Brain Scans and Biomarkers

William Jagust

Helen Wills Neuroscience Institute and School of Public Health University of California, Berkeley and Molecular Biophysics and Integrated Bioimaging Lawrence Berkeley National Laboratory 1980s onward 1970s onward Functional Imaging Structural Imaging (CTÞMRI) Metabolism (FDG-PET) Atrophy Hypometabolism

1990s onward 2000s onward Functional Imaging Molecular Imaging fMRI, resting BOLD (Amyloid/tau PET) Activation Protein aggregates and network changes

Structural, functional, and molecular imaging are all crucial components of research evaluations of dementia patients today 1984: Publication of widely accepted clinical diagnostic criteria for Alzheimer’s disease

2018: Publication of a novel research framework – Alzheimer’s disease as a biological disorder

34 years to move from a clinical diagnosis to a biological model of AD Alzheimer’s Disease: Neuropathology

b-amyloid (Ab) Plaque pathology

Brain atrophy is a sign of neurodegeneration Tau Neurofibrillary Tangles Positron Emission Tomography Central Radiopharmacy for F18

Injection into Cyclotron for Fast participant production of short- Radiochemical lived radionuclides synthesis of PET (C11, F18) tracers

PET Scan

Image reconstruction and data analysis Fluorodeoxyglucose (FDG) – PET Reduced glucose metabolism in Alzheimer’s Disease likely reflects synaptic dysfunction

Normal

Alzheimer’s Disease FDG-PET in Alzheimers and FTLD

FTLD

AD

F = frontal cortex, TP = Temporoparietal cortex, PC = posterior cingulate/precuneus Glucose Metabolism Declines in AD

Characteristic brain regions affected in Alzheimer’s disease:

Medial parietal lobe, lateral Cortical Thinning in AD temporal/parietal cortex, medial temporal lobe

Similar regional vulnerability in structure and function

Wirth et al J Neuroscience 2013 In vivo Amyloid Imaging with Pittsburgh Compound B (PIB)

H3C S CH3 N N + CH3 Fibrillar Ab CH3

Histology - Thioflavin T

Amyloid Plaques

PET Imaging - [11C]6-OH-BTA-1 (PIB) HO S 11 NH CH3 N

Chet Mathis and Bill Klunk, University of Pittsburgh Control AD Control AD

18F-Flutemetamol 18F-Florbetaben

AD AD

Control Control

18F-Florbetapir 18F-NAV4698 FDG (glucose metabolism) vs PIB (b-amyloid) Neurodegeneration vs Molecular Pathology

FDG PIB

Normal

Alzheimer’s Disease Molecular Biomarkers vs Neurodegeneration Biomarkers

Neurodegeneration Molecular Indicative of brain damage Indicative of pathology Non-specific (as to cause) Specific May be complex to interpret Relatively straightforward Correlation with symptoms May or may not correlate Questionable utility for with symptoms therapeutic testing Should be useful for therapeutic testing Multiple tau radiopharmaceuticals are now available for PET imaging

Villemagne et al, Nat Rev Neurol 2018 Tau Imaging with Flortaucipir

Young Adults PIB- Older Adults PIB+ Older Adults Alzheimer’s Disease N=5 N=17 N=16 N=15

Key regional paUern of tracer reten&on Braak staging (t-tests) Supra-threshold voxels AD-vulnerable regions (t-tests)

Schöll, Lockhart et al, Neuron 2016

Topography of tau deposition in the AD continuum reflects Braak Pathological staging

} Key regions of tau accumula&on: } medial temporal lobe (MTL) and inferior/middle temporal regions, retrosplenial cortex/posterior cingulate, precuneus, inferior parietal Utility of Tau PET in Differential Diagnosis

719 participants from Lund, Seoul, SF

Tau PET 90% sensitivity & specificity discriminating AD from other disorders

Ossenkoppele et al JAMA 2018 Biomarker Patterns: Similarities and Differences Biomarker Measurement in Autosomal Dominant AD Supports the Amyloid Hypothesis

Dominantly Inherited Alzheimer’s Disease The Dominantly Inherited Cross-sectional data on biomarkers from Alzheimer’s Disease (DIAN) Study tected in mutation carriers 15 years before expect- autosomal dominant, symptomatic and ed symptom onset (Table 2). As expected, there 2 Aβ deposition asymptomatic family members was an age-dependent decrease in hippocampal volumes in noncarriers (Fig. 1D).22 CSF tau

1 CSF Aβ42 CEREBRAL GLUCOSE METABOLISM CDR–SOB Because age-at-onset is preserved across FDG-PET measures of cerebral glucose use in the precuneus were compared with the use of an a 0 generations, biomarker values in relation to priori hypothesis of decreased metabolism in age-at-onset can be calculated mutation carriers to determine regional metabolic Hippocampal −1 volume defects. The precuneus region, which is known Standardized Difference to be an area of early deposition in both sporadic Results show that the earliest biomarker Alz heimer’s disease and autosomal dominant Alz- Glucose −2 metabolism heimer’s disease,4,23,24 was chosen for analysis of b −30 −20 −10 0 10 change is elevation of A in brain, about amyloid deposition. A significant decrease in cere- Estimated Yr from Expected Symptom Onset 20 years before expected onset bral metabolism in the precuneus was detected in mutation carriers 10 years before expected symp- Figure 2. Comparison of Clinical, Cognitive, BatemanStructural, etMetabolic, al, NEJM and 2012 tom onset (Fig. 1E and Table 2). Biochemical Changes as a Function of Estimated Years from Expected Symptom Onset. Aβ DEPOSITION TheThe normalized Amyloid differences betweenHypothesis mutation carriers and: noncarriersAmyloid are deposition is the initiating event in PIB-PET measures of fibrillar Aβ deposition25 in shown versus estimated years from expected symptom onset and plotted with a fitted curve. The order of differences suggests decreasing Aβ in the precuneus were compared with the use of an 42 AD,the CSF leading (CSF Aβ42), followed to by NFTsfibrillar Aβ deposition,-tau, then brain increased taudegeneration, and dementia a priori hypothesis of increased regional amounts in the CSF (CSF tau), followed by hippocampal atrophy and hypometabo- of amyloid deposition in mutation carriers. There lism, with cognitive and clinical changes (as measured by the Clinical De- was no detectable amyloid deposition in noncar- mentia Rating–Sum of Boxes [CDR-SOB]) occurring later. Mild dementia riers. All noncarriers had PIB-PET SUVR values of (CDR 1) occurred an average of 3.3 years before expected symptom onset. 95% confidence interval bands are shown in Figure S2 in the Supplementary less than 0.88. As compared with noncarriers, Appendix. mutation carriers had significant amyloid deposi- tion in the precuneus 15 years before expected symptom onset (Fig. 1F and Table 2). The amount of amyloid deposition in mutation carriers in- COMBINED MODEL creased as a function of estimated years from ex- The order and rate of pathophysiological changes pected symptom onset at least until clinical symp- in autosomal dominant Alz hei mer’s disease were tom onset. estimated with the use of an analysis of the rela- tionship among clinical, cognitive, imaging, and BIOCHEMICAL MEASURES biochemical measures in the DIAN cohort (Fig. 2). In mutation carriers, levels of tau in the CSF were Beginning 25 years before expected symptom on- increased 15 years before expected symptom onset set, Aβ42 concentrations in the CSF in mutation (Fig. 1G and Table 2). Concentrations of Aβ42 in carriers appeared to decline, as compared with A video showing the CSF decreased as a function of estimated years those in noncarriers. Aβ deposition as measured Aβ deposition over from expected symptom onset and were pseudo- by PIB-PET (Fig. 3; and see Video 1, available at time is available at NEJM.org normal at approximately 20 years before expected NEJM.org) was detected at least 15 years before symptom onset, reaching low levels 10 years be- expected symptom onset (Table 2). Increases in fore expected symptom onset (Fig. 1H). The de- levels of tau in the CSF and in brain atrophy were crease by half in Aβ42 in the CSF and the increase detected approximately 15 years before expected in tau in the CSF were similar in magnitude to symptom onset, followed by cerebral hypometab- those typically observed in late-onset sporadic Alz- olism and impaired episodic memory approximate- 26 heimer’s disease. Plasma Aβ42 levels were elevat- ly 10 years before expected symptom onset and ed in mutation carriers, as compared with non- global cognitive impairment starting at 5 years carriers (Fig. 1I). before expected symptom onset.

n engl j med 367;9 nejm.org august 30, 2012 801 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on September 30, 2012. For personal use only. No other uses without permission. Copyright © 2012 Massachusetts Medical Society. All rights reserved. NIA-AA Research Framework

540 Alzheimer’s diseaseC.R. Jack defined Jr. et al. / Alzheimer’s& by Dementia 3 pathological 14 (2018) 535-562 processes: b-amyloid

Table 2 deposition (A), tau deposition (T), and neurodegeneration (N) Biomarker profiles and categories

AT(N) profiles Biomarker category

A-T-(N)- Normal AD biomarkers Alzheimer’s Continuum: A+T-(N)- Alzheimer’s pathologic change Amyloid Positivity (A+) A+T+(N)- Alzheimer’s disease

Alzheimer’s A+T+(N)+ Alzheimer’s disease continuum Alzheimer’s Disease: A+T-(N)+ Alzheimer’s and concomitant suspected Amyloid and tau non Alzheimer’s pathologic change positivity (A+T+)

A-T+(N)- Non-AD pathologic change

A-T-(N)+ Non-AD pathologic change

A-T+(N)+ Non-AD pathologic change This is currently a framework meant only for research, not clinical care Abbreviation: AD, Alzheimer’s disease. Jack et al Alz & Dementia 2018 NOTE. See text for explanation of (N) notation. NOTE. Binarizing the three AT(N) biomarker types leads to eight different biomarker “profiles”. Every individual can be placed into one of Fig. 2. Preclinical Alzheimer’s pathologic change. A cognitively unim- the three general biomarker “categories” based on biomarker profiles: those paired 67-year-old man. Participant in the Mayo Clinic Study of Aging. with normal AD biomarkers (no color), those with non-AD pathologic Abnormal amyloid PET (Pittsburgh compound B, top row), no uptake on change (dark grey), and those who are in the Alzheimer’s continuum (light tau PET (with flortaucipir, middle row), no atrophy on MRI (bottom row). grey). The term “Alzheimer’s continuum” is an umbrella term that denotes Biomarker profile A1T2(N)2. either Alzheimer’s pathologic change or AD. NOTE. If an individual has an abnormal amyloid biomarker study, but a biomarker for tau is not available, then the individual is placed into the “Alz- cific for AD, are used only to stage severity not to define the heimer’s continuum”. A missing biomarker group can be labeled with an presence of the Alzheimer’s continuum. asterisk (*). For example, A1(N)1 without a T biomarker would be A1T*(N)1. 8. Biomarker profiles and categories

symptomatic phases, and for typical and atypical clinical In many research studies, it will be most appropriate to presentations. treat biomarkers of amyloid, pathologic tau, and neurodegen- eration/neuronal injury as continuous measures without using normal/abnormal cut points. However, biomarkers used in 7. Staging medicine often use a cut point denoting normal versus abnormal values to support management decisions for an in- We next developed a system for staging severity. Our dividual patient. The need for discrete categorization of guiding principles were the following. Two types of informa- biomarker continua is also obvious for AD clinical trials, tion about the research participant are staged independently where explicit cut points serve as inclusion/exclusion criteria. from each other: (1) grading disease severity using bio- The addition of a normal/abnormal cut point for each markers and (2) grading the severity of cognitive impairment. AT(N) biomarker group results in eight different AT(N) Measures used to define AD must be specific for the disease, “biomarker profiles”(Table 2, Text Box 1): whereas measures used to stage severity need not be. Thus, A1T2(N)2,A1T1(N)1,etc.Basedonthedefinitions different measures have different roles (Text Box 2). Ab bio- of Alzheimer’s pathologic change and AD outlined markers determine whether or not an individual is in the Alz- earlier, the ATN biomarker system assigns every individ- heimer’s continuum. Pathologic tau biomarkers determine if ual to one of three “biomarker categories”(Table 2, Text someone who is in the Alzheimer’s continuum has AD Box 1): (1) individuals with normal AD biomarkers; (2) because both Ab and tau are required for a neuropathologic those in the Alzheimer’s continuum (subdivided into Alz- diagnosis of the disease. Neurodegenerative/neuronal injury heimer’s pathologic change and AD); and (3) those with a biomarkers and cognitive symptoms, neither of which is spe- normal amyloid biomarker but with abnormal T or (N), or Amyloid Status Predicts Clinical Conversion

Normal à MCI conversion MCI à AD conversion

Roberts et al. JAMA Neurol 2018 AD Neuropathology in Normal Cognitive Aging

Neurofibrillary Tangles (NFTs) Tau Protein

b-amyloid plaques (Ab) MRI: Neurodegeneration

PET: Amyloid

Tau-NFTs in aging accumulate in the medial temporal lobe Ab plaques appear diffusely (atrophy on MRI, tau with throughout the brain PET: Tau PET) (visible with PET) How Well Does Ab Predict Cognition in Normal Aging?

Cross-sectional meta-analysis

4% difference (age 72)

No Ab effect for: FCSRT, 21% difference MMSE, Trails, Frontal (age 90) Assessment Battery

N = 2908 N = 313 N = 445 Jansen et al JAMA Psychaitry, 2018 Dubois et al Lancet Neurology 2018 Donohue et al JAMA 2017 Weak!!

righ t H

b-amyloid Tau Neurodegeneration righ t H

b-amyloid Tau Neurodegeneration 0 Amyloid1 (FBP/FBB)0 1 is Associated0 with1 Tau (FTP) 2.5 2.5 2.5 0 1 0 1 N = 6460 1 2.5 2.0 2.5 2.0 2.5 2.0

Entorhinal cortex 1 Neocortical tau Inferolateral1 temporal tau 1 2.0 1.5 2.0 Braak 11.5 2.0 Braak1.5 34 Braak 56 1 1 1 Unimpaired1.5 Impaired N Unimpaired1.5 N Impaired Unimpaired N Impaired 1.5 1.0 0 1.0 1 0 1.0 1 0 1 SMC SMC SMC 2.5 EMCI 3.0 N EMCI 2.5 N EMCI 1.0 1.0 N LMCI 1.0 2.5 SMC 2.5 SMC 2.5LMCI SMC LMCI AD AD AD 2.0 EMCI 2.5 EMCI EMCI Male FTP Braak 56 Braak FTP

FTP Braak 56 Braak FTP 56 Braak FTP 2.0 2.5 LMCI 2.5LMCI LMCI 2.5 2.0 1 1 1 2.0 2.0 AD 2.0 AD

AD 2 2 2

1.5 56 Braak FTP 56 Braak FTP FTP Braak 56 Braak FTP 1.5 2.0 1.5 2.0 1.5 1.5 2.0 1.5 2 2 1.0 2 N N N 1.0 1.0 1.5 1.0 1.5 1.0 SMC 1.5 1.0 SMC SMC 2.5 EMCI EMCI EMCI 0.8 1.2 1.6 2.0 0.8 1.2 1.6 2.0 3.0LMCI 2.5LMCI LMCI 1.0 1.0 0.8 1.2 1.6 2.0 0.8AD 1.21.0 1.6 2.0 0.8 1.2 1.6 2.0 0.8AD 1.2 1.6 2.0 AD

FTP Braak 1 Braak FTP Florbetapir cortical SUVR Florbetapir cortical SUVR Florbetapir cortical SUVR Female 2.0 34 Braak FTP 2.5 56 Braak FTP 0.8 1.2 1.6 2.0 0.8 1.2 1.6 2.0 0.8 1.2 1.6 2.0 0.8 1.2 1.6 2.0 0.8 1.2 1.6 2.0 0.8 1.2 1.62.0 2.0 Florbetapir cortical SUVR Florbetapir cortical SUVR 2 2.0 Florbetapir cortical SUVR 2 2 1.5 1.5 1.5 1.0 1.0 1.0 0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250 0 50 100 150 200 250 Amyloid PET Centiloids Amyloid PET Centiloids Amyloid PET Centiloids Ab within the Negative Range Affects FTP Control Subjects

0 0 0

1.0 1.0 1.0 0.5 Entorhinal cortex 0.5 0.5 Neocortical tau Inferolateral temporal tau 0 0 0 0.0 Braak 1 0.0 Braak 34 0.0 Braak 56 -0.5 -0.5 -0.5 1.75 1.75 1.75 N N N -1.0 -1.0 Age (+) (trend) -1.0 Education (+) Sex (F>M) SMC SMC SMC Edu (+) b (p=0.027) Education (+) A UW_MEM 1.0 UW_MEM 1.0 UW_MEM 1.0 Ab (p=0.046) Ab (p=0.010) 0.5 1.50 0.5 1.50 0.5 1.50 1 1 1 0.0 0.0 0.0 -0.5 -0.5 -0.5 N N N 1.25 SMC1.25 SMC1.25 SMC -1.0 -1.0 -1.0 1234 1234 1234 FTP Braak 1 Braak FTP FTP Braak 56 PVC 34 Braak FTP FTP Braak 56 PVC 56 Braak FTP FTP Braak 56 PVC

1.00 1.00 1.00

-20 -10 0 10 20 -20 -10 0 10 20 -20 -10 0 10 20 Amyloid PET Centiloids Amyloid PET Centiloids Amyloid PET Centiloids

Predictors: Age, Sex, Edu, APOE4, FTP, Ab (continuous) Age, Aβ, and tau associations among healthy elderly

Schöll, Lockhart et al, Neuron 2016 righ t H

b-amyloid Neurodegeneration Tau Medial Temporal Lobe tau is Associated with Reduced Glucose Metabolism in Amyloid+ Aging

21 Cognitively normal amyloid Glucose positive older people (78 years) Metabolism in AD

Increases of tau in MTL regions Sparse canonical are associated with correlation

analysis decreases of FDG in temporal, parietal, and orbitofrontal regions

The hypometabolic brain regions reflect downstream neural targets in medial temporal lobe pathways and are also the brain regions affected by Alzheimer’s disease

Adams et al, Cerebral Cortex, 2018 Tau Accumulates Over Time and Parallels Brain Atrophy in Cognitively Normal Older People

42 Cognitively normal adults followed for ~2 years

Longitudinal Tau Longitudinal Atrophy Tau + Atrophy CorticalCorrelation Thinning in AD

Harrison et al, Ann Neurol 2019 Baseline Tau Predicts Longitudinal Atrophy in AD (N = 26 PIB+ AD)

Renaud LaJoie Tau Correlates with AD Phenotypes Group comparison (Ab does not) cov: age, CDR-SB + cortical SUVR Marginal mean maps | centered for age (~65yo), CDR-SB (~4), and global cortical SUVR 2.8 PIB SUVR PIB 0.8

punc<0.001 pFWE<0.05 AD lvPPA PCA PCA > others lvPPA > others (N=60) (N=19) (N=18) AD > others

2.8 FTP SUVR FTP 0.8 Renaud La Joie/Gil Rabinovici Tau and Memory in Normal Older People

Cognitively normal individuals with more tau in these brain regions (entorhinal cortex) show worse performance on memory tests score - Memoryz

Entorhinal Tau Maass et al J Neuroscience 2018 Clinical Trials: Biomarkers for subject selection and target engagement Aducanumab: Mild AD and MCI Amyloid PET Scans: Cognitive Testing: Reductions in Ab Dose-Related Improvement

Antibodies directed at b- amyloid lowered PET scan Clinical benefit by 1-year: less measurements after 26 and 54 cognitive decline with higher weeks of treatment – dose doses effect Sevigny et al Nature 2016 The Future?

Therapeutic trials of amyloid lowering therapies Asymptomatic people with genetic risks Asymptomatic individuals with positive amyloid biomarkers

Other targets Tau lowering therapies Lifestyle interventions (POINTER study)

Biomarkers will play a major role in subject selection and treatment monitoring Thanks

National Institutes of Health Tau Consortium