In Preclinical Alzheimer's, p-tau217 in Blood Best Predicts Tangles
The sequential tangling up of the tau protein in different regions of the brain is part and parcel of Alzheimer’s disease progression. Being able to predict if, where, and to what extent tangles will accumulate in a given person is a crucial component of clinical drug trials that aim to derail this process. And the field is getting there. According to a paper published December 20 in JAMA Neurology, the first serial tau PET scans using the 18F-RO948 tracer detected early increases in the medial temporal lobe in people with preclinical AD, followed later by rises in the outer neocortex as clinical symptoms first emerged and then worsened. For people in the preclinical stage of the disease, however, plasma p-tau217 beat PET at predicting neurofibrillary tangles that still lay some time in their future.
- Plasma p-tau217 best predicted subsequent tau-tangle accumulation in people with preclinical AD.
- Tau PET worked better in those with symptoms.
- Screening using plasma and imaging biomarkers could cut size of clinical trials significantly.
For its part, tau PET itself had slightly greater predictive power in people who had already developed symptoms. The two measures together posted the strongest prognostic value, regardless of what disease stage a person was at, at the time of measurement. Ultimately, the researchers found, using plasma and/or imaging biomarkers could halve the needed size of trials aimed at taking down tangles.
As clinical trials increasingly include people in the preclinical or prodromal stages of AD, effective use of biomarkers—both for selecting participants and for gauging their response to therapy—can mean the difference between success or failure. Tau accumulation tracks closely with disease progression, and is an ideal biomarker for these purposes. Some trials are already employing tau PET or plasma tau biomarkers to select participants (Nov 2021 news). But researchers need more data to understand in detail which measure of tau is the most sensitive readout of pathology in the brain at each stage of this long disease, and which is the best at predicting future accumulation of tau.
First author Antoine Leuzy and colleagues addressed these questions in their study, which tracked tau accumulation in 343 participants in BioFINDER 2. This Swedish study is the first to take longitudinal tau PET measurements with the 18F-RO948 tracer, which reportedly has less off-target binding than do first-generation tracers, such as flortaucipir. Tau PET was serially measured in 107 people with Aβ accumulation, of whom 49 were cognitively normal and 58 had MCI; in 173 people without Aβ accumulation, including 137 who were unimpaired and 36 with MCI; and in 63 people with AD dementia. Participants received two tau PET scans, spaced an average of 20 months apart.
The researchers used a computational method called event-based modeling (EBM) to derive five stages of tau progression. For example, stage I comprised the entorhinal cortex, hippocampus, and amygdala, while stage II encompassed the temporal cortical regions (see image below). For people who started the study in the preclinical stage of AD—i.e., cognitively normal people with Aβ accumulation—tau tangles increased primarily in the stage I regions between their two scans. For people with Aβ accumulation and MCI, tangle increases predominated in stage II areas. For those with AD dementia at the beginning of the study, the greatest changes were seen in stage IV, which included frontal regions.
Path of Progression. Medial (top) and lateral views (bottom) show that in stage I, tau accumulates in the medial temporal lobe (red), moving into the temporal cortex by stage II and into outer neocortical regions in subsequent stages. [Courtesy of Leuzy et al., JAMA Neurology, 2021.]
For the most part, the EBM stages aligned roughly with Braak stages of tau-tangle progression, by which tau accumulates early in the medial temporal lobe before spreading out into neocortical regions as symptoms emerge. The scientists therefore observed similar trends in the rate of tau progression when using regions defined by Braak stages instead of EBM.
For Aβ-negative people, regardless of cognitive status, small increases in tau accumulation occurred in the medial temporal lobe, consistent with primary age-related tauopathy as had been reported by others (Nov 2014 news).
With these tau trajectories in hand, the researchers next asked which baseline biomarkers correlated best with subsequent progression of tangle pathology. They limited these analyses to people with Aβ accumulation. In cognitively normal people with amyloid, plasma p-tau217 correlated strongly with subsequent accumulation of tau in stage I regions. Their baseline tau and amyloid-PET scans correlated less well. In contrast, for people with MCI, baseline tau PET SUVR in stage II regions more strongly predicted subsequent tangle accumulation in those regions than did plasma p-tau217, although both were significantly associated with subsequent change. The researchers reached similar conclusions when they used plasma p-tau181 instead of p-tau217, or CSF instead of plasma. Results were similar in a separate cohort of patients in whom flortaucipir served as the tau tracer instead of 18F-RO948.
Overall, the findings support the idea that baseline biomarkers of tau accumulation could help trialists select participants who are likely to accumulate tau over the duration of the trial. By how much would these markers cut the number of participants needed to see a treatment effect on tangle progression? The researchers found that, for preclinical AD, combining plasma p-tau217 and tau PET would reduce this number by 43 percent and, for MCI, by 68 percent. Comparing individual biomarkers for each study population, the authors report that plasma p-tau217 alone would reduce the sample size in a preclinical AD trial by 31 percent, compared to 22 percent if tau PET was used. For an MCI trial, tau PET outperformed plasma p-tau217 at whittling down the numbers—dropping its size by 47 compared to 28 percent.—Jessica Shugart
No Available Further Reading
- Leuzy A, Smith R, Cullen NC, Strandberg O, Vogel JW, Binette AP, Borroni E, Janelidze S, Ohlsson T, Jögi J, Ossenkoppele R, Palmqvist S, Mattsson-Carlgren N, Klein G, Stomrud E, Hansson O. Biomarker-Based Prediction of Longitudinal Tau Positron Emission Tomography in Alzheimer Disease. JAMA Neurol. 2022 Feb 1;79(2):149-158. PubMed.
To make an annotation you must Login or Register.
This paper shows an analysis of longitudinal tau PET with the 18F-RO948 tracer, which may have less off-target binding than flortaucipir. These data show clear progression of tau PET pathology that is largely consistent with neuropathologically established Braak stages. In the normal group (amyloid-negative/cognitively unimpaired), tau PET SUVR only increased in the early regions described in this paper as being in EBM stage 1 (entorhinal cortex, hippocampus, and amygdala), which could represent either primary age-related tauopathy or very early AD. In the preclinical AD group (amyloid-positive/cognitively unimpaired), tau PET SUVR increased more rapidly in these same early regions, but tau PET SUVR also started to increase in areas all over the brain, including cortical regions. By the time individuals developed symptomatic AD (MCI or AD dementia), tau PET signal was increasing robustly throughout the brain. This data is consistent with previous reports that tau PET starts to robustly increase shortly before symptom onset.
Notably, the rates of increasing tau PET SUVR seemed to plateau at about 4 to 5 percent per year, and this plateau is reached in different regions at different stages of disease. In the preclinical AD group (amyloid-positive/cognitively unimpaired), the rate of increase in tau PET SUVR in the early regions (EBM stage 1) seemed to have already plateaued at 4 percent per year and remained at this level in the MCI and AD dementia groups. For the EBM 2 regions, the rate of increase in tau PET SUVR seemed to plateau in MCI. In contrast, the later EBM 4 and 5 regions did not reach this plateau until AD dementia. The rate of change in the later EBM 4 and 5 regions could potentially be used as a measure of disease stage from normal to AD dementia, as it appears to increase throughout the disease course. In contrast, the rate of change in the earlier regions may only be informative in distinguishing normal, preclinical AD and MCI.
The authors selected the preclinical AD and MCI groups, which would be expected to have the largest variance in tau PET SUVR rate of change, to evaluate the association of tau PET SUVR rate of change with other biomarkers, including plasma p-tau217. Plasma p-tau217 was moderately correlated with change in tau PET SUVR. It is important that a single blood test (baseline p-tau217) provided similar information to longitudinal tau PET, which is much more burdensome and expensive.
It was also interesting that baseline tau PET SUVR predicted the rate of change in tau PET SUVR, at least during preclinical AD (R2=0.13) and MCI (R2=0.33). Notably, baseline amyloid PET SUVR is correlated with change in amyloid PET SUVR during the preclinical phase of AD, but this correlation fades around the time of symptom onset. It is possible that tau PET has a similar pattern, but beginning in the transition from preclinical to symptomatic AD. It would be interesting to apply analytical approaches developed with longitudinal amyloid PET (e.g. Villemagne et al., 2013; Jack et al., 2013; Koscik et al., 2020; Schindler et al., 2021) to better understand change in tau PET. It is possible that the tau PET clock starts ticking about the time when the amyloid PET clock stops.
Villemagne VL, Burnham S, Bourgeat P, Brown B, Ellis KA, Salvado O, Szoeke C, Macaulay SL, Martins R, Maruff P, Ames D, Rowe CC, Masters CL, Australian Imaging Biomarkers and Lifestyle (AIBL) Research Group. Amyloid β deposition, neurodegeneration, and cognitive decline in sporadic Alzheimer's disease: a prospective cohort study. Lancet Neurol. 2013 Apr;12(4):357-67. Epub 2013 Mar 8 PubMed.
Jack CR, Wiste HJ, Lesnick TG, Weigand SD, Knopman DS, Vemuri P, Pankratz VS, Senjem ML, Gunter JL, Mielke MM, Lowe VJ, Boeve BF, Petersen RC. Brain β-amyloid load approaches a plateau. Neurology. 2013 Mar 5;80(10):890-6. PubMed.
Koscik RL, Betthauser TJ, Jonaitis EM, Allison SL, Clark LR, Hermann BP, Cody KA, Engle JW, Barnhart TE, Stone CK, Chin NA, Carlsson CM, Asthana S, Christian BT, Johnson SC. Amyloid duration is associated with preclinical cognitive decline and tau PET. Alzheimers Dement (Amst). 2020;12(1):e12007. Epub 2020 Feb 13 PubMed.
Schindler SE, Li Y, Buckles VD, Gordon BA, Benzinger TL, Wang G, Coble D, Klunk WE, Fagan AM, Holtzman DM, Bateman RJ, Morris JC, Xiong C. Predicting Symptom Onset in Sporadic Alzheimer Disease With Amyloid PET. Neurology. 2021 Nov 2;97(18):e1823-e1834. Epub 2021 Sep 9 PubMed.
This is nice work and is generally supportive of what we know about these biomarkers: At the asymptomatic stage, fluid biomarkers of p-tau are the most dynamic and therefore had the strongest association with early regional tau PET change in the temporal lobes. At later stages, baseline tau PET had the strongest association with subsequent tau PET change.
In general, for a screening procedure, these results suggest that starting with plasma pT217 (using the Eli Lilly-developed Meso Scale Discovery platform-based assay) could be an efficient method to determine individuals with a higher probability of having tau PET changes. But also that this is very much dependent on the population being enriched for abnormal Aβ-biomarkers (supplemental table e14).
Ultimately, it will very likely be the presence of tau PET retention outside of the entorhinal cortex plus the presence of Aβ-plaques that will determine the likelihood of tau PET progression, and this study further emphasizes the use of pT217 as a strong marker of Aβ-pathology dependent effects on neuronal dysfunction. A major limitation to this study is the lack of inclusion of cognitive performance as a predictor of longitudinal tau PET in the univariate and multivariate models.
An interesting finding is the lack of clear association of most “neurodegenerative” biomarkers (e.g., plasma NfL, MRI measures of atrophy) with longitudinal tau PET. This is somewhat in contrast to another study from this population (Ossenkoppele et al., 2021). Some of this is likely related to the individual characteristics of these measures such as variance. Even so, this lack of association continues to raise questions as to the interchangeability of biomarkers in the A/T/N classification system, and causes some pause as to what we should expect in clinical trials that have tau PET as a key outcome.
It would be more reassuring to see markers of neurodegeneration correlate with changes of tau PET in order to increase the probability that a drug specifically targeting aggregated tau will result in a clinical benefit.
Ossenkoppele R, Reimand J, Smith R, Leuzy A, Strandberg O, Palmqvist S, Stomrud E, Zetterberg H, Alzheimer's Disease Neuroimaging Initiative, Scheltens P, Dage JL, Bouwman F, Blennow K, Mattsson-Carlgren N, Janelidze S, Hansson O. Tau PET correlates with different Alzheimer's disease-related features compared to CSF and plasma p-tau biomarkers. EMBO Mol Med. 2021 Aug 9;13(8):e14398. Epub 2021 Jul 13 PubMed.
Make a Comment
To make a comment you must login or register.