Differential associations between neocortical tau pathology and blood flow with cognitive deficits in early-onset vs late-onset Alzheimer’s disease

Purpose Early-onset Alzheimer’s disease (EOAD) and late-onset Alzheimer’s disease (LOAD) differ in neuropathological burden and type of cognitive deficits. Assessing tau pathology and relative cerebral blood flow (rCBF) measured with [18F]flortaucipir PET in relation to cognition may help explain these differences between EOAD and LOAD. Methods Seventy-nine amyloid-positive individuals with a clinical diagnosis of AD (EOAD: n = 35, age-at-PET = 59 ± 5, MMSE = 23 ± 4; LOAD: n = 44, age-at-PET = 71 ± 5, MMSE = 23 ± 4) underwent a 130-min dynamic [18F]flortaucipir PET scan and extensive neuropsychological assessment. We extracted binding potentials (BPND) and R1 (proxy of rCBF) from parametric images using receptor parametric mapping, in medial and lateral temporal, parietal, occipital, and frontal regions-of-interest and used nine neuropsychological tests covering memory, attention, language, and executive functioning. We first examined differences between EOAD and LOAD in BPND or R1 using ANOVA (region-of-interest analysis) and voxel-wise contrasts. Next, we performed linear regression models to test for potential interaction effects between age-at-onset and BPND/R1 on cognition. Results Both region-of-interest and voxel-wise contrasts showed higher [18F]flortaucipir BPND values across all neocortical regions in EOAD. By contrast, LOAD patients had lower R1 values (indicative of more reduced rCBF) in medial temporal regions. For both tau and flow in lateral temporal, and occipitoparietal regions, associations with cognitive impairment were stronger in EOAD than in LOAD (EOAD BPND − 0.76 ≤ stβ ≤  − 0.48 vs LOAD − 0.18 ≤ stβ ≤  − 0.02; EOAD R1 0.37 ≤ stβ ≤ 0.84 vs LOAD − 0.25 ≤ stβ ≤ 0.16). Conclusions Compared to LOAD, the degree of lateral temporal and occipitoparietal tau pathology and relative cerebral blood-flow is more strongly associated with cognition in EOAD. Supplementary Information The online version contains supplementary material available at 10.1007/s00259-021-05669-6.


Introduction
Alzheimer's disease (AD) is characterized by depositions of amyloid-β plaques and hyperphosphorylated tau tangles [1]. The symptoms of AD are heterogeneous, partly related to age-at-onset. Late-onset AD is primarily associated with amnestic problems, while non-amnestic symptoms are more prominent when the disease manifests at younger age [2][3][4][5][6]. Atypical cognitive profiles in early-onset AD patients have previously been linked to specific spatial patterns of hypometabolism, atrophy, and more recently, tau pathology, that primarily affect the neocortex with relative sparing of the medial temporal lobe [7][8][9][10].
Age-at-onset is thus closely related to both cognitive symptoms and distinct patterns of tau pathology in AD. However, despite the notion that tau pathology is tightly linked to the degree of cognitive impairment in AD [11][12][13][14][15][16][17], it is currently unknown whether the association between tau pathology and cognitive performance is affected by age-atonset. To address this question, we conducted a dynamic [ 18 F]flortaucipir PET study. The dynamic scan protocol additionally yields a measure of R 1 . R 1 represents the rate constant for ligand transfer from blood to tissue (K 1 ) in the target region relative to the reference region, and can be interpreted as a proxy of relative cerebral blood flow [rCBF] [18][19][20][21][22][23][24]. Several studies have demonstrated strong correlations between R 1 and metabolic activity as measured with [ 18 F]FDG PET [19,20,24,25]. Furthermore, in a previous study, we showed that [ 18 F]flortaucipir R 1 is strongly associated with cognitive performance in AD, (partly) independently of tau pathology [17]. We therefore also investigated differences in rCBF and its association with cognitive performance, taking into account age-at-onset.
Accordingly, we aimed to investigate the differences between early-onset AD and late-onset AD in (1) tau pathology and rCBF and (2) the associations between tau pathology and rCBF with cognitive performance. We hypothesized that relative to late-onset AD, early-onset AD displays higher levels of tau pathology. As R 1 is tightly linked to hypometabolism measured with FDG-PET [19,20,24,25], and lower levels of metabolism have previously been reported in earlyonset AD patients, we expected to find lower levels of R 1 in early-vs late-onset AD. Furthermore, as early-onset AD may represent a more "pure" form of AD with relatively few comorbidities compared to late-onset AD [26], we hypothesized that tau pathology and cognitive performance would be more strongly associated in early-vs late-onset AD.

Participants
We included 79 subjects from the Amsterdam Dementia Cohort with either probable AD dementia [27] (n = 68) or mild cognitive impairment (MCI) due-to-AD [28] (n = 11), with positive amyloid-β biomarkers [29][30][31]. All subjects underwent a standardized dementia screening, including medical history, extensive neuropsychological assessment, physical and neurological examination, lumbar puncture, blood tests, electroencephalography, and brain MRI. Diagnosis was established by consensus in a multidisciplinary meeting. The diagnosis of MCI/AD met core clinical criteria [27,28] according to the National Institute on Aging and Alzheimer's Association (NIA-AA) and were biomarker supported, with an AD-like CSF profile (i.e., tau/Aβ42 fraction > 0.52) [32,33] and/or an Aβ-PET scan ([ 11 C]PiB or [ 18 F]florbetaben) showing substantial amyloid accumulation when visually assessed by an experienced nuclear medicine physician (BvB). When both CSF and PET measures of amyloid were available, PET visual read was used to determine amyloid status. Subjects were assigned to either the early-or late-onset AD group, based on a median split (age-at-PET). Since the median age in our sample was 66 years, this led to a similar age cut-off compared to the conventional threshold of 65 years [2,3,6,34]. The sample included eight patients meeting criteria for an atypical variant of AD, including two behavioral variant AD (bvAD) and six posterior cortical atrophy (PCA) patients [31,35]. All atypical AD patients were part of the early-onset AD group, except one bvAD patient. Exclusion criteria for all participants were (1) dementia not due-to-AD, (2) significant cerebrovascular disease on MRI (e.g., major cerebrovascular accident), (3) major traumatic brain injury, (4) major psychiatric or neurological disorders other than AD, and (5) (recent) substance abuse. The study protocol was approved by the Medical Ethics Review Committee of the Amsterdam UMC, location VU Medical center. All patients provided written informed consent.

[ 18 F]flortaucipir PET
Acquisition and processing of [ 18 F]flortaucipir PET images is described in detail elsewhere [17,25,36]. In short, dynamic 130-min [ 18 F]flortaucipir PET scans were acquired on a Philips Ingenuity TF-64 PET/CT scanner. The scanning protocol consisted of two dynamic PET scans of 60 and 50 min, respectively, with an inbetween 20-min break [36]. The first 60-min dynamic scan started simultaneously with a bolus injection of approximately 234 ± 14 MBq [ 18 F]flortaucipir (injected mass 1.2 ± 0.9 μg). The second PET scan was co-registered to the first dynamic PET scan using Vinci software (Max Planck Institute, Germany). PET data were acquired in list mode and subsequently reconstructed using 3D-RAMLA including standard corrections for dead time, decay, attenuation, random, and scatter. Patients also underwent 3DT1-weighted MRI (Ingenuity TF PET/MR, Philips Medical Systems, The Netherlands) for anatomical and tissue segmentation purposes.

Image analysis
The 3DT1-weighted MRI images were co-registered to their corresponding PET images using Vinci software. Anatomical regions-of-interest according to the Hammers template were subsequently delineated on the MR images and superimposed on the PET scan using PVElab [37]. Using receptor parametric mapping (RPM) with cerebellar gray matter as reference region, we generated binding potential (BP ND ) maps as a measure for tau pathology. Additionally, and similar to previous work [17,25], R 1 was generated as a proxy for rCBF [18]. Partial volume correction was applied to the PET images using Van Cittert iterative deconvolution methods (IDM), combined with highly constrained back-projections (HYPR) as described previously [17,38,39]. Uncorrected data are shown in the main manuscript and partial volume-corrected data are shown in the Supplementary material. In line with a previous study [17], we calculated BP ND and R 1 in frontal, occipital, parietal, medial, and lateral temporal bilateral cortical lobar regions-of-interest.
In addition to the region-of-interest analyses, we performed voxel-wise analyses to (1) create average images of the different diagnostic groups (early-and late-onset AD without atypical AD cases, PCA, and bvAD), and (2) explore more fine-grained differences in [ 18 F]flortaucipir BP ND and R 1 between early-and late-onset AD. Therefore, we warped all native space parametric BP ND and R 1 images to Montreal Neurological Institute (MNI152) space using the transformation matrixes derived from warping the coregistered T1-weighted MRI scans to MNI space using Statistical Parametric Mapping (SPM) version 12. All warped images were visually inspected for transformation errors and quality control. Average images were calculated using SPM12.

Neuropsychological assessment
We included eight neuropsychological tests (< 1 year of PET scan), including the Dutch version of the Rey Auditory Verbal Learning Test immediate recall and delayed recall (episodic memory), Digit Span forward and backward (attention/executive functioning), Trail Making Test (TMT) version A and B (attention/executive functioning), Letter Fluency test (D-A-T) (executive functioning), and Category Fluency version animals (language) [40]. TMT-A and -B scores were inverted so that lower scores indicated worse performance. In addition, the Mini-Mental State Examination (MMSE) served as a measure of global cognition [41]. For participants who had TMT-A available but were missing TMT-B, we estimated the TMT-B by multiplying the time needed to complete the TMT-A with the mean TMT-B/A ratio from the respective diagnostic group in the respective cohort. To be included in the cognition analyses, participants were required to have at least 6 out of 9 cognitive tests available, and 72 (91%) participants met this criterion (30 early-and 42 late-onset AD). Among the excluded cases, there was one PCA case (early-onset AD group). sFigure-1 shows the missing values in the cognitionsubsample for each neuropsychological test. There were on average 5% missing data in this subsample, for which we performed single imputation (with 5 iterations) for missing data, using all demographic (except for age and APOE4 status), neuropsychological, and neuroimaging variables as predictors [42,43].

Statistical analysis
Differences in demographics and clinical characteristics between groups were assessed using independent sample t-tests for continuous variables and χ 2 for dichotomous data. Differences in neuropsychological test-scores were assessed using ANOVA, adjusting for sex and education.
To assess differences in [ 18 F]flortaucipir BP ND and R 1 between early-and late-onset AD groups, we used ANOVA with each of the lobar regions-of-interest (separately) as dependent variable. We repeated these analyses, but now excluding the atypical AD cases (n = 8), to investigate whether potential age-dependent results were not driven solely by these extreme phenotypes. To assess more fine-grained differences, we additionally performed voxel-wise contrasts in SPM12 for BP ND or R 1 between early-and late-onset AD. All analyses were adjusted for sex. Results from voxel-wise analyses are displayed at both more liberal (i.e., p < 0.001, uncorrected) and more stringent (p < 0.05, family wise error (FWE) corrected) thresholds. Next, we investigated whether associations between BP ND or R 1 and cognition differed between early-and late-onset AD. We performed linear regression analyses between BP ND or R 1 and neuropsychological tests (separate models), adjusted for sex and education, including the interaction term "age-at-onset (dichotomous)*BP ND or R 1 ".
Results were considered significant if p-values were ≤ 0.05 for regional analyses. We considered a p-value ≤ 0.10 significant for interaction terms [44]. Additionally, adjustment for multiple comparisons was performed using the Benjamini-Hochberg false discovery rate (FDR) method (indicated by p FDR ) [45]. Statistical analyses were performed using R software, version 4.0.2, and the "mice" package was used for imputation.

Results
Participant characteristics are presented in Table 1. There were no differences in level of education, sex distribution, and APOE ɛ4 carriership between early-and late-onset AD (all p > 0.05), also in the cognition subsample (sTable-1). Figure 1 and sFigure-2 show average BP ND and R 1 maps, respectively, to provide insight into their spatial distribution. On average, early-onset AD patients showed highest BP ND values in medial and lateral parietal, and lateral temporal cortical areas (Fig. 1). Late-onset AD patients showed highest BP ND values in lateral temporal regions, and PCA patients in parieto-occipital regions, with BP ND values exceeding all groups (Fig. 1). Lastly, bvAD patients showed highest BP ND values in lateral frontal and temporal regions, as well as lateral and medial parietal regions (Fig. 1). For R 1 , values and patterns on the parametric images were visually similar across groups. (sFig-2).

Early-and late-onset AD differences in R 1
By contrast, [ 18 F]flortaucipir R 1 was lower in late-onset AD compared to early-onset AD in the medial temporal lobe (R 1 0.66 ± 0.05 vs 0.69 ± 0.05, p < 0.05), but not in any of the neocortical regions (p > 0.05) (Fig. 3). Voxelwise analyses showed lower R 1 in the (medial) temporal lobe (surviving FWE-correction), and subtly lower R 1 in the medial frontal cortex in late-onset compared to early-onset AD (Fig. 3G). In contrast, parieto-occipital regions showed lower R 1 in early-onset AD compared to late-onset AD, but this did not survive FWE-correction. Results remained essentially unchanged when atypical cases were excluded from the analysis (sFig-5). Results from partial volume corrected data yielded highly similar results, although R 1 values were slightly higher (sFig-6).

Early-and late-onset AD differences in the association between BP ND and cognition
Linear regression analyses showed that, in general, higher [ 18 F]flortaucipir BP ND was strongly associated with worse scores on a variety of cognitive tests (Fig. 4A). Age-at-onset moderated these associations such that in early-onset AD, associations between lateral temporal, parietal, and occipital BP ND with TMT-A, and parietal BP ND with TMT-B were stronger than in late-onset AD (all p interaction < 0.10: Fig. 4). In late-onset AD, associations between medial temporal BP ND and Digit Span backward and Letter fluency test (D-A-T) were stronger than in early-onset AD (Fig. 4). Results from partial volume corrected data were essentially comparable (sFig-7).

Early-and late-onset AD differences in the association between R 1 and cognition
In general, lower [ 18 F]flortaucipir R 1 was associated with worse scores on multiple cognitive tests (Fig. 5). Age-at-onset moderated these associations such that in early-onset AD, associations were stronger than in late-onset AD (all p interaction < 0.10: Fig. 5). More specifically, this was the case for the association between lateral temporal R 1 and Digit Span backward, TMT-A, TMT-B, and Letter fluency test (D-A-T), and the associations between occipito-parietal R 1 and TMT-A, TMT-B, and Category fluency Animals, as well as the association between occipital R 1 and MMSE (Fig. 5). Results from partial volume corrected data yielded some (test-specific) differences compared to the results from the uncorrected PET data (sFig-8).

Discussion
The main findings of this study include higher levels of neocortical tau pathology in early-onset AD compared to late-onset AD, while late-onset AD showed reduced rCBF in the medial temporal lobe compared to early-onset AD. Moreover, we found that higher levels of tau pathology and lower cerebral blood flow in lateral temporal and occipitoparietal regions were more strongly associated with cognitive impairment (mainly executive functioning domain) in early-vs late-onset AD. Previous studies demonstrated that relative to late-onset AD, early-onset AD exhibits more extensive pathological and neurodegenerative changes with respect to (among others) amyloid, metabolic activity [7,46], atrophy [47,48], functional network changes [49][50][51], and tau [52,53]. The current study extends on these previous studies not only showing greater tau load in early-onset AD (EOAD) and late-onset AD (LOAD). Indicated by the black lines on the color scale are thresholds for p < 0.001, uncorrected (t = 3.20) and for p < 0.05, FWE-corrected (t = 5.24). *p < 0.05; **p < 0.01; ***p < 0.001; †p FDR < 0.05 early-vs late-onset AD, but also providing novel insights into age-dependent differences in cerebral blood flow and demonstrates differential associations of both biomarkers with cognition. These findings contribute to our understanding of the differences between early-and late-onset AD patients and may support that younger AD patients are more suitable for inclusion in clinical trials, as the stronger link between pathology and cognitive decline suggests that a greater benefit may be achieved in this population when targeting tau pathology or cerebral blood flow compared to older AD patients, where this link is less strong. Besides, our results emphasize the importance of cerebrovascular  and cognitive test scores for early-and late-onset AD. A Significant modification of age-at-onset as assessed in the model including all AD subjects from the cognition subsample (n = 72), adjusted for age, sex, and education is depicted in grey (representing interaction terms at p ≤ 0.10) and black (representing interaction terms at p FDR ≤ 0.05) squares. Standardized regression coefficients are depicted for early-and late-onset AD separately. B Selection of scatterplots for the association between [ 18 F]flortaucipir R 1 and neuropsychological test scores health (and its potential treatment) in younger AD patients specifically.
For tau pathology, associations with worse cognitive performance across non-memory domains were stronger in early-onset AD, apart from associations in the medial temporal lobe. The medial temporal region is predominantly involved in sporadic (late-onset) AD [16,51,54], and in line with this, associations with cognitive performance for this region were stronger in late-onset AD. For rCBF, we found that lower blood flow was associated with a higher degree of cognitive impairment in non-memory domains in early-vs late-onset AD. This indicates that higher levels of neocortical tau pathology and lower rCBF have a relatively stronger influence on cognitive (dys)functioning in early-onset AD. The main hypothesis for these findings is that co-pathologies (e.g., other proteinopathies such as TDP-43 or alpha-synuclein, or vascular damage) often developing at older age, may contribute to progressive cognitive impairment in late-onset AD [55][56][57]. However, a recent study showed that co-pathologies also play an important role in the clinical phenotype of early-onset AD [26]. Another explanation could be that the brain regions most heavily affected in early-onset AD are more important for broader cognitive functioning (other than memory-specific), and comprise specific neuronal networks crucial for specific cognitive functions. The parietal lobe contains a high pathologic burden in early-onset AD (as shown by the result of the current study and by others (e.g., [7,51,53])), and is an important hub in higherorder cognitive networks [58]. If networks like these, or their cross-network relationships, segregate with older age [59][60][61], it might be that the relative importance of the regions affected by either tau pathology or decreases in rCBF decrease with older age. This would be reflected by a weakened association between tau pathology or rCBF and cognition (executive functioning specifically) in latecompared to early-onset AD.
In the current study, we used R 1 images as a measure of cerebral blood flow. This measure is tightly linked to hypometabolism measured with FDG-PET [19,24], and others have shown decreased occipitoparietal glucose metabolism in early-onset AD [7,62]. In contrast, our regional analyses did not show lower rCBF in early-vs late-onset AD. However, additional voxel-wise comparisons showed more fine-grained decreased rCBF in lateral parietal and occipital regions in early-onset AD. One explanation for these findings might be that there are no extensive or clear differential patterns of cortical reductions in rCBF in early-vs late-onset AD, or that differential reductions in rCBF may be restricted to specific cortical gyri. Another explanation could be that R 1 , which is a proxy of rCBF, is not sensitive enough to capture the differences between early-and late-onset AD. As rCBF decreases with age [63], this might suggest that early-onset AD patients are more severely affected by deficits in rCBF compared to late-onset AD patients, which in turn might be caused (in part) by the higher levels of tau pathology present. Ideally, one would investigate whether early-onset AD patients are indeed more heavily affected by reduced rCBF by including age-matched control groups, but since we had no such data available, it remains to be elucidated in future research.
Our results showed that early-onset AD patients exhibit higher levels of tau pathology in widespread neocortical regions compared to late-onset AD patients, but no differences were found in the medial temporal lobe. Others found higher levels of tau pathology primarily in (pre)frontal and (inferior) parietal cortices in early-relative to late-onset AD, and no differences in the medial temporal cortex [53].
Another study similarly showed that early-onset AD patients showed greater binding in the inferior parietal, occipital, and inferior temporal cortices [64]. The pattern of cortical involvement as found in the current study thus is highly consistent with previous findings, indicating that the development of high levels of tau pathology follows a specific spatial pattern within early-onset AD patients. Supporting this, a recent study identified four distinct spatiotemporal trajectories of tau pathology in AD, each presenting with distinct demographic and cognitive profiles and differing longitudinal outcomes. The medial temporal lobe-sparing subtype, for example, was associated with younger age, less APOE4 allele carriership, and greater overall tau burden [65], which is largely in accordance with findings in our early-onset AD study population. Others point toward genetics explaining higher tau burden, since comparable results were found in autosomal dominant mutation carriers (e.g., presenilin-1) [66,67]. The differential spatial patterns of tau pathology in early-onset AD might also be explained by genetic involvement, as early-onset AD patients are less frequently APOE4 allele carriers, and E4 genotype does influence spatial patterns of brain pathology [68][69][70]. In the present study, however, we did find slightly but non-significantly lower E4 allele carriership in early-onset AD, which makes it unlikely that E4 genotype may have explained the findings.
A strength of this study is usage of a single dynamic [ 18 F]flortaucipir PET scan to derive measures of both tau pathology (BP ND ) and relative cerebral blood flow (R 1 ). Furthermore, we repeated analyses with partial volumecorrected data. Although we did find some (test-specific) differences in the association between R 1 and cognition, core results that led to our conclusions remained. Given that the implications of applying partial volume-correction to R 1 data are currently not entirely understood, and that all other results remained essentially comparable, we feel atrophy does not bias our findings to a large extent. There were also several limitations. First, the late-onset AD patients in our study were relatively young, which might hamper the generalizability of results to older study populations. Second, seven participants did not have sufficient neuropsychological data available for cognition analyses. Although results are not expected to be influenced to a large extent, it might be that results as described in this study are slightly underestimating the differential effect of age-at-onset, given that neuropsychological data were missing in the most severely progressed patients (mean MMSE score: 20 ± 4). Third, the cross-sectional design did not allow investigating the temporal evolution of tau pathology, rCBF, and cognitive impairment. Future studies with longitudinal data available could aid in elucidating the timely pathways of these three parameters.
In conclusion, early-onset AD is characterized by higher levels of tau pathology and stronger associations between lateral temporal and occipitoparietal tau pathology or lower rCBF and cognitive impairment. These findings may have important implications for clinical trials, since effects of potential tau-or blood flow targeting therapeutic interventions might exert larger effects in early-onset AD compared to late-onset AD patients. Furthermore, differences between early-and late-onset AD patients as described in the current study should be considered in any therapeutic intervention trial where either cognition is used as an outcome measure or where indirect effects on tau pathology or rCBF are expected, to ensure correct interpretation of results and aid in the formation of appropriate patient selection criteria.