With the advent of tau PET tracers, researchers are starting to dissect how tau pathology in the brain relates to amyloid plaques and to the development of clinical Alzheimer’s disease. In the last year, the first reports from tau imaging studies have confirmed that these two hallmark pathologies spread through the brain in distinct patterns, with tangles appearing later and correlating more closely with cognitive decline. Now, a paper in the July 13 Journal of Neuroscience sheds more light on what happens at the earliest stages of disease. Researchers led by Jorge Sepulcre and Keith Johnson at Massachusetts General Hospital, Boston, analyzed the relationship between amyloid and tau deposits as well as brain atrophy in cross-sectional data from cognitively normal older adults. Using sophisticated mathematical techniques, they identified core sites where local tau or Aβ deposits correlated with widespread pathology. This hints at a physical, though perhaps long-distance, interaction that might help explain the spread of the disease across the brain, the authors suggest. They have no idea what that interaction might be. In particular, the data bolster a growing consensus in the field that the advance of tangles from the medial temporal lobe into its lateral regions and neocortex signals progression to preclinical AD. “We think the inferior lateral temporal cortex may be the gateway for propagation of the disease,” Sepulcre told Alzforum.
Other researchers praised the innovative analytical approach and said the data add to the emerging picture of how AD develops. “This convincingly shows that Aβ and tau interact in specific areas of the brain,” said David Jones at the Mayo Clinic in Rochester, Minnesota. To Christian Sorg at the Technical University of Munich, the findings are exciting because they suggest that plaques and tangles do influence each other’s spread, even though the mechanism is still unknown. “I think it will become a landmark paper,” he told Alzforum.
Unlike the snapshots of pathology provided by postmortem histology studies, imaging allows researchers to see the whole brain at once and to follow progression over time. Recent cross-sectional tau PET studies confirm previous histology findings of tangles accumulating in the medial temporal lobe of most older adults, but causing few cognitive problems. These tangles push out from the temporal lobe into the cortex only when amyloid deposits are also present in the brain (see Mar 2016 news). That spread of tau pathology seems to be a crucial step in disease progression, because widespread tangles predict cognitive decline better than does amyloid load (see May 2016 news). Tau and amyloid accumulation follow separate anatomical paths and time courses, however, creating the central puzzle of exactly how the two interact. Furthermore, no study had examined how the two PET markers together related to atrophy.
To examine these relationships, Sepulcre and colleagues analyzed data from 88 cognitively healthy participants in the Harvard Aging Brain Study, whose average age was 76. At that age about a quarter of cognitively normal people have the biomarkers of preclinical AD. All participants underwent a PiB amyloid scan, a tau scan with Eli Lilly’s tracer T807, and structural MRI to measure brain volume.
As in previous studies, the authors found that tangles accumulated mainly in the temporal lobe in these volunteers, while Aβ plaques dotted the frontal, parietal, and temporal cortices. In a majority of cases, tau or Aβ deposits in a given region associated with less gray matter there. In addition, distinct hot spots emerged where the presence of pathology correlated with a general loss of gray matter across the brain. For tangles, these were the inferior medial temporal and orbitofrontal cortices, whereas for amyloid, it was the midline frontal, parietal, and orbitofrontal regions, as well as the precuneus and posterior cingulate (see image above). The results suggest pathology in these regions drives neurodegeneration, the authors noted.
How did Aβ and tau pathology relate to each other? Locally, plaques and tangles correlated most in the entorhinal and inferior lateral temporal cortices, but they were also found together in many other cortical regions. To determine which correlations were most important, the authors applied graph theory, a mathematical technique that maps relationships between sets of paired variables in two-dimensional space. They examined the four possible combinations between local PiB or T807 uptake and brain-wide uptake. For each voxel, this analysis identified the association that occurred most frequently across the 88 different people’s brain scans or, in other words, demonstrated the strongest covariance in the cohort.
For most cortical voxels (blue areas in image below), the predominant relationship was local PiB uptake correlating with a high plaque load across the brain. This fits with the idea that amyloid spreads through the cortex. In the entorhinal cortex and other medial temporal lobe structures, generally, association occurred between local tangles (red) and extensive tangles. However, the single strongest relationship that emerged from the graph theoretical analysis was this: Tangles in the inferior lateral temporal cortex (orange) correlate with widespread amyloid.
“Tau there explains a huge amount of amyloid in the rest of the brain. That’s really important, and may provide clues to how disease propagates,” Sepulcre suggested. Others agreed. “Converging evidence [from imaging studies] suggests that tau spreading into the lateral temporal areas triggers disease progression,” said Beau Ances at Washington University in St. Louis.
The last possible relationship—local amyloid correlating with widespread tangles—occurred in only a few voxels in the posterior cingulate and posterior hippocampus (green in the image at right). However, other researchers found this result particularly intriguing, as these regions connect the hippocampal circuit to the cortex and form part of the default mode network, long deemed an early site of Aβ accumulation. “These convergence zones between amyloid and tau are a novel finding, and very interesting,” said Gil Rabinovici at the University of California, San Francisco. Sorg noted that it is still unclear how Aβ in these regions might exert effects on distant tau. Aβ can cause nearby neurons to fall silent or become hyperactive, and this might be one mechanism by which amyloid could influence distant brain regions, he speculated.
This analysis was unable to determine whether tangles drive Aβ deposition, or plaques precipitate tangles. Staging models of AD put amyloid upstream of tau pathology (see Jan 2010 webinar; Dec 2011 news). Some molecular studies support this and suggest mechanisms. For example, a recent paper reported that neuritic plaques convert tau into hyperphosphorylated, pathological forms in mouse models (see Li et al., 2016). However, tangles in some brain areas, such as the entorhinal cortex, precede plaques. The graph theory analysis seems to place tau pathology at the center of disease progression, with temporal tangles uniquely associated with brain-wide amyloid. Longitudinal imaging data may help resolve which pathology kicks off disease in given brain areas.
The findings in this study apply to very early preclinical disease, and may not hold in other cohorts, commenters noted. Rabinovici said that the relationship between protein deposits and brain atrophy, for example, likely varies with disease stage. In this preclinical cohort, both plaques and tangles appeared equally closely linked to brain atrophy, he noted. However, in ongoing studies on patients with more advanced disease, he finds that the amount of tangles, not amyloid, in each region correlates most closely with neurodegeneration. This fits with other work linking tau pathology to synapse and neuron loss. Tangles occasionally appear in regions with no neurodegeneration, but almost all areas with atrophy contain high tangles, Rabinovici added. “Tau seems to spread ahead of neurodegeneration,” he said.
Researchers will also need to measure the effect of pathology on neural networks to sort out the functional relationships between Aβ and tau, Jones suggested. Sepulcre is examining this now with resting-state fMRI connectivity studies. Tangles and amyloid seem to produce distinctive patterns of functional breakdown, he told Alzforum. In preliminary data, tau deposits disrupt the medial temporal lobe system and its connections, while Aβ accumulation spurs hyperactivity and aberrant connections in the default mode network, Sepulcre said. The ultimate aim of these imaging studies is to develop a map of brain changes that will help predict disease progression in individuals, he added. Ances noted, “These studies will lay the groundwork to enable secondary prevention trials to interpret what happens to pathology during treatment.”—Madolyn Bowman Rogers
- Tau PET Aligns Spread of Pathology with Alzheimer’s Staging
- On Multiple Marker Analysis, Tangles Track Best With Functional Decline
- Research Brief: Evidence Supports Model of AD Biomarker Progression
- Li T, Braunstein KE, Zhang J, Lau A, Sibener L, Deeble C, Wong PC. The neuritic plaque facilitates pathological conversion of tau in an Alzheimer's disease mouse model. Nat Commun. 2016 Jul 4;7:12082. PubMed.
- Neural Circuitry Goes Haywire in Both Sporadic and Familial AD
- Do Earliest Cognitive Deficits in Alzheimer's Appear in the Entorhinal Cortex?
- Cognitive Deficits Arise From Network, Not Regional, Dysfunction
- Brain Connectivity Reveals Preclinical Alzheimer’s Disease
- Communication Breakdown: Multiple Networks Decline in AD Brains
- Functional Imaging Gives Early Glimpse of AD
- Traveling Tau—A New Paradigm for Tau- and Other Proteinopathies?
- Aβ the Bad Apple? Seeding and Propagating Amyloidosis
- Insidious Spread of Aβ: More Support for Synaptic Transmission
- Cortical Hubs Found Capped With Amyloid
- Do Overactive Brain Networks Broadcast Alzheimer’s Pathology?
- Sepulcre J, Schultz AP, Sabuncu M, Gomez-Isla T, Chhatwal J, Becker A, Sperling R, Johnson KA. In Vivo Tau, Amyloid, and Gray Matter Profiles in the Aging Brain. J Neurosci. 2016 Jul 13;36(28):7364-74. PubMed.