But what about other neurodegenerative diseases? Do they also spread through distributed networks? Could that explain the distinct patterns of pathology in different diseases? The answer appears to be yes, based on new data from William Seeley and colleagues at the University of California, San Francisco, and Michael Greicius at Stanford University. In a paper in the April 16 Neuron, the researchers report how they used MRI to compare the patterns of atrophy and network activity in AD and four other diseases. In each case, the patterns delineated a different functional network. The results paint a picture of the selective vulnerability of different networks to different pathological insults, and suggest that neurodegeneration may propagate via network connections.

To compare different diseases, the researchers analyzed a total of 102 patients with one of five distinct dementia syndromes, including AD, three types of frontotemporal dementia (behavioral variant frontotemporal dementia, semantic dementia, and progressive nonfluent aphasia) and corticobasal syndrome. To better age-match healthy controls, and to focus on the beginnings of disease, the scientists chose young patients who had early-onset disease, and excluded patients with moderate or severe dementia.

Using first structural MRI, the researchers found that the patterns of atrophy for each condition were distinct, as previously described, and corresponded to the different cognitive deficits noted in each patient group. To probe for possible network connections, Seeley and colleagues chose the brain region that showed the greatest atrophy for each condition, and looked for synchronized activity with that seed region during a task-free fMRI scan in healthy controls. They found that each of the five disease-associated regions anchored a different functionally connected network. Moreover, in each case the network activity mapped closely to the distribution of atrophy observed in the analysis of diseased brains.

The functional linkage of brain regions was mirrored in structural measures, as well. Analyses of gray matter intensity across the whole brain in both normal and disease subjects revealed that the functionally linked regions also showed a high correlation of gray matter volume. As Seeley explained to ARF, “This suggests that regions that fire together, scale together. They grow and shrink together, although we don’t know exactly how that works.” Adding a structural correlation to the functional measures may be a useful approach to define networks in general, Seeley says. This is similar to previous work that used correlations of cortical thickness to map the default network (He et al., 2008 and ARF related news story).

The results support the “network degeneration hypothesis,” which holds that disease starts in small neuron populations and progressively spreads to connected areas. “The point is that once a disease gets going at a given location in the brain, the whole network or regions that are interconnected with that region are at risk, and the disease is likely to propel itself down that set of pathways somehow or the other,” Seeley explained. Because the study looks at different clinical syndromes that can be caused by a number of pathogenic proteins, this ability to propagate through networks is likely to be relevant to numerous disease proteins, including β amyloid, tau, α-synuclein, and TDP43. How the proteins might spread their damage through networks is not clear, but there are several possibilities. Pathogenic proteins like Aβ are known to disrupt synaptic connections directly, and this could weaken networks. Some disease-causing proteins disrupt axonal transport, possibly leading to a severing of connections between regions. Some might even propagate misfolding by direct transfer from one neuron to another (e.g., see Frost et al., 2009 and ARF related news story). These potential scenarios are not mutually exclusive, the authors point out. And the extensive overlap between neurodegenerative diseases, for example, AD and dementia with Lewy bodies (DLB), both clinically and pathologically, is likely to complicate this new line of research as it digs deeper.

The network map for AD agrees with previous work showing involvement of the default-mode network, but the other diseases had their own distinct patterns. That suggests that the network signatures might eventually find use clinically to diagnose early stage or even presymptomatic disease, when there are still neurons to save. “The nice thing about network analyses is that they don’t rely on atrophy,” Seeley said. “We could potentially see the disease signature before brain regions start to shrink and wither.” The measure might also serve as an objective biological endpoint to monitor disease progression and treatment. Researchers including Greicius and Randy Buckner are already working on developing these methods for AD, and the new data should spur studies of similar approaches for other diseases.

An outstanding question remains as to why different diseases, and presumably different pathogenic proteins, have a favorite network to attack. “Each disease, defined at the protein level, has a set of neurons somewhere in the brain that are most vulnerable to that protein’s misfolding. That’s the concept of selective vulnerability, and understanding what brings these problems to these cells is critical to understanding neurodegenerative disease,” Seeley said.

To that end, Seeley says he is glad to see the concepts of network biology becoming more interwoven into the fabric of neurodegeneration research. “So much of the basic research on neurodegeneration focuses on single cells and molecules, but the tide is turning on that a little bit. We’re seeing a new wave of network-oriented basic and translational neuroscience to try to understand what it is about network physiology, or gene expression, or development that makes these systems disease vulnerable. I’m excited to see that pick up.” The importance of pushing forward with explorations of network structure and function in the human brain was echoed in an accompanying commentary by Marsel Mesulam of Northwestern University in Chicago, Illinois, who pioneered anatomical study of networks important for behavior and cognition.—Pat McCaffrey.

References:
Seeley WW, Crawford RK, Zhou J, Miller BL, Greicius MD. Neurodegenerative Diseases Target Large-Scale Human Brain Networks. Neuron. 2009 April 16; 62:42-52. Abstract

Mesulam M. Defining Neurocognitive Networks in the BOLD New World of Computed Connectivity. Neuron. 2009 April 16; 62:1-3. Abstract