, Raj A; Alzheimer’s Disease Neuroimaging Initiative.

 

Abstract

Alzheimer’s disease (AD), one of the most common causes of dementia in adults, is a progressive neurodegenerative disorder exhibiting well-defined neuropathological hallmarks. It is known that disease pathology involves misfolded amyloid-β (Aβ) and tau proteins, and exhibits a relatively stereotyped progression over decades. The relationship between AD neuropathological hallmarks (Aβ, hypometabolism, and tau proteins) and imaging biomarkers (MRI, AV-45/FDG-PET) is not fully understood. In addition, biomarker pathologies are oftentimes discordant, wherein it may show varying levels of abnormality across brain regions. Evidence based on recent elucidation of trans-neuronal “prion-like” transmission and other available data already suggests that disease spread follows the brain’s fiber connectivity network. Thereby, the brain’s connectome information can be used to predict the process of disease spread in AD. A recently established mathematical model of AD pathology spread using a connectome-based network diffusion model was successful in encapsulating neurodegenerative progression. Motivated by these network-based findings, the current study explores whether and how network connectivity mediates the interactions between various AD biomarkers. We hypothesized that the structural connectivity matrix will mediate the cross-sectional association between regional AD-associated hypometabolism and Aβ deposition. Given recent reports of inherent or lifetime activity of brain regions as strong predictors of Aβ deposition in patients, we also tested whether healthy metabolism exerts a network-mediated effect on Aβ deposition and hypometabolism in AD patients. We found that regional Aβ deposition is best predicted by a linear combination of both regional healthy local metabolism and connectome-mediated regional healthy metabolism.

KEYWORDS: AV-45-PET; Alzheimer’s disease; FDG-PET; amyloid-β; biomarkers; cross-sectional; hypometabolism; metabolism; structural connectivity

PMID: 27911289

 

Supplementary information

Structural and neuropathological changes in the brain mainly associated with gray matter atrophy, Aβ deposition, and glucose hypometabolism are some of the major characteristic hallmarks of AD. We know that disease pathology involves misfolded Aβ and tau proteins, and exhibits a relatively stereotyped progression over decades [1]. Neuroimaging techniques such as MRI and PET with combination of different radiotracers are extensively used to capture these changes. However, the progression of misfolded Aβ and tau proteins through brain circuits, their misfolding, aggregattion and spread, has remained a mystery until recent elucidation of trans-neuronal “prion-like” transmission [2]. For year’s relationship between Aβ depositions and glucose metabolism are poorly understood.

Triggering events in the amyloid progression postulates that initial Aβ accumulations are assumed to cause the latter cascading breakouts of structural and functional changes in the brain [3]. Per Braak’s staging model, misfolded proteins trigger misfolding of adjacent proteins in the brain cells along certain neuronal pathways in the brain [4]. This data naturally suggest that disease spread follows the brain’s fiber connectivity network. Motivated by this we had previously established a mathematical model of AD pathology spread using a connectome-based network diffusion model which was successful in encapsulating neurodegenerative progression [5].

Based on the foundation of our earlier model, in our recent publication we explored whether and how network connectivity intermediates the interactions between various AD biomarkers. In our study we have shown that this complex relationship between regional AD-associated hypometabolism and Aβ deposition can be understood through the brain’s structural connectome. We hypothesized a natural model of progression as demonstrated in Figure 1, whereby initial brain activity results in local Aβ deposition due to dominance of axonal connections within local areas. As disease progresses towards MCI stage, we would expect Aβ deposition in remote regions with the strongest remote deposition in mature AD stage. Keeping this in mind we developed a model with stage-dependent regional association between imaging biomarkers of Aβ deposition and metabolism in healthy controls, and hypometabolism in AD patients using structural connectivity in governing these relationships.

 

 

Figure 1:  Proposed natural model of progression through stage-dependent connectome mediation. From left to right as reported in J Alzheimers Dis. 2017;55(4):1639-1657, in early stages Aβ deposition occurs in proportion to neural activity (metabolism), then begins to be deposited in remotely connected regions in the MCI stage (Effect A), and achieves the strongest remote deposition in mature AD stage (Effect B).

 

Our results show that Aβ deposition in a given region is significantly related to the healthy metabolism of both locally and remotely connected regions. Our model demonstrates that remote metabolism through long-range fiber connectivity is a better predictor of regional Aβ deposition than combined contribution of local and remote metabolism. Figure 2 illustrates regional Aβ distribution (light colored blobs) overlaid by regional hypometabolism (solid blobs) in AD patients based on the predictive ability of our model. We can see that high levels of Aβ deposition in the frontal and parietal cortex are complemented by a commensurate decrease in glucose metabolism, suggestive of our model prediction that the local effect of Aβ on metabolism might be mediated by region-specific toxic effects of Aβ.

 

This model will help unravel processes involved in understanding this network-mediated relationship which is of critical importance in understanding spatiotermporal dynamics of AD related biomarkers. Quantitative assessment of these connectivity-mediated relationships between biomarkers in AD and AD related pathologies like tau hold importance for future clinical applications.

 

 

Figure 2: Glass brain representations of Aβ (light colored blobs) on the predictive ability of the model overlaid by hypometabolism (solid blobs) in AD. Each point represents one of the GM regions, color-coded by lobe membership (frontal = purple, parietal = red, occipital = orange, temporal = cyan, and subcortical = green).

 

References

[1] Goedert M, Klug A, Crowther RA (2006) Tau protein, the paired helical filament and Alzheimer’s disease. J Alzheimers Dis 9, 195-207

[2] Frost B, Diamond MI (2010) Prion-like mechanisms in neurodegenerative diseases. Nat Rev Neurosci 11, 155-159.

[3] Palop JJ, Mucke L (2010) Amyloid-beta-induced neuronal dysfunction in Alzheimer’s disease: From synapses toward neural networks. Nat Neurosci 13, 812-818.

[4] Braak H, Braak E (1996) Evolution of the neuropathology of Alzheimer’s disease. Acta Neurol Scand Suppl 165, 3-12.

[5] Raj A, Kuceyeski A, Weiner M (2012) A network diffusion model of disease progression in dementia. Neuron 73, 12041215