Alzheimer’s & Dementia - (2018) 1-7

55
1 56
2 Perspective 57
3 58
4 Q1 The National Institute on Aging—Alzheimer’s Association Framework 59
5 60
6 on Alzheimer’s disease: Application to clinical trials 61
7 62
8 63
9 Q4 Jeffrey Cummings* 64
10 Cleveland Clinic Lou Ruvo Center for Brain Health, Las Vegas, NV, USA 65
11 66
12 67
13 68
Abstract Introduction: The National Institute on Aging—Alzheimer’s Association Research Framework on
14 69
Alzheimer’s disease (AD) represents an important advance in the biological characterization of the
15 70
AD spectrum.
16 71
Methods: The National Institute on Aging—Alzheimer’s Association Framework is considered as it
17 72
applies to clinical trials.
18 73
Results: Using the combination of amyloid (A), tau (T), and neurodegeneration (N) biomarkers, the
19 74
Framework provides a means of defining the state of patients with regard to Alzheimer pathologic
20 75
change. The Framework is relevant to clinical trials of disease-modifying agents, allowing partici-
21 76
pants to be characterized biologically at baseline. The ATN Framework can also inform trial out-
22 77
comes. The preclinical phase of the disease after amyloid deposition is defined by A1T2N2, and
23 78
the transition to prodromal disease and dementia is characterized by the addition of T and N. Most
24 79
symptomatic patients in clinical trials are in the class of A1T1N2 and A1T1N1.
25 80
Discussion: The National Institute on Aging—Alzheimer’s Association Framework on AD repre-
26 81
sents progress in providing biomarker profiles of participants in the AD spectrum that can be used
27 82
to help design clinical trials.
28 83
Ó 2018 Published by Elsevier Inc. on behalf of the Alzheimer’s Association. This is an open access
29 84
article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
30 85
31 Keywords: Alzheimer’s disease; Clinical trials; NIA-AA Framework; Biomarkers; Disease modification; Amyloid; Tau; Neu- 86
32 rodegeneration 87
33 88
34 89
35 90
36 1. Introduction markers—amyloid (A), tau (T), and neurodegeneration (N). 91
37 Each of the types of biomarkers has several representatives 92
38
The National Institute on Aging—Alzheimer’s Associa- 93
(Table 1), including cerebrospinal fluid (CSF) measures of
tion (NIA-AA) Alzheimer’s Diagnostic Framework [1]
39 amyloid b-protein 42 (Ab42) and amyloid imaging for A; 94
40 seeks to define the Alzheimer’s spectrum in terms of CSF phospho-tau (p-tau) and tau positron emission tomogra- 95
41 biomarkers and to distinguish Alzheimer’s disease (AD)
phy (PET) imaging for T; and magnetic resonance imaging 96
42 from non-Alzheimer causes of cognitive impairment by 97
(MRI) atrophy, CSF total tau (t-tau), and fluorodeoxyglu-
43 biomarker criteria. The classification uses three types of bio- 98
cose (FDG) PET for N [1,2]. The Framework redefines 99
44
45
AD and its progression in terms of the biomarker profile of 100
46 the condition. 101
47 Declaration of interest: J.C. has provided consultation to Axovant, BiO- 102
48 asis Technologies, Biogen, Boehringer-Ingelheim, Bracket, Dart, Eisai, 103
49
2. NIA-AA Framework application to clinical trials 104
Genentech, Grifols, Intracellular Therapies, Kyowa, Lilly, Lundbeck, Med-
50 Avante, Merck, Neurotrope, Novartis, Nutricia, Orion, Otsuka, Pfizer, Pro- 105
The NIA-AA Framework has the great advantage of
51 biodrug, QR, Resverlogix, Servier, Suven, Takeda, Toyama, and United 106
focusing on the biology of the AD that comprises the reper-
52 Neuroscience companies. 107
*Corresponding author. Tel.: 702-483-6029; Fax: 702-722-6584. toire of targets for pharmacologic treatment of AD with 108
53
54 E-mail address: cumminj@ccf.org disease-modifying therapies (DMTs). There is a synergy 109
https://doi.org/10.1016/j.jalz.2018.05.006
1552-5260/Ó 2018 Published by Elsevier Inc. on behalf of the Alzheimer’s Association. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).

SSU 5.5.0 DTD
JALZ2626_proof
20 June 2018
8:41 pm
ce

 2 J. Cummings / Alzheimer’s & Dementia - (2018) 1-7

110 Table 1 whose tau is confined to temporal lobe regions and an 171
111 Biomarkers included in the amyloid, tau, neurodegeneration (A,T,N) outcome would be a drug-placebo difference in spread to 172
112 classification of the NIA-AA Framework
wider neocortical regions. Alternatively, atrophy can be 173
113 Biomarker class CSF marker Imaging marker quantitated and drug-placebo differences in progression to 174
114 175
Amyloid (A) CSF Ab42 Amyloid imaging more severe atrophy captured as an outcome. Such biolog-
115 176
Tau (T) CSF phospho-tau Tau imaging ical measures have rarely been used to construct trials to
116 Neurodegeneration (N) CSF total tau MRI; FDG PET 177
117 date; the Framework provides a plausible means of begin- 178
118 Abbreviations: Ab42, amyloid-b protein 42 amino acid length; CSF, cere- ning to plan such trials. 179
brospinal fluid; FDG, fluorodeoxyglucose; MRI, magnetic resonance imag- The NIA-AA Framework emphasizes the presence of
119 180
ing; PET, positron emission tomography.
120 Alzheimer-related pathologic change (ATN) but does not 181
121 exclude additional types of pathology that may complicate 182
122 between the proposed classification and biologically based planning and outcomes in clinical trials. A1T2N1 can be 183
123 drug development for AD. Likewise, there are parallels be- mixed dementia with N being attributable to a concomitant 184
124 tween the Framework and the new AD staging system non-AD condition. A1T1N1 can also be mixed dementia 185
125 with the neurodegeneration coming from AD or an associ- 186
proposed by the U.S. Food and Drug Administration [3].
126 187
Here, the advantages and challenges of using this classifica- ated condition. Thus, the Framework will be supplemented
127 188
tion in clinical trials are presented. by other approaches such as excluding patients with cerebro-
128 189
129 vascular disease by MRI to accurately interpret N. Similarly, 190
130 many other types of pathology co-occur with AD changes 191
2.1. Biomarker-based participant selection
131 and are not addressed in the ATN Framework. Alpha- 192
132 Of the eight categories of the NIA-AA Framework, one synuclein TAR DNA-binding protein 43, hippocampal scle- 193
133 applies to primary prevention before any Alzheimer pathobi- rosis, and vascular lesions are frequently present in patients 194
134 ology is present (A2T2N2) and three apply to Alzheimer’s with AD, particularly older individuals. Only 10 to 30 195
135 pathologic change or AD (Table 2). The other four cate- percent of patients with AD have pure Alzheimer’s patho- 196
136 gories do not meet criteria for Alzheimer’s pathology logic changes [13–16]. Additional biomarkers for these co- 197
137 (A2) or have Alzheimer’s pathology plus some non-AD occurring pathologies will be needed if they prove to be 198
138 199
cause of neurodegeneration (A1T2N1). important from a treatment perspective.
139 200
Primary prevention trials may be appropriate for secre- The NIA-AA Framework provides a means of excluding
140 201
141 tase inhibitors such as b-site amyloid precursor protein patients with Alzheimer-related pathologies from trials 202
142 cleavage enzyme inhibitors, where the participants have no focusing on non-AD dementias such as vascular dementia, 203
143 AD-type pathologic changes, and the goal of the trial is to frontotemporal dementia, progressive supranuclear palsy, 204
144 prevent the initiation of the cascade of events leading to and corticobasal degeneration (Table 2). All these cases 205
145 AD in the future [4]. Primary prevention trials could use would be in the A2 category. Some may be T1 by tau 206
146 the Framework to define A2T2N2 participants. This might PET (e.g., tau-related frontotemporal dementia, progressive 207
147 include amyloid PET–negative participants with known supranuclear palsy, and corticobasal degeneration) [17,18], 208
148 disease-causing mutations (presenilin 1, presenilin 2, amy- and most will be N1 unless diagnosed at very early stages 209
149 loid precursor protein, and Down syndrome) or apolipopro- of disease. Diagnosis of these disorders will depend on 210
150 211
tein epsilon 4 (ApoE4) homozygotes at high risk for recognizing a compatible phenotype and supportive brain
151 212
developing AD [5]. This group of participants is character- imaging findings. In this setting, the Framework
152 213
153 ized by the absence of any state biomarker of Alzheimer’s contributes as a means of identifying AD-phenocopies of 214
154 pathologic change (A2). non-AD disorders [12,19]. The ATN approach does not 215
155 Three categories of the Framework apply to AD trials: prioritize any specific aspect of the biology and allows 216
156 A1T2N2; A1T1N2; and A1T1N1. These three cate- systems level analyses of the elements in response to trial 217
157 gories embrace the spectrum of changes from Alzheimer’s interventions [20]. 218
158 pathologic change. Tau imaging has shown that the 219
159 A1T2N2 and A1T1N2 characterize much of the preclin- 220
160 2.2. Biomarker-based participant section Q2
221
ical period of AD with increasing tau signal shown by PET to
161 appear in later preclinical and prodromal phases of the A critical aspect of the Framework approach is to insure 222
162 illness [6,7]. Two classes, A1T1N2 and A1T1N1, that the target pathology is present in the trial population. 223
163 224
account for most individuals with symptomatic AD [8–10]. Anti-amyloid immunotherapies and some small molecules
164 225
165 Although the Framework will assist in clinical trials, require the presence of excessive Ab in the brain, and reduc- 226
166 further biological differentiation will be needed to allow bio- tion of the Ab burden is evidence of target engagement. The 227
167 logical targets to be meaningfully related to disease progres- clinical phenotype is not sufficient to insure the accuracy of 228
168 sion. This might be done by rating severity of the ATN the AD diagnosis and the presence of the key pathology 229
169 changes or by adding additional phase-specific biomarkers [12,21] Amyloid biomarkers are important for diagnostic 230
170 [11,12]. For example, trials could begin with participants confirmation even in trials of nonamyloid therapies. 231

SSU 5.5.0 DTD
JALZ2626_proof
20 June 2018
8:41 pm
ce

 J. Cummings / Alzheimer’s & Dementia - (2018) 1-7 3

232 Table 2 293

233 ATN classification of neurodegenerative disorders including those related to Alzheimer’s disease (AD) 294
234 Amyloid (A) Tau (T) Neurodegeneration (N) Comment Trial population 295
235 296
Negative Negative Negative Normal Primary prevention trials; before amyloid is
236 297
present
237 Positive Negative Negative Alzheimer pathology; this defines preclinical Secondary prevention trials; amyloid is
298
238 AD before any changes associated with present, tau is not; delay of tau spread as a 299
239 amyloid have begun potential outcome 300
240 Positive Positive Negative AD; amyloid and tau changes are present; no Secondary prevention trials; amyloid and tau 301
241 effect on neurodegeneration are present, neurodegeneration is not; 302
242 delay in tau spread or development of 303
243 neurodegeneration are potential outcomes 304
244Q3 Positive Positive Positive AD; amyloid, tau, and neurodegeneration Treatment trials; all three basic biomarkers 305
245 This category will also include mixed are present; slowing of progression or 306
dementia where AD co-exists with other delay to milestone are appropriate designs
246 307
brain disorders such as cerebrovascular Combination treatment trials could include
247 disease. Comorbid conditions contribute this population; for example, trials
308
248 to the neurodegeneration component. including AD and CVD 309
249 Positive Negative Positive Alzheimer pathology plus some other cause Combination treatment trials of anti-amyloid 310
250 of neurodegeneration agent and drugs addressing concomitant 311
251 pathology may be warranted 312
252 Negative Negative Positive Not AD; neurodegeneration only Non-AD trials such as VaD, FTD, PSP, CBD 313
253 Negative Positive Negative Not AD; elevated tau without Non-AD trials of CVD, prion disease, or 314
254 neurodegeneration early tauopathies 315
255 Negative Positive Positive Not AD; elevated tau and neurodegeneration Non-AD trials of VaD or prion disease 316
256 Abbreviations: CBD, corticobasal degeneration; CVD, cerebrovascular disease; FTD, frontotemporal dementia; PSP, progressive supranuclear palsy; VaD, 317
257 vascular dementia. 318
258 NOTE. The three categories most relevant to clinical trials for AD are shaded. The type of trial associated with each group in the classification is noted. 319
259 320
260 321
261
2.3. Clinically based participant selection between amyloid measures and cognition is weak. Correla- 322
262 tions between tau (tau PET and CSF p-tau and clinical mea- 323
Clinical trials cannot depend exclusively on the biolog- sures) are significant in the symptomatic phases of the illness
263 ical profile of the participants and must include characteriza- 324
264 [23,24]. 325
tion of the clinical syndrome; clinical benefit or the ability to
265 Recent studies show that tau PET accounts for approxi- 326
266
predict clinical benefit will be required for drug approval. A mately 30% of the variance on a composite memory test 327
267 clinical benefit is the principal outcome of clinical trials of score in the AD Neuroimaging Initiative cohort [25]. Corre- 328
268 interest to participants and their partners, clinicians, regula- lations between MRI atrophy and ADAS-cog scores in pro- 329
269 tors, and payers. The Framework emphasizes biomarkers dromal AD and AD dementia are in the range of 0.45 [24]. 330
270 over clinical features, but clinical trials will require consid- These observations suggest that other brain pathologies 331
271 eration of both aspects of AD [12]. 332
such as inflammation, oxidation, a-synuclein, TAR DNA-
272 Table 3 uses the numeric clinical stage of individuals on 333
273
binding protein 43, cerebrovascular changes, and host- 334
the Alzheimer continuum proposed by the Framework and related compensatory factors (e.g., cerebral reserve)
274 shows the typical clinical assessments that would be used 335
275 contribute importantly to the profile of cognitive changes 336
to characterize the participants in a clinical trial. The clinical observed in prodromal AD and AD dementia. These pathol-
276 337
stages correspond to the staging system of the Alzheimer’s
277 ogies and the corresponding cognitive deficits may not 338
278
spectrum proposed by the U.S. Food and Drug Administra- respond to anti-amyloid or anti-tau therapies. The NIA-AA 339
279 tion [3]. The ATN classification is juxtaposed with the clin- Framework can establish more homogenous treatment 340
280 ical stage and the assessment instruments. The biological groups; biological heterogeneity and biomarker/clinical dis- 341
281 Framework of the Alzheimer’s continuum is helpful for parities are not eliminated. 342
282 characterizing the biomarker state of the participants and 343
283 will be integrated with clinical staging to construct clinical 344
2.4. Prediction of progression
284 trials that have populations with sufficiently homogeneous 345
285 clinical characteristics to allow trial planning including in- Faster disease progression on clinical measures allows 346
286 clusion and exclusion factors, outcome assessments, recruit- detection of a drug-placebo difference in clinical trials 347
287 ment, and sample size determination. with smaller sample sizes. This translates into faster 348
288 349
The clinical linkages to the ATN Framework are modest. decision-making and could ultimately accelerate getting
289 350
290
Amyloid status (CSF Ab42 or amyloid PET) has a minor new therapies to AD patients and those at risk for AD. The 351
291 impact on cognition over the long preclinical period in NIA-AA Framework provides support for predicting pro- 352
292 which amyloid is present in the brain [22]; correlation gression [1]. The highest rates of short-term progression 353

SSU 5.5.0 DTD
JALZ2626_proof
20 June 2018
8:41 pm
ce

 4 J. Cummings / Alzheimer’s & Dementia - (2018) 1-7

354 Table 3 415

355 Numeric clinical stages of individuals on the Alzheimer continuum, examples of the clinical assessments that would be used to characterize the participants in a 416
356 clinical trial, and corresponding ATN characteristics at each stage 417
357 Numeric clinical Cognitive/composite Functional Behavioral 418
358 staging Clinical phase assessment assessment assessment ATN classification 419
359 1 Cognitively normal with no PACC; APCC No functional deficit No established A1T2N2
420
360 indication of decline behavioral change 421
361 2 Cognitively normal with PACC; APCC No functional deficit No established A1T1N2 or 422
362 indication of decline behavioral change A1T1N1 423
363 3 Prodromal AD CDR-sb; ADCOMS; ADCS-ADL MBI/NPI A1T1N1 424
364 iADRS; NTB (MCI version) 425
365 4 Mild AD dementia ADAS-cog; NTB; CDR-sb ADCS-ADL NPI A1T1N1 426
366 5 Moderate AD dementia ADAS-cog; NTB; CDR-sb ADCS-ADL NPI A1T1N1 427
367 6 Severe AD dementia SIB; CDR-sb ADCS-ADL NPI A1T1N1 428
(severe version)
368 429
369 Abbreviations: A, amyloid; ADAS-cog, Alzheimer’s Disease Assessment Scale—cognitive subscale; ADCOMS, AD Composite Score; ADCS-ADL (MCI), 430
370 Alzheimer’s Disease Cooperative Study—Activities of Daily Living scale (mild cognitive impairment version); APCC, Alzheimer Prevention Initiative Cogni- 431
371 tive Composite; CDR-sb, Clinical Dementia Rating–Sum of Boxes; FDA, Food and Drug Administration; iADRS, Integrated AD Rating Scale; MBI, Minimal 432
372 Behavioral Impairment scale; N, neurodegeneration; NPI, Neuropsychiatric Inventory; NTB, Neuropsychological Test Battery; PACC, Preclinical Alzheimer 433
Cognitive Composite; SIB, severe impairment battery; T, tau.
373 434
NOTE. The Neurological Clinical Staging is consistent with the recent FDA guidance of AD stages (U.S. Food and Drug Administration. Early Alzheimer’s
374 435
disease: developing drugs for treatment; draft guidance for industry. Federal Register; 2018. p. 7060–1).
375 436
376 437
377 are in the A1T1N2 and A1T1N1 classes of AD. CSF presence of the target pathology and provides the opportu- 438
378 p-tau levels predict MRI progression in longitudinal studies nity for the ATN classification to be used to demonstrate 439
379 of patients with mild AD dementia and in cognitively normal target engagement of A or T. 440
380 participants [26]. FDG PET and t-tau (markers of N in the 441
381 442
ATN classification) predict clinical progression to AD de-
382 2.6. Outcomes for trials of DMTs 443
383
mentia in patients with MCI at baseline [27]. Tau PET at 444
384 baseline predicts progressive cognitive decline, but this ef- Successful DMTs will exert neuroprotection [30,31]. 445
385 fect is greater in older than younger individuals [12,28]. Neurodegeneration in the ATN Framework is assessed by 446
386 Choosing participants with the progression-prediction ATN MRI atrophy, CSF t-tau, for FDG PET. FDG PET can be 447
387 biotypes will help insure measurable cognitive decline in impacted by symptomatic therapies, and these effects must 448
388 the trial period and an improved chance of observing a be considered when interpreting observations considered 449
389 drug-placebo difference. 450
indicative of disease modification [32,33]. Drug-placebo dif-
390 451
391
ferences in MRI atrophy or t-tau at trial termination would 452
392 2.5. Target engagement provide the most compelling evidence of disease modifica- 453
393 tion. MRI has often performed irregularly as an outcome 454
The NIA-AA Framework is aimed at establishing the
394 in clinical trials, with greater atrophy in the treatment group 455
biomarker characteristics of the Alzheimer’s spectrum and
395 [34–36], and dependence on this measure as an outcome to 456
not with other uses of the biomarkers in AD drug develop-
396 support disease modification has uncertainties. More 457
397 ment [1]. Development of DMTs depends on demonstrating 458
measures of successful amelioration of neurodegeneration
398 target engagement in phase II trials to insure that near- and 459
are needed to serve as outcomes in DMT trials.
399 intermediate-term steps critical to disease modification are 460
The Framework provides guidance for uniform reporting
400 being achieved [29]. A successful DMT must exert neuro-
of biological outcomes in clinical trials, which is critical to 461
401 protection, and this goal will be reflected most closely in a 462
building understanding of the relationship of drug mecha-
402 drug-placebo difference on N [30,31]. A and T are 463
403
nisms to their biological consequences [37]. 464
intermediate targets whose modification may result in
404 neuroprotection and disease modification through linked 465
405 mechanisms. 466
406 3. Discussion 467
Amyloid plaque burden on amyloid PET, tau aggregation
407 468
on tau PET, and CSF measures of Ab42 or p-tau can function The NIA-AA Framework is a useful advance in
408 469
409 as intermediate measures of target engagement. Measure- biomarker classification of the Alzheimer’s spectrum [1]. 470
410 ment of tau in trial participants allows tau PET or CSF p- It uses the ATN biomarkers to establish the continuum, al- 471
411 tau to be used as measures of target engagement. Measures lows staging of trial participants, insures the presence of 472
412 could include prevention of tau accumulation in A1T2 in- target pathology, provides a framework of considering 473
413 dividuals, spread in A1T1 participants, or reduction of ATN as outcomes supportive of disease modification, and fa- 474
414 tau in A1T1 participants. The Framework establishes the cilitates a means of excluding Alzheimer’s spectrum 475

SSU 5.5.0 DTD
JALZ2626_proof
20 June 2018
8:41 pm
ce

 J. Cummings / Alzheimer’s & Dementia - (2018) 1-7 5

476 individuals from trials intended to address non-Alzheimer clinical diagnosis of AD unconfirmed by biomarkers in 537
477 disorders. some studies. Progress has been made on both of these 538
478 Two forms of Ab measures are used in the ATN fronts. 539
479 approach—CSF Ab42 and amyloid PET. There is substantial Trial planners must be aware of the limits of biomarkers 540
480 541
evidence that the most neurotoxic species of Ab comprise as reflections of the state of the pathology in the brain; the
481 542
oligomers with variable lengths of amino acids from dimers latter must be impacted by effective biological therapies.
482 543
483 to dodecomers and higher order pre-fibrillar species [38–40]. 544
484 There is no consensus measure of these forms of Ab, and the 545
4. Summary
485 relationship of oligomers to the ATN classification remains 546
486 to be clarified. Some approaches of anti-amyloid therapy The NIA-AA Research Framework [1] is an important 547
487 may depend on an effect on oligomers. advance in using biomarkers to define the AD spectrum. 548
488 The Framework poses a CSF and a brain imaging alterna- This approach will be useful in designing clinical trials 549
489 tive for each ATN category [1] (Table 1). This is useful and and has an important role in characterizing the biomarker 550
490 may facilitate trial recruitment by allowing participants with profile of participants entering trials. The Framework will 551
491 access to either technology to enter the trial (this is espe- 552
assist in excluding Alzheimer spectrum participants from tri-
492 553
cially important in global trials where access to amyloid als of non-Alzheimer dementias. The Framework facilitates
493 554
PET is limited). The CSF measures of amyloid and amyloid considerations of target engagement and trial outcomes. The
494 555
495 PET are highly correlated [41], but they measure different Framework effectively captures the current understanding of 556
496 forms of amyloid and use of the either/or approach may biomarkers of the Alzheimer’s spectrum using the ATN 557
497 increase the heterogeneity of the trial population with approach. The eight types of ATN biomarker profiles include 558
498 unknown consequences. Similarly, the correlation of CSF one with no abnormalities, four with amyloid changes (one 559
499 p-tau to tau PET is significant in AD dementia, where tau with features of mixed dementia and one limited to amyloid 560
500 PET shows extensive changes but not in preclinical AD changes characteristic of Alzheimer’s pathological changes 561
501 where CSF p-tau levels are similar to those of AD dementia without AD), and three with non-AD type profiles; most of 562
502 but PET abnormalities are geographically limited the symptomatic forms comprise two types (A1T1N2; 563
503 [23,42,43]. Tau PET Standard Uptake Volume Ratios 564
A1T1N1). Mixed dementias and the complex neuropa-
504 565
account for about one-third of the variance of hippocampal thology of AD are not addressed in the ATN Framework
505 566
atrophy and about 20% of the variance of cortical atrophy [13–15]. Characterizing trial participants with multiple
506 567
507 among individuals with positive amyloid imaging [44]. biomarkers (e.g., specific levels of T and N in A1 568
508 MRI atrophy correlates with t-tau in some investigations, individuals) may complicate recruitment while improving 569
509 but the correlations have not been observed in all studies the biological definition of the trial population. The 570
510 [45–47]. Agreement between MRI volumetric measures imperfect relationships between the two amyloid, two tau, 571
511 and FDG PET (2 N measures) is inconsistent across and three neurodegeneration markers may contribute to 572
512 studies [48]. Tau and FDG PET (2 N measures) are not trial population heterogeneity, and the less-than-complete 573
513 well correlated [49,50]; and agreement between MRI correlations between biomarkers and brain pathology chal- 574
514 atrophy and t-tau (N markers) is limited [48]. Trial design lenge researchers to refine the Framework. Finally, the 575
515 will need to anticipate the effects of applying the ATN 576
limited correlation of cognitive decline with ATN biomarker
516 577
criteria by CSF, imaging, or mixed approaches. changes demonstrates that factors outside this repertoire are
517 578
Biomarkers are the best window on the biology of Alz- likely contributing to the cognitive impairment. Biomarkers
518 579
519 heimer’s pathological changes as the brain itself is inacces- of these other pathologies and treatments to address them 580
520 sible. Investigators use CSF measures of Ab42 and amyloid may play an important role in the quest to find DMTs for 581
521 PET as surrogates for brain amyloid; CSF p-tau and tau those with or at risk for AD. The NIA-AA Framework is a 582
522 PET as surrogate measures of brain tau pathology; and key advance in establishing a clinically viable biologically 583
523 t-tau, FDG PET, and MRI as surrogates for neurodegenera- defined characterization of the Alzheimer’s spectrum with 584
524 tion. Autopsy studies support these relationships but also application to clinical trials. 585
525 reveal that correlations between biomarkers and brain 586
526 changes are imperfect. Neurofibrillary tangles correlate 587
527 Acknowledgments 588
with changes in hippocampal measures observed on MRI
528 in some studies but not others [46,51,52]. Plaques do not 589
J.C. acknowledges support of a COBRE grant from the NIH/
529 590
correlate with MRI atrophy measures [46,51]. CSF p-tau NIGMS (P20GM109025) and Keep Memory Alive.
530 591
531 and t-tau correlate with neurofibrillary tangle burden in 592
532 some studies by not others [46,53]. Variability in the References 593
533 correlation of CSF Ab42 to extracellular amyloid plaques 594
534 observed at autopsy has also been observed [53,54]. Some [1] Jack CR Jr, Bennett DA, Blennow K, Carrillo MC, Dunn B, Elliott C, 595
535 of the inconsistencies in the literature may reflect et al. NIA-AA research framework: towards a biological definition of 596
536 challenges with CSF assay standardization and the use of Alzheimer’s disease. Alzheimers Dement 2018;14:535–62. 597

SSU 5.5.0 DTD
JALZ2626_proof
20 June 2018
8:41 pm
ce

 6 J. Cummings / Alzheimer’s & Dementia - (2018) 1-7

598 [2] Jack CR Jr, Bennett DA, Blennow K, Carrillo MC, Feldman HH, [21] Sevigny J, Suhy J, Chiao P, Chen T, Klein G, Purcell D, et al. Amyloid 659
599 Frisoni GB, et al. A/T/N: An unbiased descriptive classification PET screening for enrichment of early-stage Alzheimer disease clin- 660
600 scheme for Alzheimer disease biomarkers. Neurology 2016; ical trials: experience in a phase 1b clinical trial. Alzheimer Dis Assoc 661
601 87:539–47. Disord 2016;30:1–7. 662
[3] U.S. Food and Drug Administration. Early Alzheimer’s Disease: [22] Donohue MC, Sperling RA, Petersen R, Sun CK, Weiner MW,
602 663
Developing Drugs for Treatment; Draft Guidance for Industry. Federal Aisen PS. Association between elevated brain amyloid and subsequent
603 664
Register; 2018;. p. 7060–1. cognitive decline among cognitively normal persons. JAMA 2017;
604 [4] Kennedy ME, Stamford AW, Chen X, Cox K, Cumming JN, 317:2305–16. 665
605 Dockendorf MF, et al. The BACE1 inhibitor verubecestat (MK- [23] Brier MR, Gordon B, Friedrichsen K, McCarthy J, Stern A, 666
606 8931) reduces CNS beta-amyloid in animal models and in Alzheimer’s Christensen J, et al. Tau and Abeta imaging, CSF measures, and cogni- 667
607 disease patients. Sci Transl Med 2016;8:363ra150. tion in Alzheimer’s disease. Sci Transl Med 2016;8:338ra66. 668
608 [5] Reiman EM, Langbaum JB, Fleisher AS, Caselli RJ, Chen K, [24] Pontecorvo MJ, Devous MD Sr, Navitsky M, Lu M, Salloway S, 669
609 Ayutyanont N, et al. Alzheimer’s Prevention Initiative: a plan to accel- Schaerf FW, et al. Relationships between flortaucipir PET tau binding 670
610 erate the evaluation of presymptomatic treatments. J Alzheimers Dis and amyloid burden, clinical diagnosis, age and cognition. Brain 2017; 671
611 2011;26 Suppl 3:321–9. 140:748–63. 672
[6] Cho H, Choi JY, Hwang MS, Lee JH, Kim YJ, Lee HM, et al. Tau PET [25] Tosun D, Landau S, Aisen PS, Petersen RC, Mintun M, Jagust W, et al.
612 673
in Alzheimer disease and mild cognitive impairment. Neurology 2016; Association between tau deposition and antecedent amyloid-beta
613 674
87:375–83. accumulation rates in normal and early symptomatic individuals.
614 [7] Johnson KA, Schultz A, Betensky RA, Becker JA, Sepulcre J, Rentz D, Brain 2017;140:1499–512. 675
615 et al. Tau positron emission tomographic imaging in aging and early [26] Tarawneh R, Head D, Allison S, Buckles V, Fagan AM, Ladenson JH, 676
616 Alzheimer disease. Ann Neurol 2016;79:110–9. et al. Cerebrospinal fluid markers of neurodegeneration and rates of 677
617 [8] Fortea J, Vilaplana E, Alcolea D, Carmona-Iragui M, Sanchez- brain atrophy in early Alzheimer disease. JAMA Neurol 2015; 678
618 Saudinos MB, Sala I, et al. Cerebrospinal fluid beta-amyloid and 72:656–65. 679
619 phospho-tau biomarker interactions affecting brain structure in pre- [27] Choo IH, Ni R, Scholl M, Wall A, Almkvist O, Nordberg A. Combi- 680
620 clinical Alzheimer disease. Ann Neurol 2014;76:223–30. nation of 18F-FDG PET and cerebrospinal fluid biomarkers as a better 681
621 [9] Baxter LC, Sparks DL, Johnson SC, Lenoski B, Lopez JE, Connor DJ, predictor of the progression to Alzheimer’s disease in mild cognitive 682
et al. Relationship of cognitive measures and gray and white matter in impairment patients. J Alzheimers Dis 2013;33:929–39.
622 683
Alzheimer’s disease. J Alzheimers Dis 2006;9:253–60. [28] Koychev I, Gunn RN, Firouzian A, Lawson J, Zamboni G, Ridha B,
623 684
[10] Kang JM, Lee SY, Seo S, Jeong HJ, Woo SH, Lee H, et al. Tau positron et al. PET tau and amyloid-beta burden in mild Alzheimer’s disease:
624 emission tomography using [18F]THK5351 and cerebral glucose hy- divergent relationship with age, cognition, and cerebrospinal fluid bio- 685
625 pometabolism in Alzheimer’s disease. Neurobiol Aging 2017; markers. J Alzheimers Dis 2017;60:283–93. 686
626 59:210–9. [29] Cummings J, Ritter A, Zhong K. Clinical trials for disease-modifying 687
627 [11] Grothe MJ, Barthel H, Sepulcre J, Dyrba M, Sabri O, Teipel SJ, et al. therapies in alzheimer’s disease: a primer, lessons learned, and a blue- 688
628 In vivo staging of regional amyloid deposition. Neurology 2017; print for the future. J Alzheimers Dis. in press. 689
629 89:2031–8. [30] Cummings JL, Fox N. Defining disease modification for Alzheimer’s 690
630 [12] Knopman DS, Haeberlein SB, Carrillo MC, Hendrix JA, Kerchner G, disease clinical trials. J Prev Alzheimers Dis 2017;4:109–15. 691
631 Margolin R, et al. The National Institute on Aging and the Alzheimer’s [31] Cummings J. Disease modification and Neuroprotection in neurode- 692
Association Research Framework for Alzheimer’s disease: perspec- generative disorders. Transl Neurodegener 2017;6:25.
632 693
tives from the research roundtable. Alzheimers Dement 2018; [32] Mega MS, Cummings JL, O’Connor SM, Dinov ID, Reback E, Felix J,
633 694
14:563–75. et al. Cognitive and metabolic responses to metrifonate therapy in Alz-
634 [13] Abner EL, Kryscio RJ, Schmitt FA, Fardo DW, Moga DC, heimer disease. Neuropsychiatry Neuropsychol Behav Neurol 2001; 695
635 Ighodaro ET, et al. Outcomes after diagnosis of mild cognitive impair- 14:63–8. 696
636 ment in a large autopsy series. Ann Neurol 2017;81:549–59. [33] Mega MS, Dinov ID, Porter V, Chow G, Reback E, Davoodi P, et al. 697
637 [14] James BD, Wilson RS, Boyle PA, Trojanowski JQ, Bennett DA, Metabolic patterns associated with the clinical response to galant- 698
638 Schneider JA. TDP-43 stage, mixed pathologies, and clinical Alz- amine therapy: A fludeoxyglucose f 18 positron emission tomographic 699
639 heimer’s-type dementia. Brain 2016;139:2983–93. study. Arch Neurol 2005;62:721–8. 700
640 [15] Schneider JA, Arvanitakis Z, Leurgans SE, Bennett DA. The neuropa- [34] Fleisher AS, Truran D, Mai JT, Langbaum JB, Aisen PS, 701
641 thology of probable Alzheimer disease and mild cognitive impairment. Cummings JL, et al. Chronic divalproex sodium use and brain atrophy 702
Ann Neurol 2009;66:200–8. in Alzheimer disease. Neurology 2011;77:1263–71.
642 703
[16] Boyle PA, Yu L, Wilson RS, Leurgans SE, Schneider JA, Bennett DA. [35] Fox NC, Black RS, Gilman S, Rossor MN, Griffith SG, Jenkins L, et al.
643 704
Person-specific contribution of neuropathologies to cognitive loss in Effects of Abeta immunization (AN1792) on MRI measures of cere-
644 old age. Ann Neurol 2018;83:74–83. bral volume in Alzheimer disease. Neurology 2005;64:1563–72. 705
645 [17] Hall B, Mak E, Cervenka S, Aigbirhio FI, Rowe JB, O’Brien JT. [36] Novak G, Fox N, Clegg S, Nielsen C, Einstein S, Lu Y, et al. Changes 706
646 In vivo tau PET imaging in dementia: pathophysiology, radiotracer in brain volume with bapineuzumab in mild to moderate Alzheimer’s 707
647 quantification, and a systematic review of clinical findings. Ageing disease. J Alzheimers Dis 2015;49:1123–34. 708
648 Res Rev 2017;36:50–63. [37] Silverberg N, Elliott C, Ryan L, Masliah E, Hodes R. NIA commentary 709
649 [18] Passamonti L, Vazquez Rodriguez P, Hong YT, Allinson KS, on the NIA-AA Research Framework: towards a biological definition 710
650 Williamson D, Borchert RJ, et al. 18F-AV-1451 positron emission to- of Alzheimer’s disease. Alzheimers Dement 2018;14:576–8. 711
651 mography in Alzheimer’s disease and progressive supranuclear palsy. [38] Jongbloed W, Bruggink KA, Kester MI, Visser PJ, Scheltens P, 712
Brain 2017;140:781–91. Blankenstein MA, et al. Amyloid-beta oligomers relate to cogni-
652 713
[19] Shim YS, Roe CM, Buckles VD, Morris JC. Clinicopathologic study tive decline in Alzheimer’s disease. J Alzheimers Dis 2015;
653 714
of Alzheimer’s disease: Alzheimer mimics. J Alzheimers Dis 2013; 45:35–43.
654 35:799–811. [39] Kayed R, Lasagna-Reeves CA. Molecular mechanisms of amyloid 715
655 [20] Khachaturian AS, Hayden KM, Mielke MM, Tang Y, Lutz MW, oligomers toxicity. J Alzheimers Dis 2013;33 Suppl 1:S67–78. 716
656 Gustafson DR, et al. Future prospects and challenges for Alzheimer’s [40] Viola KL, Klein WL. Amyloid beta oligomers in Alzheimer’s disease 717
657 disease drug development in the era of the NIA-AA Research Frame- pathogenesis, treatment, and diagnosis. Acta Neuropathol 2015; 718
658 work. Alzheimers Dement 2018;14:532–4. 129:183–206. 719

SSU 5.5.0 DTD
JALZ2626_proof
20 June 2018
8:41 pm
ce

 J. Cummings / Alzheimer’s & Dementia - (2018) 1-7 7

720 [41] Li QX, Villemagne VL, Doecke JD, Rembach A, Sarros S, Varghese S, [48] Alexopoulos P, Kriett L, Haller B, Klupp E, Gray K, Grimmer T, et al. 781
721 et al. Alzheimer’s disease normative cerebrospinal fluid biomarkers Limited agreement between biomarkers of neuronal injury at different 782
722 validated in PET amyloid-beta characterized subjects from the Austra- stages of Alzheimer’s disease. Alzheimers Dement 2014;10:684–9. 783
723 lian Imaging, Biomarkers and Lifestyle (AIBL) study. J Alzheimers [49] Vukovich R, Perneczky R, Drzezga A, Forstl H, Kurz A, 784
Dis 2015;48:175–87. Riemenschneider M. Brain metabolic correlates of cerebrospinal fluid
724 785
[42] Gordon BA, Friedrichsen K, Brier M, Blazey T, Su Y, Christensen J, beta-amyloid 42 and tau in Alzheimer’s disease. Dement Geriatr Cogn
725 786
et al. The relationship between cerebrospinal fluid markers of Alz- Disord 2009;27:474–80.
726 heimer pathology and positron emission tomography tau imaging. [50] Pascoal TA, Mathotaarachchi S, Mohades S, Benedet AL, Chung CO, 787
727 Brain 2016;139:2249–60. Shin M, et al. Amyloid-beta and hyperphosphorylated tau synergy 788
728 [43] Mattsson N, Smith R, Strandberg O, Palmqvist S, Scholl M, Insel PS, drives metabolic decline in preclinical Alzheimer’s disease. Mol Psy- 789
729 et al. Comparing (18)F-AV-1451 with CSF t-tau and p-tau for diag- chiatry 2017;22:306–11. 790
730 nosis of Alzheimer disease. Neurology 2018;90:e388–95. [51] Burton EJ, Mukaetova-Ladinska EB, Perry RH, Jaros E, Barber R, 791
731 [44] Wang L, Benzinger TL, Su Y, Christensen J, Friedrichsen K, Aldea P, O’Brien JT. Quantitative neurodegenerative pathology does not 792
732 et al. Evaluation of tau imaging in staging Alzheimer disease and explain the degree of hippocampal atrophy on MRI in degenerative de- 793
733 revealing interactions between beta-amyloid and tauopathy. JAMA mentia. Int J Geriatr Psychiatry 2012;27:1267–74. 794
Neurol 2016;73:1070–7. [52] Csernansky JG, Hamstra J, Wang L, McKeel D, Price JL, Gado M,
734 795
[45] Apostolova LG, Hwang KS, Andrawis JP, Green AE, Babakchanian S, et al. Correlations between antemortem hippocampal volume and post-
735 796
Morra JH, et al. 3D PIB and CSF biomarker associations with mortem neuropathology in AD subjects. Alzheimer Dis Assoc Disord
736 hippocampal atrophy in ADNI subjects. Neurobiol Aging 2010; 2004;18:190–5. 797
737 31:1284–303. [53] Engelborghs S, Sleegers K, Cras P, Brouwers N, Serneels S, De 798
738 [46] de Souza LC, Chupin M, Lamari F, Jardel C, Leclercq D, Colliot O, Leenheir E, et al. No association of CSF biomarkers with APOEepsi- 799
739 et al. CSF tau markers are correlated with hippocampal volume in Alz- lon4, plaque and tangle burden in definite Alzheimer’s disease. Brain 800
740 heimer’s disease. Neurobiol Aging 2012;33:1253–7. 2007;130:2320–6. 801
741 [47] Ossenkoppele R, Mattsson N, Teunissen CE, Barkhof F, Pijnenburg Y, [54] Brunnstrom H, Hansson O, Zetterberg H, Londos E, Englund E. Cor- 802
742 Scheltens P, et al. Cerebrospinal fluid biomarkers and cerebral atrophy relations of CSF tau and amyloid levels with Alzheimer pathology in 803
743 in distinct clinical variants of probable Alzheimer’s disease. Neurobiol neuropathologically verified dementia with Lewy bodies. Int J Geriatr 804
Aging 2015;36:2340–7. Psychiatry 2013;28:738–44.
744 805
745 806
746 807
747 808
748 809
749 810
750 811
751 812
752 813
753 814
754 815
755 816
756 817
757 818
758 819
759 820
760 821
761 822
762 823
763 824
764 825
765 826
766 827
767 828
768 829
769 830
770 831
771 832
772 833
773 834
774 835
775 836
776 837
777 838
778 839
779 840
780 841

SSU 5.5.0 DTD
JALZ2626_proof
20 June 2018
8:41 pm
ce