Association of Midlife Depressive Symptoms with Regional Amyloid-β and Tau in the Framingham Heart Study

Abstract

Background:

Depressive symptoms predict increased risk for dementia decades before the emergence of cognitive symptoms. Studies in older adults provide preliminary evidence for an association between depressive symptoms and amyloid-β (Aβ) and tau accumulation. It is unknown if similar alterations are observed in midlife when preventive strategies may be most effective.

Objective:

The study aim was to evaluate the association between depressive symptoms and cerebral Aβ and tau in a predominately middle-aged cohort with examination of the apolipoprotein (APOE) ε4 allele as a moderator.

Methods:

Participants included 201 adults (mean age 53±8 years) who underwent ¹¹C-Pittsburgh Compound B amyloid and ¹⁸F-Flortaucipir tau positron emission tomography (PET) imaging. Depressive symptoms were evaluated with the Center for Epidemiological Studies Depression Scale (CES-D) at the time of PET imaging, as well as eight years prior. Associations between depressive symptoms at both timepoints, as well as depression (CES-D≥ 16), with regional Aβ and tau PET retention were evaluated with linear regression adjusting for age and sex. Interactions with the APOE ε4 allele were explored.

Results:

Depressive symptoms and depression were not associated with PET outcomes in the overall sample. However, among APOE ε4 allele carriers, there was a significant cross-sectional association between depressive symptoms and increased tau PET uptake in the entorhinal cortex (β = 0.446, SE = 0.155, p = 0.006) and amygdala (β = 0.350, SE = 0.133, p = 0.012).

Conclusion:

Although longitudinal studies are necessary, the results suggest that APOE ε4 carriers with depressive symptoms may present with higher susceptibility to early tau accumulation in regions integral to affective regulation and memory consolidation.

INTRODUCTION

Late-life depression, a prevalent and potentially treatable psychiatric condition that manifests after age 60, has been associated with a four-fold higher risk of incident Alzheimer’s disease and related dementias (ADRDs) [1]. Late life depressive symptoms, even below the threshold for a clinical diagnosis, have been associated with cognitive decline and accelerated cerebral atrophy [2, 3]. Depression that develops earlier in the lifespan has also been associated with increased dementia risk [4]. Notably, depressive symptoms have been found to predict incident ADRD up to 25 years before overt cognitive symptoms emerge [5], suggesting that neuropsychiatric symptoms may be an etiological risk factor or an early manifestation of neurodegenerative disease. Enhanced understanding of the neurobiological mechanisms linking depressive symptoms with ADRDs in the prodromal stage may advance the development of novel prevention strategies [6].

Elevations in amyloid-β (Aβ), which can emerge decades prior to dementia diagnosis, are a harbinger for progressive cognitive decline and incident AD [7]. Support for a potential pathogenic link between Aβ deposition and depression is derived from postmortem studies of AD reporting a higher burden of hippocampal Aβ plaques and neurofibrillary tangles in individuals with a lifetime history of depression [8]. Major depression and depressive symptoms in late-life have also been associated with Aβ levels assessed in plasma, cerebrospinal fluid (CSF), and positron emission tomography (PET) imaging [9, 10]. Moreover, multiple cohort studies of older adults have reported that higher cerebral Aβ levels predict increased depressive symptomatology over time [11–13]. Nonetheless, not all studies have reported an association between depression and Aβ levels [9, 14]. A recent investigation from the ADNI Depression study reported that older adults with late-life depression had poorer memory performance, yet lower cerebral Aβ levels than age-matched non-depressed controls [15]. The findings suggest that the association between depression and cognitive decline may be governed, at least in part, by mechanisms outside of the Aβ pathway [14].

As compared to Aβ, tau deposition more strongly correlates with phenotypic ADRD features such as cerebral hypometabolism and cognitive decline [16–19]. As such, tau may also be more closely coupled with neuropsychiatric symptoms. Contrary to this hypothesis, a meta-analytic study failed to detect an association between depression diagnosis and CSF levels of total and phosphorylated tau [20]. To date, only a few studies have examined the association between depression and tau with PET imaging. In the Harvard Aging Brain Study, increased depressive symptoms were related to higher tau PET deposition in the inferior temporal lobe and entorhinal cortex within cognitively unimpaired older adults [21]. A recent study by Babulal et al. reported that cognitively intact older adults with elevated tau PET retention in a composite region comprised of the amygdala, inferior temporal lobe, entorhinal cortex, and the lateral occipital lobe were two times more likely to meet diagnostic criteria for depression [14].

While support exists for a mechanistic link between depression and Aβ and tau accumulation, the findings have been inconsistent across studies. Thus, further research is needed, particularly leveraging PET imaging to better characterize spatial patterns of deposition. The primary aim of the current study was to examine the association of depressive symptoms, as well as the presence of more significant symptoms meeting criteria for depression, with regional Aβ and tau PET retention. While prior literature has focused on older adults [11, 14, 21], the present study examined a predominately middle-aged sample. As ADRDs have a protracted pre-symptomatic period that extends across decades [7], examinations conducted at midlife may shed insight on the initial pathophysiological changes and facilitate preventive efforts. In addition to examining cross-sectional relationships, we also evaluated the association between change in depressive symptomatology over time with PET imaging outcomes. Furthermore, as apolipoprotein (APOE) ε4 carriage has been linked with neuropathological protein accumulation [12, 22, 23] and increased risk of depression in some cohorts, we examined interactions with APOE ε4 carrier status. Based on previous studies conducted in older adult samples [11, 14, 21], we hypothesized that regional Aβ and tau PET retention levels would be associated with depressive symptoms at baseline and change over time. We further hypothesized a moderating effect of APOE with stronger associations in ε4 carriers relative to non-carriers. Finally, we conducted sensitivity analyses to examine if our findings remain consistent with adjustment for antidepressant usage. In animal models, selective serotonin reuptake inhibitors (SSRIs) have been associated with lower amyloid beta burden with some converging evidence in human studies [24, 25]. However, population-based research has produced conflicting results with antidepressant use being linked to both increased and decreased risk of incident dementia [26–28]. Therefore, additional research exploring the potentially modifying role of antidepressant usage on PET imaging ing outcomes is necessary.

MATERIALS AND METHODS

Participants

The Framingham Heart Study (FHS) was initiated in 1948 with enrollment of the Original Cohort, followed by the enrollment of the Offspring cohort (children of the Original Cohort and their spouses) in 1971 and enrollment of the Third Generation cohort (grandchildren of the Original Cohort and children of the Offspring Cohort) in 2002 [29]. The Third Generation cohort completed their first examination between 2003–2005, which was followed by their second examination between 2008 to 2011, and their third examination between 2016 to 2019. At the third examination, a subset of the Third Generation cohort completed their first amyloid and tau PET scans for the study. These individuals were included in the current analyses.

Eligibility for the PET imaging sub-study, as well as for inclusion in the current analyses, included age 32 to 75 years at the time of the third examination, prior completion of FHS examination two, and absence of significant neurological conditions including clinical stroke, dementia, and multiple sclerosis. As previously described, FHS participants undergo routine cognitive screening and comprehensive monitoring for continual surveillance of dementia [30], which was exclusionary for participation in the PET imaging sub-study. Depressive symptoms were assessed at examinations 2 and 3. All participants provided written informed consent at the time of enrollment. The study was approved by the Institutional Review Board at Boston University Medical Center and the Massachusetts General Hospital.

Assessment of depressive symptoms

Depressive symptoms were evaluated using the Center for Epidemiological Studies Depression Scale (CES-D) at the second and third examinations [31]. The CES-D is a 20-item self-report questionnaire with response options ranging from 0 to 3 for each item (0 = Rarely or None of the Time, 1 = Some or Little of the Time, 2 = Moderately or Much of the time, 3 = Most or Almost All the Time). The total score range is between 0 to 60 points with higher scores indicating greater depressive symptomatology. In accordance with previous studies, a CES-D score of 16 or greater was utilized as the cut-off for depression [32], which has been shown to have strong convergent validity (100% sensitivity, 88% specificity) with diagnostic interviews [33].

Antidepressant use

The use of antidepressant medication (yes or no) was determined by self-report at examination three using the Anatomical Therapeutic Chemical (ATC) classification N06A.

APOE genotype

Leukocyte DNA was acquired from 5–10 ml of whole blood and APOE genotype was performed with PCR as previously described [34]. Participants were classified as APOE ε4 carriers if they had at least one copy of the ε4 allele.

¹¹C-Pittsburgh Compound B (PiB) Aβ and ¹⁸F-Flortaucipir (FTP) tau PET imaging

PiB Aβ and FTP tau PET imaging were conducted on a Siemens ECAT HR + scanner (3D mode; 63 image planes; 15.2 cm axial field of view; 5.6 mm transaxial resolution; and 2.4 mm slice interval), following previously published methodology [21]. Briefly, PiB PET images were acquired with a 10 to 15 mCi bolus injection followed by a 60 min dynamic acquisition. FTP PET images were obtained across 80 to 100 min using 4 × 5 min frames after a single 9 to 11 mCI bolus injection. PiB and FTP images were co-registered to a structural T1-weighted brain MRI using SPM8. FreeSurfer v6.0 was used to derive regions of interest (ROIs) [35]. PiB retention was expressed as the distribution volume ratio (DVR) using the cerebellar cortex as a reference region [36]. FTP retention was expressed as the standardized uptake value ratio (SUVr) using the cerebellar cortex as a reference. PET data were evaluated without partial volume correction given the relatively young age of the sample with minimal atrophy. A PiB summary measure, frontal, lateral, and retrosplenial (FLR), was derived from the mean of superior frontal, inferior frontal, rostral middle frontal, rostral anterior cingulate, medial orbitofrontal, inferior and middle temporal, inferior parietal, and precuneus regions [36]. Regional precuneus PiB was also examined given the region’s susceptibility to early Aβ accumulation [37]. FTP retention was evaluated in regions with vulnerability to early tau deposition including the entorhinal cortex, inferior temporal lobe, amygdala, precuneus, parahippocampus, and hippocampus [38]. An FTP summary measure was derived from the mean of fusiform gyrus, entorhinal cortex, and inferior and middle temporal regions [39]. In addition, PiB and FTP PET retention were also examined in two regions widely implicated in depression, the anterior cingulate and the middle frontal gyms [2, 40]. Across ROIs, values from the left and right hemisphere were averaged.

Statistical analysis

Data normality was visually ascertained. CES-D scores at examinations 2 and 3 and Aβ PET retention in the FLR and precuneus were skewed and were natural log transformed to normalize their distributions. All PET variables were standardized prior to analyses. We calculated annualized change in depressive symptoms as change in raw CES-D score from examination 2 to 3 divided by the time between exams. APOE ε4 carrier and non-carrier group diferences in demographic and clinical characteristics were assessed with independent samples t-tests or Mann-Whitney U tests for continuous variables and the chi-squared statistic for categorical variables. Associations between depressive symptoms, annualized change in depressive symptoms, and depression at examination 3 with regional Aβ and tau PET retention were evaluated with linear regression models adjusting for age and sex. Interactions between APOE ε4 status with depressive symptoms and annualized change in depressive symptoms for the PET imaging outcomes were evaluated with linear regression with covariates for age and sex. For outcomes with a significant APOE ε4 interaction, stratified analyses by APOE ε4 carrier status were performed using linear regression adjusted for age and sex. Due to the negligible number of APOE ε4 carriers meeting criteria for depression (N = 2), interactions with APOE ε4 were not examined with depression at examination 3. Given the possible confound of antidepressant medication usage, sensitivity analyses were performed for all regression models examining main effects and interactions with APOE ε4 for PET imaging outcomes with inclusion of an additional covariate for antidepressant use (yes or no). Statistical tests were 2-sided and the criterion for significance was set at p-value of < 0.05. Regression analyses examining main and interactive effects were FDR-corrected for multiple comparisons and the adjusted p-values are reported. Analyses were performed using SAS version 9.4.

RESULTS

Participant characteristics

The study sample included 201 participants, mean age 53 ± 8 years, of which 47% were female (Table 1). Seven individuals were missing data on APOE genotype and were not included in the stratified analyses. Depressive symptoms were assessed with the CES-D at examinations two and three, which occurred a mean of 7.8 years apart. At examination 3, the median CES-D score was 5 (interquartile range 1, 8). Twelve percent (N = 25) of the sample met criteria for depression and 18% (N = 37) of the total sample was on at least one antidepressant medication. Approximately a quarter of the sample were APOE ε4 carriers. There were no significant differences in demographic and clinical characteristics between the APOE ε4 carrier and non-carrier groups.

Table 1.

Demographics and clinical characteristics of the sample by apolipoprotein E (APOE) ε4 carrier status

	Total Sample (N = 201)	APOE ε4 Non-carriers (N = 147)	APOE ε4 Carriers (N = 47)	p
Age, y	53.4 ± 8.0	53.5 ± 7.8	53.3 ± 8.8	0.887a
Female, N (%)	94 (47%)	73 (50%)	19 (40%)	0.270c
Education, N (%)
High School Degree	22 (11%)	18 (12%)	4 (9%)	0.699c
Some College	57 (29%)	44 (30%)	13 (28%)
College Degree	115 (59%)	85 (58%)	30 (64%)
Depressive symptoms (CES-D) at Examination 2, Median (Q1-Q3)	3 (1–8)	3 (1–9)	3 (1–7)	0.445b
Depression (CES-D ≥ 16) at Examination 2, N (%)	7 (4%)	6 (5%)	0	-
Depressive Symptoms (CES-D) at Examination 3, Median (Q1-Q3)	5 (2–9)	5 (2–10)	5 (1–9)	0.320b
Depression (CES-D ≥ 16) at Examination 3, N (%)	25 (12%)	22 (15%)	2 (4%)	-
Time Between Examinations 2 and 3, y	7.8 ± 0.4	7.8 ± 0.4	7.8 ± 0.4	0.778a
Annualized Change in CES-D Between Examinations 2 and 3, Median (Q1-Q3)	0.1 (−0.1–0.5)	0.1 (−0.1–0.5)	0.0 (−0.1–0.5)	0.492b
Antidepressant Use at Exam 3	37 (18%)	27 (18%)	8 (17%)	0.763c
Aβ PET Composite (FLR)†	1.40 ± 0.02	1.40 ± 0.02	1.40 ± 0.02	0.135a
Tau PET Composite (Temporal Lobe)	1.09 ± 0.06	1.09 ± 0.06	1.11 ± 0.06	0.072a

Group differences were assessed with independent t-tests ^a or Mann-Whitney U tests^b for continuous variables and the chi-squared statistic^c for categorical variables.

^†Natural log transformed.

Values represent mean ± standard deviationunless otherwise noted. APOE, apolipoprotein E; CES-D, Center for Epidemiological Studies Depression Scale; Aβ, amyloid-β; FLR, frontal, lateral and retrosplenial.

Cross-sectional associations between depressive symptoms, depression, and regional Aβ and tau PET retention

Depressive symptoms, evaluated at examination 3 using the CES-D, were not associated with regional Aβ or tau PET (Table 2). However, there was a significant interaction between depressive symptoms and the APOE ε4 allele in their effect on tau PET retention in the entorhinal cortex (interaction effect size = 0.522), amygdala (interaction effect size = 0.507), and hippocampus (interaction effect size = 0.453). Stratified analyses indicated significant associations between depressive symptoms and higher regional tau PET retention in the entorhinal cortex and amygdala in the APOE ε4 carrier group (Table 3). In contrast, the APOE ε4 non-carrier group evidenced a significant association between depressive symptoms and lower hippocampal tau PET retention

Table 2.

Associations of depressive symptoms with standardized regional amyloid-β and tau positron emission tomography (PET) retention

	Depressive Symptoms (CES-D†)a		Interaction Between Depressive Symptoms (CES-D†) and APOE ε4b
	β (SE)	p	p
FLR Aβ †	−0.031 ± 0.068	0.836	0.223
Precuneus Aβ†	−0.017 ± 0.071	0.885	0.118
Anterior Cingulate Aβ	−0.033 ± 0.071	0.836	0.320
Middle Frontal Gyrus Aβ	−0.082 ± 0.068	0.683	0.394
Entorhinal Tau	0.032 ± 0.074	0.836	0.025*
Inferior Temporal Tau	0.039 ± 0.072	0.836	0.230
Amygdala Tau	−0.046 ± 0.073	0.836	0.025*
Precuneus Tau	−0.127 ± 0.074	0.683	0.637
Parahippocampus Tau	−0.029 ± 0.075	0.836	0.215
Hippocampus Tau	−0.067 ± 0.056	0.683	0.046*
Anterior Cingulate Tau	−0.101 ± 0.074	0.683	0.223
Middle Frontal Gyrus Tau	0.003 ± 0.070	0.968	0.478

^†Natural log transformed.

^aβ ± SE and FDR-corrected p-values derived from linear regression models examining the associations between depressive symptoms (natural log transformed Center for Epidemiological Studies Depression Scale score) with regional Aβ and tau PET retention, adjusting for age and sex.

^bFDR-corrected p-values derived from linear regression models examining the interaction between depressive symptoms (natural log transformed Center for Epidemiological Studies Depression Scale) and APOE ε4 for regional Aβ and tau PET retention with adjustment for age, sex, and APOE ε4.

CES-D, Center for Epidemiological Studies Depression Scale; APOE, apolipoprotein E; FLR, frontal, lateral and retrosplenial; Aβ, amyloid-β.

Table 3.

Associations of depressive symptoms with standardized regional tau positron emission tomography (PET) retention stratified by APOE ε4 carrier status

	Depressive Symptoms (CES-D†)a
	β±SE	p
Tau entorhinal
APOE ε4 positive	0.446 ± 0.155	0.006*
APOE ε4 negative	−0.076 ± 0.083	0.364
Amygdala Tau
APOE ε4 positive	0.350 ± 0.133	0.012*
APOE ε4 negative	−0.157 ± 0.088	0.078
Tau hippocampus
APOE ε4 positive	0.244 ± 0.131	0.070
APOE ε4 negative	−0.209 ± 0.090	0.022*

^aβ±SE and FDR-corrected p-values derived from linear regression models stratified by APOE ε4 carrier status examining the associations between depressive symptoms (natural log transformed Center for Epidemiological Studies Depression Scale score) with regional tau PET retention, adjusting for age and sex.

CES-D, Center for Epidemiological Studies Depression Scale; APOE, apolipoprotein E.

Depression, as defined by CES-D ≥ 16, was not associated with regional Aβ or tau PET retention (Table 4).

Table 4.

Associations of depression with standardized regional amyloid-β and tau positron emission tomography (PET) retention

	Depression (CES-D score ≥16)a
	β ± SE	p
FLR Aβ†	0.183 ± 0.197	0.664
Precuneus Aβ†	0.241 ± 0.205	0.664
Anterior Cingulate Aβ	0.042 ± 0.207	0.839
Middle Frontal Gyrus Aβ	0.075 ± 0.197	0.839
Entorhinal Tau	−0.133 ± 0.213	0.721
Inferior Temporal Tau	0.128 ± 0.209	0.721
Amygdala Tau	−0.219 ± 0.211	0.664
Precuneus Tau	−0.249 ± 0.215	0.664
Parahippocampus Tau	−0.187 ± 0.216	0.664
Hippocampus Tau	−0.052 ± 0.164	0.664
Anterior Cingulate Tau	−0.238 ± 0.214	0.664
Middle Frontal Gyrus Tau	0.059 ± 0.201	0.839

^†Natural log transformed for normality.

^aβ ± SE and FDR-corrected p-values derived from linear regression models examining the association between depression, as defined by Center for Epidemiological Studies Depression Scale score >16, with regional Aβ and tau PET retention, adjusting for age and sex.

CES-D, Center for Epidemiological Studies Depression Scale; FLR, frontal, lateral and retrosplenial; Aβ, amyloid-β.

Association of annualized change in depressive symptoms with regional Aβ and tau PET retention

Annualized change in depressive symptoms across examinations 2 and 3 was not associated with regional Aβ or tau PET retention (Table 5). There were no significant interactions with APOE ε4 carrier status.

Table 5.

Associations of annualized change in depressive symptoms and regional amyloid-β and tau positron emission tomography (PET) retention derived from linear regression

	Annualized Change in Depressive Symptoms (CES-D†)a		Interaction of Annualized Change in Depressive Symptoms (CES-D) × APOE ε4b
	β ± SE	p	p
FLR Aβ†	0.051 ± 0.070	0.947	0.903
Precuneus Aβ†	0.066 ± 0.073	0.947	0.903
Anterior Cingulate Aβ	0.081 ± 0.073	0.947	0.903
Middle Frontal Gyrus Aβ	0.050 ± 0.070	0.947	0.903
Entorhinal Tau	−0.011 ± 0.076	0.979	0.217
Inferior Temporal Tau	0.023 ± 0.074	0.979	0.896
Amygdala Tau	0.006 ± 0.075	0.979	0.270
Precuneus Tau	−0.071 ± 0.076	0.947	0.903
Parahippocampus Tau	−0.006 ± 0.077	0.979	0.516
Hippocampus Tau	−0.051 ± 0.058	0.947	0.217
Anterior Cingulate Tau	−0.002 ± 0.076	0.979	0.743
Middle Frontal Gyrus Tau	0.027 ± 0.071	0.979	0.711

^†Natural log transformed for normality.

^aβ ± SE and FDR-corrected p-values derived from linear regression models examining the associations between annualized change in depressive symptoms (derived from Center for Epidemiological Studies Depression Scale raw scores) with regional Aβ and tau PET retention, adjusting for age and sex.

^bFDR-corrected p-values derived from linear regression models examining the interaction between annualized change in depressive symptoms (derived from Center for Epidemiological Studies Depression Scale raw scores) and APOE ε4 for regional Aβ and tau PET retention with adjustment for age, sex, and APOE ε4.

CES-D, Center for Epidemiological Studies Depression Scale; APOE, apolipoprotein E; FLR, frontal, lateral and retrosplenial; Aβ, amyloid-β.

Sensitivity analysis

With additional adjustment for antidepressant use, the main effects of the associations between depressive symptoms, annualized change in depressive symptoms, and depression with regional Aβ and tau PET retention remain unchanged (Supplementary Tables 1, 2, and 3). Additionally, the interactions between APOE ε4 with both depressive symptoms and annualized change in depressive symptoms remain consistent. For stratified analyses by APOE ε4 groups (Supplementary Table 4), the associations between depressive symptoms with regional entorhinal and amygdala tau PET retention in APOE ε4 carriers persist. However, the association between depressive symptoms and hippocampal tau PET retention in APOE ε4 non-carriers was no longer significant.

DISCUSSION

The current study examined associations between depressive symptoms and regional Aβ and tau PET retention in a predominately middle-aged cohort. In cross-sectional analyses, depressive symptoms and depression were not associated with PET imaging outcomes. Additionally, annualized change in depressive symptoms across a mean of 8 years prior to PET imaging was not associated with Aβ or tau PET retention. However, there was a significant interaction between depressive symptoms and APOE ε4 carriage in their effect on regional tau PET deposition in the entorhinal cortex, amygdala, and hippocampus. Within the hippocampus, depressive symptoms were associated with lower tau PET retention in APOE ε4 non-carriers. However, the association was no longer significant with adjustment for antidepressant use. Within the entorhinal cortex and the amygdala, depressive symptoms were associated with higher tau PET retention only within APOE ε4 carriers. The results suggest that individuals with depressive symptoms and higher genetic risk for AD due to the APOE ε4 allele may represent a subgroup at increased vulnerability for neuropathological accumulation in regions susceptible to early ADRD pathogenesis [38]. Longitudinal studies will be necessary to evaluate if higher entorhinal cortex and amygdala tau PET retention confers risk for accelerated cognitive decline and incident dementia in individuals with depressive symptoms.

In our study, depressive symptoms were only associated with higher entorhinal and amygdala tau PET retention in individuals with at least one copy of the APOE ε4 allele. The APOE ε4 allele is the most robust known genetic risk factor for sporadic AD with a single copy of the allele increasing the relative risk of AD by 2–4 fold [41]. The APOE ε4 allele contributes to increased production and reduced clearance of the Aβ protein [42], resulting in higher rates of Aβ PET positivity at earlier ages of onset compared to the ε2 and ε3 alleles [23]. The APOE ε4 allele also propagates tau hyperphosphorylation and APOE ε4 carriage is associated with increased tau PET retention independent of Aβ [22, 42, 43]. Therefore, in our predominately middle-aged sample, higher tau PET retention may have only been detectable in APOE ε4 carriers, who typically display elevations in Aβ and tau at earlier timepoints [42]. Interestingly, some prior studies of older adults have reported that APOE ε4 carriage is associated with increased likelihood of depression particularly in the presence of Aβ PET positivity [12]. Additionally, within individuals with late-life depression, the APOE ε4 allele has been associated with hippocampal, frontal, and occipital lobe atrophy [44]. A longitudinal study by Qiu et al. reported that older adults with depression and elevated plasma Aβ 40/42 ratios were at increased risk for incident AD with a modifying effect of APOE ε4 [45]. Specifically, APOE ε4 carriers with depression and elevated plasma Aβ levels had a 40% increased risk of AD as compared to only 4% in non-carriers. Overall, the results suggest that depressive symptoms, in the context of the APOE ε4 allele, may confer risk for incident dementia. As compared to prior research demonstrating associations with Aβ [12, 45], our study identified an interaction between depressive symptoms and APOE ε4 carriage for tau PET retention. Additional studies will be necessary to determine if this finding replicates in other cohorts.

Within our study, APOE ε4 carriers with higher depressive symptoms displayed increased tau PET retention in the entorhinal cortex and the amygdala. Neuropathological and more recently, PET imaging studies have indicated that the entorhinal cortex and amygdala are early sites of tau accumulation in both normal aging and AD [16, 38, 43]. The medial temporal lobe has a fundamental role in memory consolidation [46]. In older adults without dementia, subtle elevations in entorhinal tau PET retention have been associated with poorer memory performance independent of A [47, 48]. The APOE ε4 allele has been found tomoderate this associationwith stronger effects observed in ε4 carriers relative to non-carriers [49]. The entorhinal cortex has reciprocal connections with the amygdala [50], a region governing affective modulation. As the amygdala has diffuse connections with brainstem, allocortical, and neocortical regions, it is hypothesized to play a pivotal role in the dispersion of neuropathological proteins, potentially harbingering a progressive downward cognitive trajectory [51]. The entorhinal cortex and amygdala also comprise an affective network and their dysregulation is implicated in the pathogenesis of depressive symptoms [52]. Prior neuroimaging studies of depression have reported accelerated atrophy and altered functional connectivity in the entorhinal cortex and the amygdala [53, 54]. In older adults without dementia, depressive symptoms and repetitive negative thoughts have been associated with elevated entorhinal tau PET retention [55]. A recent study reported that higher neuroticism, a personality trait that increases vulnerability to depression [56], was associated with increased entorhinal cortex, amygdala, and inferior temporal lobe tau PET retention in older adults [57]. The current study advances the literature by indicating that depressive symptoms in the presence of the APOE ε4 allelemay be associated with increased regional temporal and limbic tau PET retention as early as midlife.

The mechanisms that may link depressive symptoms with higher entorhinal cortex and amygdala tau PET retention in APOE ε4 carriers are not fully understood. Tau accumulation in regions critical to affective regulation may precipitate depressive symptoms [21]. However, it is important to note that annualized change in depressive symptoms did not have significant main or interactive effects with Aβ or tau PET retention. Depressive symptoms may propagate tau accumulation through dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis [6]. Depression has been linked with impaired negative feedback of the HPA axis and elevated glucocorticoid levels [58]. In animal models, glucocorticoids have been found to accelerate tau pathology [59]. Depressive symptoms have also been associated with higher inflammatory cytokine levels in plasma and CSF [60, 61]. A proteomic study reported that inflammatory processes may be a causal pathway underlying cognitive decline in late-life depression [62]. Inflammatory processes have also been found to accelerate tau phosphorylation [63]. Thus, inflammation may be a mechanism linking depressive symptoms and tau accumulation in individuals with higher genetic susceptibility to AD due to the APOE ε4 allele.

In addition to our findings of higher entorhinal cortex and amygdala tau PET retention in association with depressive symptoms in APOE ε4 carriers, we also report an association between depressive symptoms and lower hippocampal tau PET retention in APOE ε4 non-carriers. However, the finding was attenuated in sensitivity analyses adjusting for antidepressant use. Observational studies have produced conflicting findings with evidence of both increased [26] and lowered incident dementia risk [27, 28] associated with antidepressant use. A meta-analytical study reported that antidepressant use was associated with more than a two-fold increase in risk for incident cognitive impairment or dementia with stronger effects for those who initiated antidepressant use before the age of 65 [64]. In a sample of cognitively unimpaired older adults, individuals taking antidepressant medication were more likely to present with elevated tau on PET imaging relative to individuals not taking antidepressant medication [14]. Moreover, there was an interaction between elevated tau PET retention and antidepressant use for increased likelihood of depression. Longitudinal studies will be necessary to evaluate if antidepressant usage is associated with higher tau accumulation over and above the effects of depressive symptoms alone.

In our study, neither depressive symptoms nor depression were associated with Aβ PET retention. Depressive symptoms have been inconsistently linked with Aβ levels in plasma, CSF, and PET imaging [9, 12, 15]. Population-based studies have provided stronger converging evidence of a longitudinal association of Aβ PET positivity and increased depressive symptomatology over time [11–13]. Similar to our findings, two cross-sectional studies conducted in cognitively unimpaired older adults detected significant associations between depressive symptoms and tau PET retention without concomitant changes in Aβ [14, 21]. The evidence suggests that depressive symptoms may be more tightly coupled with tau than Aβ deposition even in individuals without overt cognitive symptoms.

Our study had several strengths including longitudinal assessments of depression, inclusion of Aβ and tau PET imaging, and a predominately middle-aged sample. Nonetheless, the results must also be considered in the context of the study limitations. While our largely middle-aged cohort provided a unique opportunity to examine early changes in neuropathological protein accumulation associated with depressive symptoms, Aβ and tau deposition at midlife is typically more restricted than within older age groups [65]. Attenuated ranges and regional distributions of Aβ and tau within our sample may have diminished power for detecting significant effects and regions included, such as the hippocampus, may be vulnerable to off-target binding effects [66]. Furthermore, the clinical significance of continuous ranges of Aβ and tau PET retention in middle-aged adults is unknown. However, elevated cerebral Aβ and tau have been found to predict accelerated cognitive decline even in late midlife [67]. Future longitudinal studies within our cohort will be essential for evaluating the clinical relevance of our findings. Our study examined depressive symptoms and depression in a population-based cohort. A meta-analysis examining depression prevalence using the CES-D within similar settings reported a median prevalence of 8.8% with a range of 3.8 to 12.6% [68], which is highly consistent with the rates observed in our study. While our population-based cohort enabled assessment of subsyndromal depressive symptoms that are widely prevalent, we had more limited power for evaluating associations with annualized change in depressive symptomatology and major depression, which may have contributed to the null findings. In addition, our study does not include formal assessment of depression diagnosis and is unable to disentangle the effects of different subtypes of depression, such as seasonal affective disorder and bipolar disorder. In addition, our study lacks data on anxiety symptomatology. which is frequently comorbid with depression and has been previously associated with Aβ PET imaging outcomes in older adults [39]. Finally, the association between depressive symptoms and PET imaging outcomes may be mediated or moderated by several medical and lifestyle factors such as physical activity engagement, sleep quality, and cardiovascular disease burden [69]. While outside of the scope of the current study, future research with larger sample sizes and longer follow-up periods will be critical for elucidating the causal pathways linking depressive symptoms with tau accumulation.

In summary, our study did not detect associations between depressive symptoms and regional Aβ and tau PET retention within a predominately middle-aged sample. However, depressive symptoms were significantly associated with higher entorhinal cortex and amygdala tau PET retention among APOE ε4 carriers. While future longitudinal studies are necessary, the results suggest that APOE ε4 carriers with depressive symptoms may present with higher susceptibility to early tau accumulation in regions integral to affective regulation and memory consolidation [47, 48, 51, 53]. As such, the results highlight the importance of ongoing monitoring of depressive symptoms within tertiary care settings. Finally, with further validation, the findings may have relevance for improving detection of individuals at elevated risk for tau accumulation in midlife, who may benefit from preventive trials.