One burning question on the minds of people who study preclinical detection in different populations and with different methods is how the data might all fit together. Johnson offered this hypothesis:
By that scenario, the long road to disease would begin with amyloid-β pathology. Age and genetics (i.e., ApoE genotype) influence a person’s propensity to deposit amyloid, and both CSF Aβ and PIB PET can detect it. This would lead to synaptic dysfunction, which impairs cognition directly, and also indirectly over time as the neurons bearing malfunctioning synapses slowly die. It is widely accepted that brain and cognitive reserve influence how long a person can withstand the damaging effects of the encroaching pathology. Bill Klunk, of the University of Pittsburgh Medical School in Pennsylvania, noted that vascular disease is another important environmental factor that acts at this level. The PIB+ clinically normal people currently draw intense interest among researchers. Are they the ones who are on this path to AD? Imaging methods from functional MRI to FDG glucose PET and volumetric MRI exist to visualize and quantify these stages. The standard clinical tests used routinely in diagnosis pick up the latest stage on this flow chart, but increasingly, new findings are converging around more sensitive cognitive measures to pick up earlier, subtler changes (see Part 3 and below).
One way of probing whether this suggested signature of AD holds true is to test whether amyloid deposition in clinically and cognitively normal people relates to the other components in this diagram. “We do this by looking at other measures of AD in PIB+ controls,” Johnson said. Here are four examples of doing so that Johnson presented in his talk:
- Cortical thickness: In MGH studies of cognitively normal PIB+ people, increasing amyloid is associated with a thinning cortex in the precuneus, anterior and posterior cingulate, and some temporo-parietal and medial frontal cortical areas. These areas have come to represent a network that represents functional connection. “These areas talk to each other,” Johnson said.
- Hippocampal volume: In the MGH/BWH studies, this measure does not seem to relate to amyloid deposition in the cognitively normal volunteers but does by mildly impaired stages of CDR 0.5 and 1. “This suggests to us that the cortex thins first and the hippocampus a bit later,” Johnson said.
- Brain connectivity: A recent study reported that the more highly linked an area is—in other words, if it is a hub of intense connectivity—the more amyloid it tends to have (Buckner et al., 2009). This month, the researchers reported an amyloid-related disruption of intrinsic connectivity among asymptomatic elderly (Hedden et al., 2009). “This suggests a disconnection where one area is unable to talk with other, previously connected areas anymore,” Johnson said. In July, the same researchers reported that using fMRI, they found that brain amyloid in non-demented people mapped to their inability to deactivate the default network when focusing on a task (Sperling et al., 2009).
- Cognitive function: It is becoming increasingly clear that amyloid deposition is tied to cognitive impairment. IQ and education are important here because people high on those measures are able to tolerate amyloid pathology longer. People with low IQ/education scores show a much stronger link between amyloid in their precuneus (a particularly early area) and how they perform on neuropsychological tests (Rentz et al., Annals of Neurology, in press). And preliminary analyses of longitudinal data showed a detectable decline in a delayed word recall test. People declined along with increasing precuneus PIB. These are subtle effects on difficult tests, Johnson noted; the first difference the MGH/BWH studies see in those tests is that people with PIB in precuneus lose the practice effect, i.e., they don't improve anymore with repeated testing. “We really don’t understand in biological terms what cognitive reserve is and how it preserves function,” Johnson noted.
The Society for Neuroscience conference, held 17-21 October 2009 in Chicago, featured even newer data along these lines. To quote but one example, Deepti Putcha, working with Reisa Sperling at Brigham and Women’s Hospital in Boston, reported that cognitively normal older people who performed poorly on a challenging memory test (RVALT) had hyperactivation of their hippocampus along with cortical thinning in AD-vulnerable areas of the medial temporal lobe. These data are fresh out of the scanner, and concomitant amyloid and connectivity imaging data on this new aging cohort still needs to be analyzed, Putcha said.
These and other emerging clues on how multiple preclinical markers link up together in the same people will hopefully bring into focus a predictive signature over the course of the next five years, as longitudinal studies under way at several institutions progress. But even now, the field has advanced far enough to start planning prevention research, argued Eric Reiman of the Banner Alzheimer's Institute in Phoenix, Arizona. Evaluating promising presymptomatic treatments in the general population takes too long and is too costly, but doing so in high-risk groups such as ApoE4 carriers and presymptomatic carriers of dominantly inherited Alzheimer disease (DIAD)/eFAD mutations has come entirely within reach, Reiman said. Adding to the other speakers’ presentations a brief summary of his group’s research, Reiman noted that he has followed ApoE- genotyped volunteers—that is, people at three defined levels of risk—with brain imaging for well over a decade. People with the ApoE4 allele have decreased glucose utilization and subtle metabolic changes already as young adults, predating even the earliest biochemical marker to date, i.e., CSF Aβ. These FDG PET abnormalities do not progress continuously from early adulthood, but they flag early on where the AD changes will develop later, Reiman said (Reiman et al., 2001; Caselli et al., 2004; Reiman et al., 2005; Reiman et al., 2009).
The wealth of imaging and biomarker data the field has accumulated to date provides a sufficient foundation for proof-of-concept trials of biomarker outcomes, not clinical outcomes, with several hundred subjects, Reiman said. He spoke forcefully for starting an era of prevention research. This will not be a quick achievement, but it is time to begin gathering drug/biomarker data that can open a new regulatory path to approval. Such trials would also show research volunteers who have allowed scientists to poke, question, and scan them for decades that the scientists, in turn, will use all that data toward something tangible that truly matters to the volunteers and their families.—Gabrielle Strobel.
This is Part 6 of a seven-part series on presymptomatic detection. See also Parts 1, 2, 3, 4, 5, and 7.