Amyloid Imaging: Laying PIB Concerns to Rest
Consider it the final nail. That’s what coauthor Bill Klunk calls a new study in which he and colleagues show direct correlation between postmortem amyloid pathology and live brain imaging using Pittsburgh Compound B (PIB), a PET tracer being developed for earlier diagnosis and drug testing of Alzheimer disease. Using PIB to track amyloid in the brain of a 76-year-old man with dementia, researchers at Massachusetts General Hospital and others (including Klunk) had already reported last year the first correspondence between PIB-PET imaging and amyloid distribution revealed at autopsy in the same person (see ARF related news story). However, there was a catch: that patient had a clinical diagnosis not of AD but of dementia with Lewy bodies (DLB), a disorder that shares characteristics with both Alzheimer and Parkinson diseases. In the current study, published March 12 in the journal Brain online, researchers at the University of Pittsburgh, Pennsylvania, confirm the PIB-postmortem connection in a patient with definitive AD.
“It’s sort of an anti-climactic thing,” admitted Klunk, who co-developed PIB with chemist Chester Mathis. Since human studies with the radiotracer began six years ago, PIB has been used on more than 2,000 subjects at 40 research centers worldwide. “But this was the last piece,” Klunk told ARF. “The final nail, if you will.”
To build a more solid case for PIB as a tool for monitoring amyloid changes during disease progression, first author Milos Ikonomovic and colleagues used PIB-PET to peer into the brain of a 64-year-old woman with severe AD. After her death 10 months later, the researchers removed her brain for biochemical analyses on frozen sections from the right hemisphere and, from the left hemisphere, histological studies. The team also collected brain tissue postmortem from 27 other patients who did not undergo PIB-PET imaging while alive but were diagnosed with AD at autopsy.
In histological studies of brain samples from the autopsy cases, the scientists found that 6-CN-PIB, a highly fluorescent PIB analog, specifically labeled amyloid-β (Aβ) plaques across multiple brain regions in patterns very similar to those of anti-Aβ antibodies. To assess PIB’s binding specificity more quantitatively, the team performed ELISA and observed a direct correlation between [3H]PIB binding and levels of insoluble Aβ peptide in postmortem frontal and occipital cortices. Importantly, the researchers found no significant correlation between [3H]PIB binding and soluble Aβ, which does not form plaques.
From the AD patient who underwent PIB-PET imaging 10 months before death, the scientists were able to compare in vivo PIB uptake with region-matched postmortem measurements of amyloid pathology in the same brain. To make these comparisons, the team probed some 25 regions of interest (ROIs)—1x1-cm tissue cubes prepared from multiple brain regions displaying distinctive patterns of Aβ plaques and neurofibrillary tangles, the emblematic markers of AD. Experiments to quantitate PIB binding, Aβ plaque load, and Aβ peptide levels in these ROIs were then compared with PIB uptake during PIB-PET imaging. An MRI scan taken just prior to PIB-PET imaging allowed the researchers to define specific brain areas that correspond to the postmortem ROIs.
“The in vivo PIB signal correlates very well with the amyloid load in the brain determined at autopsy,” Klunk said of the typical AD case study. Based on their findings, he and his colleagues propose that PIB binding is highly specific for insoluble, fibrillar Aβ but not for neurofibrillary pathology. This conclusion was bolstered by the experiments using autopsy samples from the other 27 AD patients who did not undergo PIB-PET scanning while alive. PIB “behaves the same way in all these AD cases,” Klunk said, “so the inferences we’re making from this one patient seem to be representative.”
Christopher Rowe, director of the Centre for PET at Austin Health in Melbourne, Australia, agreed. "This paper is the icing on the cake for the validation of what is proving to be an extraordinarily good PET tracer," he wrote in an e-mail to ARF (see further comments below).
En route to its much-hoped-for entry into the clinical realm as an AD diagnostic marker, PIB is being used as a readout for drug effect in human trials. A team led by Agneta Nordberg at the Karolinska Institute in Stockholm, Sweden, has just published the first such study—using PIB to track brain amyloid load in 20 patients with mild AD during the course of treatment with phenserine. Based on their findings, which appeared in the Annals of Neurology last month (Kadir et al., 2008), PIB can be useful in evaluating other anti-amyloid drug therapies, Nordberg told ARF. However, one factor in the current trial that might hamper sound assessment of PIB’s effectiveness is that phenserine is an acetylcholinesterase inhibitor whose Aβ-reducing activity is only a secondary effect (see comment below). Also ongoing are small, unpublished trials in Europe using PIB to measure amyloid load in AD patients receiving immunotherapy with AAB-001 (bapineuzumab), a humanized monoclonal antibody that binds to and clears Aβ peptide (see ARF clinical trials update).
While Rowe describes PIB’s march toward clinical use as “tantalizingly close,” Klunk concedes that the real value of PIB will depend on the eventual availability of much more effective AD treatments than are available now. “If you don’t have a drug to treat the disease, there’s really no point screening for it,” Klunk said, noting that a typical PET scan costs $1,500 to $2,000 dollars. “But if you have a drug…that’s a drop in the bucket.”—Esther Landhuis
Esther Landhuis is a science journalist in Dublin, California.
This paper is the icing on the cake for the validation of what is proving to be an extraordinarily good PET tracer. PIB has shown robust results all around the world in a multitude of centers no matter what analysis method has been applied, ranging from simple visual inspection of a 20-minute scan to complex modeling of 90-minute acquisitions with arterial metabolite-corrected blood sampling. PIB has challenged preconceptions about the distribution of β amyloid plaque and won the arguments. For example, there were many disbelievers that the high frontal lobe binding was due to high concentration of amyloid plaque in this area. This paper confirms that this is indeed the case. The paper also plugs a few gaps by throwing some light on the potential mechanism for the observed marked binding of PIB in the striatum but not the cerebellar cortex when both areas have few neuritic plaques. The current work has demonstrated that PIB does not bind to diffuse plaque in the cerebellum, confirming that the use of the cerebellar cortex to normalize cortical PIB binding measures is valid. The extraordinary specificity of PIB for β amyloid plaque has been repeatedly demonstrated in vitro and now further confirmed by the correlation of histopathology with in-vivo imaging.
We can now say with great confidence that we have an excellent tool for in-vivo measurement of insoluble β amyloid deposition, thanks to the inspired and determined efforts of the University of Pittsburgh researchers over the last 10 or more years. The great remaining challenge is to determine the clinical significance of β amyloid deposition in apparently normal elderly individuals and those with mild cognitive impairment. We should have the answer to this vital question within the next year or two as longitudinal follow-up data accumulate from individual institutions and large multicenter studies such as the Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Australian Imaging, Biomarkers and Lifestyle study of the elderly (AIBL).
It is clear that PIB PET can greatly assist differential diagnosis of dementia and earlier diagnosis of Alzheimer disease. But perhaps most importantly, the results in the paper and the myriad of recent PIB publications show that the much-sought tool for preclinical detection of Alzheimer disease that will greatly assist the development of preventative and early intervention therapies is tantalizingly close, if not already here.View all comments by Christopher Rowe
This is the first time an amyloid imaging ligand such as PIB has been used in the evaluation of a drug treatment study in mild AD patients. This is a double-blind study. It shows that in phenserine-treated AD patients, there is in some patients a reduction in PIB retention after three months of up to 15 percent that is more than the test-retest value (less than 5 percent). Interestingly enough, there is an increase in CSF Aβ40 in the phenserine-treated AD patients, and there is a negative correlation between CSF Aβ40 and PIB retention. This indicates that there seems to be a reciprocal change in amyloid in brain versus CSF in the individual patients.
There is also a correlation between CSF Aβ40 and cognition, as well as with cerebral glucose metabolism. This is promising in that PIB can be useful in the further evaluation of anti-amyloid drug therapy in AD, including immunization therapy.View all comments by Agneta Nordberg
It has been six years now since the first Alzheimer patient was scanned with PIB in Sweden (Klunk et al., 2004). It was quite obvious already from the first PIB scans that there was a robust difference in PIB retention in cortical brain regions of mild AD patients compared to age-matched controls (Klunk et al., 2004). Ikonomovic et al. now report in-vitro PIB binding in the brain of an AD patient who 10 months earlier underwent a PIB scan. A correlation was observed between PIB retention measured in vivo and PIB in-vitro binding. Diffuse plaques in caudate nucleus and presubiculum were labeled, and amorphous Aβ plaques were not detectable in vitro with PIB. Importantly, no binding was observed to neurofibrillary tangles. A direct correlation was observed between in-vitro PIB binding and levels of insoluble Aβ. This study supports the assumption from earlier in-vivo studies that PIB is useful in measuring amyloid plaque load in AD. Further studies are now needed to understand the role of other pathological processes in AD brains. If we can develop imaging ligands to label neurofibrillary tangles, inflammatory processes, (micro)glia, and pre-plaque forms of Aβ, e.g., oligomers, then we will learn and understand much more about the AD pathology that will help us develop new AD drugs and evaluate the underlying mechanisms.
Klunk WE, Engler H, Nordberg A, Wang Y, Blomqvist G, Holt DP, Bergström M, Savitcheva I, Huang GF, Estrada S, Ausén B, Debnath ML, Barletta J, Price JC, Sandell J, Lopresti BJ, Wall A, Koivisto P, Antoni G, Mathis CA, Långström B. Imaging brain amyloid in Alzheimer's disease with Pittsburgh Compound-B. Ann Neurol. 2004 Mar;55(3):306-19. PubMed.
Ikonomovic MD, Klunk WE, Abrahamson EE, Mathis CA, Price JC, Tsopelas ND, Lopresti BJ, Ziolko S, Bi W, Paljug WR, Debnath ML, Hope CE, Isanski BA, Hamilton RL, Dekosky ST. Post-mortem correlates of in vivo PiB-PET amyloid imaging in a typical case of Alzheimer's disease. Brain. 2008 Jun;131(Pt 6):1630-45. PubMed.View all comments by Agneta Nordberg
New diagnostics go hand in hand with new therapies. Are we there yet! Roll on the summer months!