18 April 2007. Five years ago this spring, the molecular amyloid imaging probe PIB for the first time entered the veins and brains of people at the first independent site outside of the developer’s home turf at the University of Pittsburgh. The volunteers were patients at Agneta Nordberg’s outpatient clinic at the Karolinska University Hospital Huddinge in Stockholm, and they submitted to the first European test of this substance at the Uppsala PET center/Imanet in Uppsala, Sweden. At the 8th International Conference AD/PD, held last month in Salzburg, Austria, Nordberg recalled that right from the get-go in 2002, it was clear to the Swedish investigators that there was a stark difference in PIB uptake between people with and without AD. Even so, back then discussing amyloid imaging was a challenge and met much skepticism, said Nordberg, who is also at the Karolinska Institute. That has changed thoroughly since then, as some 25 groups around the world have begun their own independent studies to verify claims made for PIB, and to explore amyloid imaging more broadly and deeply. At this point, Nordberg said, the field can conclude that amyloid imaging works. It is also clear that amyloid PET imaging is not a great burden. Procedures vary slightly among the different groups, but typically a person receives a PIB injection into an arm vein while lying in the scanner, and is then scanned for 60 to 90 minutes. It is, however, still expensive. Several companies have jumped in to develop various ligands, and the scientific debate has moved on to subtler questions. This story summarizes some of the 15 presentations on amyloid imaging at the AD/PD conference. The talks and posters offered news morsels on the following issues:
Biology of PIB binding
Beyond PIB: other ligands
Imaging inflammation and amyloid
First, the emerging consensus. Worldwide, more than 500 patients have been scanned with PIB to date, 200 of them in Australia alone. Several presenters noted that basic observations at the different centers in large part converge. For example, PIB reliably binds to β amyloid in the nanomolar range, and it visualizes amyloid deposits in a characteristic pattern of brain areas in different people with AD (see, e.g., Ng et al., 2007). Importantly, PIB imaging in the hands of several different groups indicates that a person’s brain amyloid load has built up more or less fully before the person is diagnosed with AD. The longest-standing studies are now beginning to report repeat scans of people who have progressed in their AD. They are finding that while the patients over time decline clinically, and also decline in terms of brain atrophy as measured by MRI and glucose utilization as measured by positron emission tomography (FDG PET), their PIB amyloid load stays unchanged (e.g. see Engler et al., 2006). This was surprising, as many colleagues had expected a linear PIB increase with ongoing disease. “The amyloid just seems to sit there stably as the patient gets worse,” said David Brooks of Imperial College, London. It could be that there is a steady state with active amyloid turnover that PIB imaging cannot detect, but the patient’s total value of PIB imaging tends to stay the same even years into the disease. This means that PIB is not going to be a marker for disease progression, and indeed several presenters noted that PIB load does not correlate with cognitive symptoms in clinically diagnosed AD. At the same time, this also suggests that PIB or other amyloid imaging ligands could be a useful marker for the effect of amyloid-removing therapies. This is a helpful refinement of what PIB can and cannot do, as a candidate biomarker typically cannot serve all of the potential purposes scientists initially envision for it—that is, marker for prediction, diagnosis, progression, treatment effect (see ARF related news story).
The realization that a person’s amyloid load has built up fully by the time of diagnosis comes only partly from repeat scans of AD patients who are being followed over time. It comes also from scans of people with MCI. In Salzburg, several groups reported data similar to what the Pittsburgh researchers have seen. Brooks’s Imanet center in London, Nordberg’s Swedish group, as well as Christopher Rowe’s group at Austin Hospital near Melbourne, Australia, and a fourth group led by Riitta Parkkola at the University of Turku’s Pet Center in Finland, all showed that about 60 percent of their MCI cases scanned with PIB so far (24 in London, 21 in Stockholm, 40 in Melbourne, 13 in Turku) have a PIB reading like that of AD cases. “My guess is these are the people who will get AD,” said Brooks, who also works for G.E. Healthcare, which develops PIB commercially. Groups around the world are following MCI cohorts to answer this question. Rowe noted that during the MCI stage, the PIB load correlates strongly with impaired cognitive function, and that this correlation only disappears once dementia is established as AD. In the Australian healthy aging group, people with high PIB readouts had slightly lower episodic memory performance and were frequently judged to be "declining," whereas people with low PIB readouts were judged to be “stable.” Rowe cautioned that this is an early observation based on too few patients to be statistically significant. Nordberg reported that in her group of 21 longitudinally followed MCI patients, 11 had high PIB and seven of them have since converted to AD. Of the people who had low PIB, none have converted (Forsberg et al., in press at Neurobiology of Aging). All three speakers noted that they expect amyloid imaging to help them with the challenge of predicting who will go on to develop AD from among the heterogeneous group of people who are classified as having MCI.
At the research level, the effort to predict future AD reaches even further back into groups of cognitively normal volunteers. The Australian group reported in Salzburg that about 20-25 percent of healthy people around age 70-75 begin to show significant amyloid loads in PIB PET scans. This matches postmortem findings. “It is tantalizing to speculate that these are the people who will have AD by age 85,” said Rowe. Other investigators, too, are asking themselves whether they are looking, literally, at the prodromal stage of AD in these people. Nordberg said that on this issue, the field would do well to explore more deeply the emerging correlations between PIB data and cerebrospinal fluid biochemistry of Aβ and tau proteins. She noted that Swedish data linking PIB and CSF appear to strengthen the prediction along the same lines as reported recently by researchers at Washington University, St. Louis, Missouri (Fagan et al., 2007). In addition, the significant percentage of high PIB levels in cognitively normal people implies that clinical studies must take care not to recruit people with subjective memory complaints or elevated AD risk factors into “healthy control” groups to avoid blurring the necessary distinctions between the trial groups, Nordberg added.
Budding Drug Marker?
Can amyloid imaging report on the success or failure of anti-amyloid drugs? This research is in its infancy, but Nordberg reported a first glimpse in Salzburg. A 1-year placebo-controlled Swedish trial of phenserine in 20 patients with mild AD incorporated PIB PET scans along with FDG PET, CSF and plasma biochemistry, and cognitive readouts. Phenserine is an experimental acetylcholinesterase inhibitor reported to modulate Aβ production. Between baseline and the 6-month time point, the scientists noted an improvement in MMSE and in an attention test in the patients on study drug. Some of the other markers also changed in these people: glucose metabolism increased, PIB retention nudged downward by about 12 percent (test-retest variation of PIB is below 5 percent), and CSF Aβ nudged upward. CSF tau stayed unchanged. “This is the first time we have related PIB measurements and CSF to a drug effect. The people on phenserine did slightly better and their biomarkers are moving, though we don’t understand it fully yet,” said Nordberg. For news on differential diagnosis, PIB biology, other amyloid imaging ligands, and imaging inflammation, see Part 2 of this story.—Gabrielle Strobel.
- Ng S, Villemagne VL, Berlangieri S, Lee ST, Cherk M, Gong SJ, Ackermann U, Saunder T, Tochon-Danguy H, Jones G, Smith C, O'Keefe G, Masters CL, Rowe CC. Visual assessment versus quantitative assessment of 11C-PIB PET and 18F-FDG PET for detection of Alzheimer's disease. J Nucl Med. 2007 Apr;48(4):547-52. PubMed.
- Engler H, Forsberg A, Almkvist O, Blomquist G, Larsson E, Savitcheva I, Wall A, Ringheim A, Långström B, Nordberg A. Two-year follow-up of amyloid deposition in patients with Alzheimer's disease. Brain. 2006 Nov;129(Pt 11):2856-66. PubMed.
- Fagan AM, Roe CM, Xiong C, Mintun MA, Morris JC, Holtzman DM. Cerebrospinal fluid tau/beta-amyloid(42) ratio as a prediction of cognitive decline in nondemented older adults. Arch Neurol. 2007 Mar;64(3):343-9. PubMed.
- Translational Biomarkers in Alzheimer Disease Research, Part 1
- Salzburg Overview: 8th International Conference Draws Record Number
- Salzburg: New Proteins Redefine Frontotemporal Dementias
- Salzburg: Getting to Know Progranulin, the New Heavy
- Salzburg: The Pony in There—Can ACAT Inhibition Work for AD?
- Salzburg: Protein Trio Offers Handle on Elusive ERAD