Skip to main content
NCBI home page

This is a preprint.

It has not yet been peer reviewed by a journal.

The National Library of Medicine is running a pilot to include preprints that result from research funded by NIH in PMC and PubMed.

Research Square logoLink to Research Square
[Preprint]. 2026 Aug 14:rs.3.rs-10145165. [Version 1] doi: 10.21203/rs.3.rs-10145165/v1

Brain metabolic resilience conferred by the APOE3 Christchurch variant in autosomal dominant Alzheimer’s disease and across the adult lifespan

Yakeel Quiroz 1, Catarina Tristão-Pereira 2, Vincent Malotaux 3, Yesica Zuluaga 4, Claudia Muñoz 5, George Silva 6, Bing He 7, Diana Alzate 8, Yamile Bocanegra 9, Alejandro Guerrero 10, Laura Serna 11, Carolina Ospina 12, Monica Vidal 13, Sergio Alvarez 14, Rafael Posada-Duque 15, Joseph Arboleda-Velasquez 16, David Aguillon 17
PMCID: PMC13640913  PMID: 42845753

Abstract

The APOE3 Christchurch (APOE3Ch) is a rare genetic variant associated with marked protection against autosomal-dominant Alzheimer’s disease (ADAD) due to the PSEN1E280A variant, by delaying the onset of clinical symptoms. In this study, we investigate age-related alterations in cerebral glucose metabolism, measured with 18F-fluorodeoxyglucose positron emission tomography, and cognitive performance in: 1) APOE3Ch-PSEN1E280A carriers (n = 7) compared to typical PSEN1E280A carriers (n = 48) across the ADAD continuum; and 2) cognitively-unimpaired APOE3Ch carriers (n = 29) compared to non-carriers (n = 43) across the adult lifespan. Across ADAD, APOE3Ch-PSEN1E280A carriers had a 58% smaller age-related reduction in glucose metabolism in the precuneus, along with a 50% smaller reduction in episodic memory, compared to typical PSEN1E280A carriers. Across the adult lifespan, APOE3Ch carriers had remarkably preserved glucose metabolism in the anterior cingulate cortex, compared to non-carriers (slope: 0.001 versus – 0.009 /year). In this region, higher glucose metabolism in APOE3Ch carriers correlated with greater global cognition and episodic memory, as well as with fewer subjective memory complaints and depressive symptoms. Our findings suggest that the APOE3Ch variant may be protective against brain metabolic changes both in ADAD and across the disease-free adult lifespan, with corresponding benefits in cognitive function. Understanding the pathways through which APOE3Ch confers metabolic protective effects could inform novel therapeutic avenues to enhance resilience against AD and promote healthy brain aging.

Introduction

The APOE3 Christchurch (APOE3Ch) is a rare variant (R136S) that may protect against neurodegeneration and Alzheimer’s disease (AD). The first evidence of this protection was demonstrated in autosomal dominant AD (ADAD) due to the PSEN1E280A genetic variant from the Colombian ADAD kindred. In typical PSEN1E280A carriers, the onset of mild cognitive impairment (MCI) occurs by age 44 and the onset of dementia by age 491. The disease trajectory is characterized by progressive amyloid-β (Aβ) accumulation beginning approximately 20 years before the onset of cognitive symptoms, followed by tau pathology, cerebral glucose hypometabolism and neuronal injury2,3, largely mirroring the late-onset AD progression4.

Despite the near-complete penetrance of this PSEN1 variant, the homozygous APOE3Ch carrier found in the Colombian kindred developed MCI about 30 years after the expected age of onset5. Notably, at age 75, this individual had a high burden of Aβ but limited tau pathology and preserved cerebral glucose metabolism. Within the kindred, 27 PSEN1E280A carriers were also found to be heterozygous for APOE3Ch and showed a 5-year delay in cognitive decline, suggesting that the APOE3Ch variant may confer a dose-dependent protection against ADAD6. Imaging outcomes were similar in individuals with one copy of the APOE3Ch variant compared to the case carrying two copies, consistent with mechanisms that reduce tau pathology and neurodegeneration, even in the presence of high Aβ burden7. However, imaging results derive from samples of only one or two individuals and larger sample sizes are warranted to elucidate APOE3Ch-related protective phenotypes.

While protective in ADAD, whether the APOE3Ch variant may confer protection against age-related decline in individuals without the PSEN1 variant remains unknown. Across the adult lifespan, cerebral glucose metabolism, measured by 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET), has been shown to decline with age, even in the absence of overt pathology8,9. In addition, regional hypometabolism precedes and predicts subtle cognitive changes in healthy older adults, prior to cognitive impairment10-12. These findings support the utility of FDG-PET as a sensitive biomarker to gradual age-related changes and its potential to evaluate APOE3Ch-related protective effects against hypometabolism.

Therefore, this study investigated age-related alterations in cerebral glucose metabolism and in 1) heterozygous APOE3Ch-PSEN1E280A carriers compared to typical PSEN1E280A carriers across the ADAD continuum; and in 2) cognitively unimpaired heterozygous APOE3Ch carriers compared to non-PSEN1E280A carriers across the adult lifespan. We hypothesized that the APOE3Ch variant would provide brain metabolic protection, such that both dual APOE3Ch-PSEN1E280A and APOE3Ch carriers would show greater preservation of cerebral glucose metabolism with age compared to their respective control groups. Additionally, we examined associations between cerebral glucose metabolism and cognitive performance across groups. Given the rarity of the APOE3Ch variant, with a 0.004% allele frequency in a UK Biobank sample of 455,306 individuals13 and a 0.003% global frequency in gnomAD, the Colombian kindred represents a unique opportunity to study its biological and clinical significance.

Results

Study demographics

Among PSEN1E280A carriers, 48 were non-carriers for APOE3Ch (Mage: 39 [32, 45] years, 63% females, 35% cognitively impaired) and 7 were heterozygous for APOE3Ch (Mage: 49 [35, 53], 29% females, 29% cognitively impaired) (Table 1). There were no statistically significant differences in demographics or cognitive metrics between groups, possibly due to imbalanced and small sample sizes. Among cognitively unimpaired non-PSEN1E280A carriers, 43 were non-carriers for APOE3Ch (Mage: 40 [36, 44] years, 60% females) and 29 were heterozygous for APOE3Ch (Mage: 43 [35, 50], 52% females). APOE3Ch carriers had less years of education (11 [5, 13] versus 12 [8, 15] years, p = 0.034), lower processing speed (higher Trail Making Test Part A [TMTA] time: 49 [39, 62] versus 40 [30, 50] s, p = 0.016) and greater burden of depressive symptoms (higher Geriatric Depressive Scale [GDS] scores: (2 [1, 4] versus 0 [0, 2], p = 0.002) than non-carriers. There were no significant differences in age, sex distribution or other cognitive domains.

Table 1.

Sample demographics by group.

PSEN1E280A

Carriers
N = 481 		APOE3Ch-
PSEN1E280A
Carriers
N = 71 		p-value2 Non-Carriers
N = 431 		APOE3Ch

Carriers
N = 291 		p-value2
Age (years) 39 (32, 45) 49 (35, 53) 0.10 40 (36, 44) 43 (35, 50) 0.3
Female sex (n) 30 (63%) 2 (29%) 0.12 26 (60%) 15 (52%) 0.5
Education (years) 11 (7, 13) 11 (7, 11) 0.2 12 (8, 15) 11 (5, 13) 0.034
Cognitively impaired (n) 17 (35%) 2 (29%) >0.9 0 (0%) 0 (0%) >0.9
MMSE score 27 (23, 29) 27 (19, 28) 0.7 28 (28, 29) 28 (28, 29) >0.9
CERAD WLL delayed recall score 7 (2, 8) 5 (1, 8) 0.5 8 (6, 9) 7 (6, 8) 0.3
Naming score 13 (12, 14) 13 (12, 15) 0.7 13 (12, 14) 13 (11, 14) 0.4
TMTA time (s) 45 (36, 85)† 63 (39, 70)† 0.8 40 (30, 50) 49 (39, 62) 0.016
TMTB time (s) 94 (55, 164)† 108 (58, 258)† 0.8 80 (51, 119)† 107 (74, 151)† 0.077
Raven score 10 (7, 10)† 10 (8, 11) 0.6 10 (8, 11) 10 (9, 11) 0.4
SMC score 12 (5, 24)† 7 (4, 27) 0.5 6 (2, 8)† 5 (3, 9)† 0.9
GDS score 1.0 (0, 3.5) 1.0 (0, 1.0) 0.2 0 (0, 2) 2 (1, 4) 0.002
Open in a new tab
1

Median (Q1, Q3); n / N (%)

2

Wilcoxon rank sum test; Pearson's χ2 test

†

Missing values in PSEN1E280A carriers for TMTA time (n=1), TMTB time (n=19), abstract reasoningscore (n=1) and subjective memory complaints score (n=1); in APOE3Ch + PSEN1E280A carriers forTMTA time (n=1) and TMTB time (n=3); in non-carriers for TMTB time (n=8) and subjective memorycomplaints score (n=1); and in APOE3Ch carriers for TMTB time (n=8) and subjective memorycomplaints score (n=5)

Age-related alterations in cerebral glucose metabolism and cognitive performance in APOE3Ch-PSEN1E280A carriers across ADAD

To investigate age-related alterations in cerebral glucose metabolism, we first selected a priori the precuneus and hippocampus as our regions of interest (ROI), due to their metabolic role in AD pathology and aging9,14, and then used an exploratory voxelwise approach. Our adjusted ROI-based analyses within ADAD revealed a significant age-by-group interaction for FDG SUVR in the precuneus (β = 0.011 [0.004, 0.018], p = 0.003; Figure 1A), indicating that the association between age and precuneus glucose metabolism differed between groups. Specifically, APOE3Ch-PSEN1E280A carriers had a 58% smaller age-related reduction in precuneus glucose metabolism with age (β = −0.008 [−0.012, −0.004]), compared to a steeper reduction in typical PSEN1E280A carriers (β = −0.019 [−0.027, −0.011]). No significant age-by-group interaction was observed for hippocampal FDG SUVR (β = 0.002 [−0.007, 0.012], p = 0.629; Figure 1B). As expected due to our limited sample size, exploratory voxelwise analyses did not survive correction for multiple comparisons. Though FDR-uncorrected clusters should be interpreted with caution, the positive age-by-group contrast suggested a trend for slope differences in the precuneus and frontotemporal regions, consistent with greater preservation of regional glucose metabolism with age in APOE3Ch-PSEN1E280A carriers compared to typical PSEN1E280A carriers (Figure 1C).

Figure 1.

Figure 1

Open in a new tab

Age-related alterations in cerebral glucose metabolism and cognitive performance in APOE3Ch-PSEN1E280A carriers compared to PSEN1E280A carriers across ADAD. Scatterplots showing the correlations between age and precuneus (A) and hippocampus FDG SUVR (B) per group. FDR-uncorrected voxelwise maps of the slope differences in age-related FDG SUVR changes between groups (C). Warm colors represent larger effect sizes for positive interaction effects, representing smaller age-related reduction in FDG SUVR in APOE3Ch-PSEN1E280A carriers relative to typical PSEN1E280A carriers. No clusters survived negative interactions. Scatterplots showing the correlations between age and MMSE (D), CERAD Word List delayed recall (E), SMC (F) and GDS (G) scores between groups. β represent the regression coefficient for interaction effects, i.e., slope differences between groups, from age, sex and education-adjusted robust regression models. The dashed vertical line represents the estimated age of MCI onset of typical PSEN1E280A carriers from the Colombian kindred.

Across cognitive performance scores, we examined a battery of neuropsychological tests to assess global cognition, delayed recall, subjective memory complaints and depressive symptoms. There was not a significant age-by-group interaction for global cognition (Mini-Mental State Examination [MMSE] scores: β = 0.129 [−0.098, 0.357], p = 0.259; Figure 1D). However, significant slope differences in the Consortium to Establish a Registry for Alzheimer’s disease (CERAD) Word List scores between groups (β = 0.159 [0.050, 0.269], p = 0.005; Figure 1E) indicated that delayed recall had a 50% smaller age-related reduction in APOE3Ch-PSEN1E280A carriers (β = −0.158 [−0.264, −0.051]) than in typical PSEN1E280A carriers (β = −0.317 [−0.407, −0.227]). Memory complaints tended to occur later life in APOE3Ch-PSEN1E280A carriers than in typical PSEN1E280A carriers, with similar Subjective Memory Complaints (SMC) scores slopes between groups (β = −0.185 [−0.778, 0.408], p = 0.534; Figure 1F). The interaction for GDS scores was also not statistically significant (β = −0.048 [−0.168, 0.072], p = 0.426; Figure 1G), but we observed a trend for lower and more stable depressive symptoms with age in APOE3Ch-PSEN1E280A carriers compared to typical PSEN1E280A carriers.

Age-related alterations in cerebral glucose metabolism and cognitive performance in cognitively unimpaired APOE3Ch carriers across the adult lifespan

In adjusted ROI-based analysis within the adult lifespan, we did not observe a significant age-by-group interaction for FDG SUVR in the precuneus (β = 0.003 [−0.004, 0.010], p = 0.424; Figure 2A). However, there were significant slope differences in hippocampal FDG SUVR (β = 0.003 [0.000, 0.006], p = 0.039; Figure 2B), with APOE3Ch carriers exhibiting a stable hippocampal glucose metabolism over age (β = 0.001 [−0.000, 0.002]), whereas non-carriers for the APOE3Ch or PSEN1E280A variants showed a significant age-related reduction in metabolism (β = −0.002 [−0.005, −0.000]). FDR-corrected voxelwise analysis further showed a positive interaction between groups, in a statistical cluster that primarily included the anterior cingulate cortex (ACC; Figure 2C). Post-hoc ROI-based analysis of FDG SUVR in the ACC confirmed an age-by-group interaction (β = 0.008 [0.004, 0.012], p < 0.001), in which glucose metabolism did not decrease with age in APOE3Ch carriers (β = 0.001 [−0.001, 0.002]) in contrast to the age-related reduction in glucose metabolism in non-carriers (β = −0.008 [−0.011, −0.004]). The ACC FDG SUVR model was additionally adjusted for GDS scores and ACC volume and the interaction effect remained significant (β = 0.009 [0.004, 0.013], p < 0.001).

Figure 2.

Figure 2

Open in a new tab

Age-related alterations in cerebral glucose metabolism and cognitive performance in cognitively unimpaired APOE3Ch carriers compared to non-carriers across the adult lifespan. Scatterplots showing the correlations between age and precuneus (A) and hippocampus FDG SUVR (B) per group. FDR-corrected voxelwise maps of the slope differences in age-related FDG SUVR changes between groups (C). Warm colors represent larger effect sizes for positive interaction effects, representing smaller age-related reduction in FDG SUVR in APOE3Ch carriers relative to non-carriers. No clusters survived negative interactions. Scatterplots showing the correlations between age and MMSE (D), CERAD Word List delayed recall (E), SMC (F) and GDS (G) scores between groups. β represent the regression coefficient for interaction effects, i.e., slope differences between groups, from age, sex and education-adjusted robust regression models.

As for cognitive performance scores, a significant positive age-by-group interaction for MMSE scores (β = 0.092 [0.009, 0.175], p = 0.031; Figure 2D) demonstrated that global cognition decreased with age in non-carriers (β = −0.072 [−0.128, −0.016]) but not in APOE3Ch carriers (β = −0.020 [−0.046, 0.086]). Though with similar trends as MMSE scores, there were no significant slope differences in delayed recall, subjective memory complaints or depressive symptoms between groups (CERAD Word List: β = 0.032 [−0.069, 0.132], p = 0.530; SMC: β = −0.167 [−0.534, 0.201], p = 0.368; GDS: β = −0.123 [−0.262, 0.016], p = 0.082; Figure 2E-G). There were also no significant differences in secondary cognitive domains over time between groups, including scores assessing language and word retrieval abilities, processing speed and attention, executive function and abstract reasoning (Supplementary Figure 1A-D).

Associations between cerebral glucose metabolism and cognitive performance in cognitively unimpaired APOE3Ch carriers across the adult lifespan

Given that cerebral glucose metabolism and cognitive performance were stable over time in APOE3Ch carriers, we computed age-adjusted Z-scores for regional FDG SUVR values and cognitive scores in APOE3Ch carriers relative to the non-carriers. Adjusted robust regression analyses revealed significant associations between ACC FDG SUVR z-scores and cognitive z-scores, indicating that greater cerebral glucose metabolism in the ACC in APOE3Ch carriers compared to non-carriers was associated with greater global cognition (MMSE: β = 0.494 [0.074, 0.914], p = 0.023; Figure 3A), greater delayed recall (CERAD Word List: β = 0.280 [0.062, 0.497], p = 0.014; Figure 3B), lower subjective memory complaints (SMC: β = −0.312 [−0.539, −0.086], p = 0.009; Figure 3C) and lower burden of depressive symptoms (GDS: β = −0.312 [−0.620, −0.005], p = 0.047; Figure 3D) in APOE3Ch carriers compared to non-carriers. Greater ACC glucose metabolism in APOE3Ch carriers compared to non-carriers was not correlated with secondary cognitive domains (Supplementary Figure 1E-H). Similar models were not computed across the ADAD population due to the limited sample size.

Figure 3.

Figure 3

Open in a new tab

Associations between cerebral glucose metabolism and cognitive performance in cognitively unimpaired APOE3Ch carriers across the adult lifespan. Scatterplots showing the correlation between FDG SUVR z-scores in the anterior cingulate cortex (ACC) and MMSE (A), CERAD Word List delayed recall (B), SMC (C) and GDS (D) age-adjusted z-scores in APOE3Ch carriers relative to non-carriers. Dashed horizontal and vertical lines represent the quadrants for positive and negative age-adjusted z-scores, with positive z-scores representing higher values for APOE3Ch carriers than non-carriers. β represent the regression coefficient for the main effect from age, sex and education-adjusted robust regression models.

Discussion

In this study, we evaluated the effect of the APOE3Ch variant on age-related changes in cerebral glucose metabolism and cognitive performance in two distinct populations: one group with ADAD due to the PSEN1E280A variant and one group without the PSEN1E280A variant representing the normal adult lifespan.

We observed that dual APOE3Ch-PSEN1E280A carriers had a smaller age-related reduction in glucose metabolism in the precuneus, accompanied by a smaller age-related reduction in episodic memory, compared to typical PSEN1E280A carriers. Among cognitively unimpaired APOE3Ch carriers without the PSEN1E280A variant, we found a preservation in glucose metabolism in the ACC, compared to non-carriers. This increased glucose metabolism in APOE3Ch carriers relative to non-carriers correlated with higher global cognition and episodic memory performance and with lower subjective memory complaints and depressive symptoms. Our findings suggest that the APOE3Ch variant may be protective against brain metabolic changes in the ADAD continuum and across the disease-free lifespan, representing a promising target for enhancing resilience against neurodegenerative diseases.

Within the Colombian kindred, typical PSEN1E280A carriers exhibit reduced cerebral glucose metabolism approximately 15 years prior to the onset of MCI, affecting primarily the precuneus and posterior cingulate cortex, followed by frontotemporal regions as the disease advances14. Here, we compared age-related cerebral glucose metabolism alterations in PSEN1E280A carriers from the kindred with and without the APOE3Ch variant. Dual APOE3Ch-PSEN1E280A carriers exhibited a 58% slower reduction in glucose metabolism in the precuneus compared to typical PSEN1E280A carriers (−0.008 versus – 0.019). The regional pattern trending toward smaller glucose metabolism reductions in APOE3Ch-PSEN1E280A carriers matched the hypometabolic pattern seen in ADAD and late-onset AD cases14-16. These metabolic slope differences were concurrent with a 50% slower rate of episodic memory decline in APOE3Ch-PSEN1E280A carriers compared to typical PSEN1E280A carriers (−0.158 versus – 0.317). Despite the small sample size due to the low frequency of the APOE3Ch variant, our results support and expand upon the findings from the first APOE3Ch cases found in the kindred5,6.

Limited tau pathology in the presence of high Aβ burden has been proposed to result from increased microglial response to Aβ and reduced APOE3Ch binding to heparan sulfate proteoglycans, which facilitate tau uptake and propagation7,17,18. Considering the temporal cascade of AD, our results suggest that the absence of tau pathology in APOE3Ch-PSEN1E280A carriers may hinder the decline in glucose uptake, which would otherwise be near-certain due to the high penetrance of the PSEN1 variant. Notably, glucose metabolism has been shown to be preserved in response to early Aβ burden in young healthy adults as a neural reserve process19. Thus, limited tau pathology and preserved glucose metabolism in APOE3Ch-PSEN1E280A carriers may synergistically or independently protect against clinical symptoms, potentially through less severe neurodegeneration, delaying the age of onset. This potential brain metabolic protection conferred by the APOE3Ch variant against ADAD progression should be interpreted with caution given the limited sample and warrants further investigation. The underlying mechanisms of resilience to disease progression could inform the development of therapeutic approaches, such as the novel disease-modifying APOECh-mimetic antibody20.

Less is known about the effect of the APOE3Ch variant in the absence of deterministic pathology. To our knowledge, this is the first study to explore imaging biomarkers in APOE3Ch carriers without genetic predisposition to AD. FDG-PET is a highly sensitive biomarker to investigate age-related changes across the disease-free adult lifespan, as cerebral glucose metabolism decreases with healthy aging8,9. Indeed, we found that glucose metabolism in the precuneus and hippocampus decreased with age in non-carriers, despite relatively young age. Slope differences between groups were observed in the hippocampus, where glucose metabolism remained preserved with age in APOE3Ch carriers (0.001 versus – 0.002). Our exploratory analysis further revealed that the ACC was among the most metabolically preserved brain regions in APOE3Ch carriers (0.001 versus – 0.008). This pattern of metabolic maintenance is notably different than that of dual APOE3Ch-PSEN1E280A carriers, suggesting a dissociation between processes involved in resilience to AD-specific pathology versus healthy aging.

In line with our findings, a study on cognitively intact adults over 80 years reported a signature of preserved metabolism that included primarily the ACC21. We theorize that genetic protection might act through similar pathways, only decades earlier, though longitudinal follow-up of APOE3Ch carriers will provide better insights into their aging process. The ACC is a frontal limbic brain region that serves as a key hub connecting cognitive and affective networks22,23. This region is characterized by specialized von Economo neurons, which support social-emotional and self-related functions24. In our study, increased ACC glucose metabolism in APOE3Ch carriers relative to non-carriers was associated with greater global cognitive performance, better episodic memory, lower subjective memory complaints and lower burden of depressive symptoms. This suggests that the preservation of glucose metabolism in the ACC may underlie the preservation of cognitive-behavioral function. While we cannot draw causal conclusions in our study, higher glucose metabolism has been causally linked to improved cognitive performance in animal models and genetic human studies25-27. Further exploring the functional and structural role of the ACC in this population may shed light on cognitive reserve and aging-resilience mechanisms, which could promote healthy aging.

Of note, brain metabolic as well as cognitive protection in APOE3Ch carriers was age-dependent, as baseline cerebral glucose metabolism and cognitive scores were lower in APOE3Ch carriers than non-carriers at younger ages. One possible explanation is the lower years of education and higher GDS scores in APOE3Ch carriers compared to non-carriers, which could influence cerebral glucose metabolism through brain reserve remodeling28,29. Lower baseline cerebral glucose metabolism could also stem from compensatory mechanisms necessary to sustain metabolic function over time, such as neurovascular or neuroinflammation function30. More studies are needed to understand cell-specific metabolic protection processes, specifically astrocytic and microglia alterations as well as changes in lipid metabolism pathways. In the UK Biobank sample, APOE3Ch carriers showed lower polygenic risk for AD compared to non-carriers, even though the APOE3Ch variant is not included in genome-wide association studies13. This finding suggests that epistatic interactions between APOE3Ch and other genetic variants may contribute to the protection observed.

Several limitations in our study should be noted. While the high prevalence of APOE3Ch in the Colombian kindred is the key strength of our study, the relatively small sample of APOE3Ch carriers does not allow for complex statistical modeling and multiple comparisons or stratification into sex and APOE genotype. APOE4 alleles are linked to cerebral glucose hypometabolism, while APOE2 alleles are associated with preservation of metabolism31,32. The cross-sectional and observational design of this study limits inferences about disease progression and aging trajectories and precludes causal interpretations of APOE3Ch effects. Future biomarker studies in the Colombian kindred will focus on systemic and central vascular changes. Animal models and genetic studies are further necessary to complement our findings with mechanistic insight.

Overall, we present novel evidence consistent with a metabolic resilience phenotype conferred by the APOE3Ch variant across AD-related pathology as well as in the disease-free adult lifespan. Brain metabolic protection and corresponding benefits in cognitive function may contribute to cognitive reserve through compensatory mechanisms. Understanding the pathways through which APOE3Ch confers protective effects could inform novel therapeutic avenues to enhance resilience to AD and promote healthy brain aging.

Methods

Study participants

In the Colombian kindred, our group has identified a total of 121 APOE3Ch carriers, among whom 1 was homozygous for APOE3Ch and heterozygous for PSEN1E280A, 27 were heterozygous for both APOE3Ch and PSEN1E280A variants and 93 were heterozygous for APOE3Ch but were not PSEN1E280A carriers6. The determination of the APOE3Ch and PSEN1E280A status has been described previously5,6. Of these, 7 APOE3Ch-PSEN1E280A carriers and 33 APOE3Ch carriers underwent FDG-PET imaging and cognitive examinations. To study APOE3Ch-related alterations in the adult lifespan (aim 2), we excluded 4 cognitively impaired APOE3Ch carriers. Clinical status was assessed using the Functional Assessment Staging Tool (FAST), with scores of FAST > 3 indicating objective cognitive impairment33. In this study, we additionally included 48 typical PSEN1E280A carriers and 43 non-carriers for PSEN1E280A without the APOE3Ch variant who had available clinical assessments and FDG-PET data, to serve as control groups for each aim, respectively (Supplementary Fig. 2). All participants or their representatives provided written informed consent for all study procedures. Clinical and neuroimaging assessments were conducted by trained staff blinded to the genetic status at the Grupo de Neurociencias de Antioquia, Universidad de Antioquia, Medellín, Colombia and the Hospital Pablo Tobón Uribe, Medellín, Colombia. This study was approved by the Universidad de Antioquia institutional review board.

Clinical assessments

A comprehensive battery of neuropsychological assessments adapted to this Colombian population was conducted in Spanish. In our study, the primary cognitive performance assessments were selected for their sensitivity to detecting early changes in typical PSEN1E280A carriers across the different cognitive-behavioral domains34. Specifically, we included the Mini-Mental State Examination (MMSE, range: 0–30) to assess global cognition35; the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) Word List Learning to assess delayed recall (range: 0–10)36; the self-reported Memory Complaint Scale to assess subjective memory complaints (SMC, range: 0–45)37; and the Geriatric Depression Scale (GDS, range: 0–15) to assess burden of depressive symptoms38. To complement these domains, secondary and exploratory cognitive assessments included the CERAD Boston Naming Test (range: 0–15) to assess language and word retrieval abilities; the Trail Making Parts A and B (TMTA and TMTB) times to assess processing speed and attention as well as executive function, respectively; and Raven’s Matrices (range: 0–15) to assess abstract reasoning34.

Neuroimaging acquisition and processing

FDG-PET scans were acquired on a 64-section PET/computed tomography (CT) scanner (Biograph mCT; Siemens), within a median (Q1,Q3) of 2.7 (0.2, 7.1) months from neuropsychological assessment. Intravenous injection of 5 mCi of FDG was administered in fasting conditions followed by a 30-minute radiotracer uptake. A dynamic emission scan of six 5-minute frames (matrix: 400x400x81, voxel size: 1x1x2 mm3) was reconstructed with CT attenuation correction. Frames were aligned, averaged and spatially normalized using affine transformations to a subject-specific group template in the Montreal Neurological Institute (MNI) reference space (isotropic 2 mm3 voxel size). The inverse transformation was applied to the Automated Anatomical Labeling 3 (AAL3) atlas to align with each subject’s FDG-PET native space39. All images were visually inspected to ensure proper alignment. Next, FDG uptake was normalized to that of the pons, a widely used reference region to minimize intra-subject variability in FDG-PET40. The resulting standardized uptake value ratio (SUVR) maps were finally smoothed with a Gaussian kernel of 8 mm full width at half maximum. The mean SUVR was computed for bilateral regions of interest (ROI) of the AAL3 atlas (90 regions). We selected a priori the precuneus and the hippocampus as ROIs, considering the regional metabolic changes that occur in AD and aging9,14. In addition to ROI-based analysis, SUVR maps in the MNI space were computed to allow for exploratory voxelwise analysis. This approach informed post-hoc analyses of other ROIs.

Magnetic resonance imaging (MRI) was acquired on a 3T Tim Trio (Siemens). The structural T1-weighted image (3D MPRAGE sequence, repetition time: 2300 ms, echo time: 2 ms, inversion time: 900 s, flip angle: 8°, matrix: 192x256x256, voxel size: 0.9x0.9x0.9 mm3) was normalized to the MNI template space using non-linear transformation. The inverse deformation field was applied to the AAL3 atlas to obtain bilateral regional volumes in the same ROIs.

Statistical analysis

Study characteristics were compared between APOE3Ch-PSEN1E280A carriers and typical PSEN1E280A carriers (aim 1) and between non-PSEN1E280AAPOE3Ch carriers and non-carriers (aim 2) using Wilcoxon rank sum test and Pearson's χ2 test as appropriate. We compared age-related alterations in cerebral glucose metabolism and cognitive performance between groups within each aim. For this purpose, we used robust linear regression models using an interaction term between age and group as main predictor. To explore cerebral glucose metabolism, precuneus or hippocampus FDG SUVR were used as outcome. A similar model was computed at the voxelwise level, after threshold-free enhancement correction and false discovery rate (FDR) correction. To explore cognitive performance, MMSE, CERAD Word List delayed recall, SMC and GDS scores were used as outcomes. For statistically significant interaction effects, which indicate slope differences between groups, we further estimated marginal trends to obtain the outcome-age correlation for each group separately. Robust regression was chosen because it is less sensitive to departures from model assumptions and reduces the influence of outlying observations compared to ordinary least squares regression. Models of FDG SUVR and cognitive performance were all adjusted for age, sex (female/male) and years of education. A post-hoc FDG SUVR model was additionally adjusted for regional volume and GDS scores.

Within aim 2, we transformed the variables of interest into age-adjusted z-scores for APOE3Ch carriers relative to non-carriers. Positive z-scores indicated higher values for APOE3Ch carriers than non-carriers, while negative z-scores indicated lower values for APOE3Ch carriers than non-carriers. The association between FDG SUVR z-scores and cognitive metrics z-scores was computed using robust linear regression adjusted for age, sex and years of education. Exploratory analyses investigated age-related alterations in secondary cognitive metrics as well as their correlations with FDG SUVR. Due to the relatively small sample size of APOE3Ch-PSEN1E280A carriers, this analysis was not computed for aim 1. Across the study, we report regression coefficients (β), 95% confidence intervals (CI) and p-values. Statistical significance was set at α < 0.05. All analyses were performed using R version 2024.04.2 + 764.

Supplementary Material

This is a list of supplementary files associated with this preprint. Click to download.

• APOE3ChFDGSupplementaryFinal.docx

Acknowledgements

We thank the Colombian families from the Colombian ADAD kindred for contributing their time to this study and their continued commitment to this research. We also thank Lusiana Martinez from Boston University for her assistance in coordinating data transfers across study sites. The authors also wish to acknowledge Dr. Francisco Lopera for his lifelong commitment and dedication to studying Alzheimer’s disease and other dementias, as well as his invaluable contributions to the Colombian kindred during his tenure as coordinator of the Grupo de Neurociencias of Universidad de Antioquia.

Competing Interests

YTQ received funding from the NIA (no. R01AG054671; RF1AG077627), the Massachusetts General Hospital (MGH) Executive Committee on Research (MGH Research Scholar Award), the Alzheimer’s Association and the National Institute of Neurological Disorders and Stroke (NINDS) (grant RM1NS132996). YTQ gratefully acknowledges support from the Dan and Diane Riccio Charitable Fund, which partially supported this research. The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. JAV is a co-founder of Epoch Biotech, a company developing protected case-inspired therapeutics. YTQ serves as consultant for Biogen. All other co-authors have nothing to disclose.

Contributor Information

Yakeel Quiroz, Massachusetts General Hospital, Harvard Medical School.

Catarina Tristão-Pereira, Boston University.

Vincent Malotaux, Boston University.

Yesica Zuluaga, Grupo de Neurociencias de Antioquia, Facultad de Medicina, Universidad de Antioquia, Medellín, Colombia.

Claudia Muñoz, Grupo de Neurociencias de Antioquia.

George Silva, Boston University.

Bing He, Boston University.

Diana Alzate, Grupo de Neurociencias de Antioquia, Facultad de Medicina, Universidad de Antioquia, Medellín, Colombia.

Yamile Bocanegra, Universidad de Antioquia.

Alejandro Guerrero, Universidad de Antioquia.

Laura Serna, Universidad de Antioquia.

Carolina Ospina, Universidad de Antioquia.

Monica Vidal, Hospital Pablo Tobón Uribe.

Sergio Alvarez, Hospital Pablo Tobón Uribe.

Rafael Posada-Duque, Universidad de Antioquia.

Joseph Arboleda-Velasquez, Schepens Eye Research Institute of Mass Eye and Ear and Department of Ophthalmology, Harvard Medical School, Boston, USA.

David Aguillon, Universidad de Antioquia.

Data availability

Data are not publicly available due to privacy or ethical restrictions. Anonymized clinical, genetic and imaging data supporting the findings of this study are available upon request, subject to an internal review by the corresponding author to ensure that anonymity, confidentiality, and genetic status are protected. Data requests will be considered based on a proposal review, and completion of a data sharing agreement, in accordance with the University of Antioquia.

References

  • 1.Acosta-Baena N. et al. Pre-dementia clinical stages in presenilin 1 E280A familial early-onset Alzheimer’s disease: a retrospective cohort study. The Lancet Neurology 10, 213–220 (2011). [DOI] [PubMed] [Google Scholar]
  • 2.Fleisher A. S. et al. Associations Between Biomarkers and Age in the Presenilin 1 E280A Autosomal Dominant Alzheimer Disease Kindred: A Cross-sectional Study. JAMA Neurol 72, 316 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Malotaux V. et al. Associations of plasma biomarkers with age in the presenilin-1 E280A autosomal dominant Alzheimer’s disease kindred. The Journal of Prevention of Alzheimer’s Disease 13, 100578 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Jack C. R. et al. NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimer’s & Dementia 14, 535–562 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Arboleda-Velasquez J. F. et al. Resistance to autosomal dominant Alzheimer’s disease in an APOE3 Christchurch homozygote: a case report. Nat Med 25, 1680–1683 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Quiroz Y. T. et al. APOE3 Christchurch Heterozygosity and Autosomal Dominant Alzheimer’s Disease. N Engl J Med 390, 2156–2164 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Sepulveda-Falla D. et al. Distinct tau neuropathology and cellular profiles of an APOE3 Christchurch homozygote protected against autosomal dominant Alzheimer’s dementia. Acta Neuropathol 144, 589–601 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Ishibashi K. et al. Longitudinal effects of aging on 18F-FDG distribution in cognitively normal elderly individuals. Sci Rep 8, 11557 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Allocca M. et al. Evaluation of Age and Sex-Related Metabolic Changes in Healthy Subjects: An Italian Brain 18F-FDG PET Study. JCM 10, 4932 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Hanseeuw B. J. et al. Fluorodeoxyglucose metabolism associated with tau-amyloid interaction predicts memory decline. Annals of Neurology 81, 583–596 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Mayblyum D. V. et al. Comparing PET and MRI Biomarkers Predicting Cognitive Decline in Preclinical Alzheimer Disease. Neurology 96, (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Segobin S. et al. FDG-PET Contributions to the Pathophysiology of Memory Impairment. Neuropsychol Rev 25, 326–355 (2015). [DOI] [PubMed] [Google Scholar]
  • 13.He K. Y. et al. Characterization of APOE Christchurch carriers in 455,306 UK Biobank participants. Mol Neurodegener 18, 92 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Tristão-Pereira C. et al. Contribution of local amyloid-β and tau burden to hypometabolism in autosomal-dominant Alzheimer’s disease. Brain Commun fcaf 508 (2026) doi: 10.1093/braincomms/fcaf508. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Minoshima S., Cross D., Thientunyakit T., Foster N. L. & Drzezga A. 18F-FDG PET Imaging in Neurodegenerative Dementing Disorders: Insights into Subtype Classification, Emerging Disease Categories, and Mixed Dementia with Copathologies. Journal of Nuclear Medicine 63, 2S-12S (2022). [DOI] [PubMed] [Google Scholar]
  • 16.Bateman R. J. et al. Clinical and Biomarker Changes in Dominantly Inherited Alzheimer’s Disease. N Engl J Med 367, 795–804 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Nelson M. R. et al. The APOE-R136S mutation protects against APOE4-driven Tau pathology, neurodegeneration and neuroinflammation. Nat Neurosci 26, 2104–2121 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Chen Y. et al. APOE3ch alters microglial response and suppresses Aβ-induced tau seeding and spread. Cell 187, 428–445.e20 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Goyal M. S. et al. Brain aerobic glycolysis and resilience in Alzheimer disease. Proceedings of the National Academy of Sciences 120, e2212256120 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Marino C. et al. APOE Christchurch-mimetic therapeutic antibody reduces APOE-mediated toxicity and tau phosphorylation. Alzheimer’s & Dementia 20, 819–836 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Arenaza-Urquijo E. M. et al. The metabolic brain signature of cognitive resilience in the 80+: beyond Alzheimer pathologies. Brain 142, 1134–1147 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Lin F. et al. The cingulate cortex of older adults with excellent memory capacity. Cortex 86, 83–92 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Cera N., Esposito R., Cieri F. & Tartaro A. Altered Cingulate Cortex Functional Connectivity in Normal Aging and Mild Cognitive Impairment. Front Neurosci 13, 857 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Allman J. M. et al. The von Economo neurons in fronto-insular and anterior cingulate cortex. Ann N Y Acad Sci 1225, 59–71 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Minhas P. S. et al. Restoring hippocampal glucose metabolism rescues cognition across Alzheimer’s disease pathologies. Science 385, eabm6131 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Ueda M. et al. Cerebral glucose uptake and cognitive functions: A phenome-wide Mendelian randomization study. J Alzheimers Dis 110, 895–908 (2026). [DOI] [PubMed] [Google Scholar]
  • 27.Tristão-Pereira C., Casquero-Veiga M., Malotaux V. & Quiroz Y. T. Uncovering the causal link between cerebral glucose metabolism and cognitive function in Alzheimer’s disease. Journal of Alzheimer’s Disease 13872877261452181 (2026) doi: 10.1177/13872877261452181. [DOI] [PubMed] [Google Scholar]
  • 28.Ewers M., Insel P. S., Stern Y., Weiner M. W., & For the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Cognitive reserve associated with FDG-PET in preclinical Alzheimer disease. Neurology 80, 1194–1201 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Su L. et al. Cerebral metabolism in major depressive disorder: a voxel-based meta-analysis of positron emission tomography studies. BMC Psychiatry 14, 321 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Pineda-Lopez L. et al. Plasma extracellular vesicles from APOE3 Christchurch carriers display a protective phenotype in early stages of autosomal dominant Alzheimer’s disease. Alzheimers Dement 22, e71148 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Mosconi L. et al. MCI conversion to dementia and the APOE genotype: A prediction study with FDG-PET. Neurology 63, 2332–2340 (2004). [DOI] [PubMed] [Google Scholar]
  • 32.Guerrero A. et al. Longevity and cognitive resilience in a Colombian family carrying the APOE ε2 variant. Journal of Alzheimer’s Disease 13872877261449392 (2026) doi: 10.1177/13872877261449392. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Reisberg B. Functional assessment staging (FAST). Psychopharmacol Bull 24, 653–659 (1988). [PubMed] [Google Scholar]
  • 34.Giudicessi A. et al. Cognitive Outcomes in Autosomal-Dominant Alzheimer’s Disease: A Comprehensive Review from a Colombian Kindred with the Presenilin-1 E280A Mutation. JAD 101, 397–415 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Folstein M. F., Folstein S. E. & McHugh P. R. “Mini-mental state”. Journal of Psychiatric Research 12, 189–198 (1975). [DOI] [PubMed] [Google Scholar]
  • 36.Aguirre-Acevedo D. C. et al. Validity and reliability of the CERAD-Col neuropsychological battery. Rev Neurol 45, 655–660 (2007). [PubMed] [Google Scholar]
  • 37.Norton D. J. et al. Subjective memory complaints in preclinical autosomal dominant Alzheimer disease. Neurology 89, 1464–1470 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Bacca A. M., González A. & Uribe Rodríguez A. F. Validación de la Escala de Depresión de Yesavage (versión reducida) en adultos mayores colombianos. 1, 53–64 (2005). [Google Scholar]
  • 39.Rolls E. T., Huang C.-C., Lin C.-P., Feng J. & Joliot M. Automated anatomical labelling atlas 3. NeuroImage 206, 116189 (2020). [DOI] [PubMed] [Google Scholar]
  • 40.Verger A., Doyen M., Campion J. Y. & Guedj E. The pons as reference region for intensity normalization in semi-quantitative analysis of brain 18FDG PET: application to metabolic changes related to ageing in conventional and digital control databases. EJNMMI Res 11, 31 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data are not publicly available due to privacy or ethical restrictions. Anonymized clinical, genetic and imaging data supporting the findings of this study are available upon request, subject to an internal review by the corresponding author to ensure that anonymity, confidentiality, and genetic status are protected. Data requests will be considered based on a proposal review, and completion of a data sharing agreement, in accordance with the University of Antioquia.

Articles from Research Square are provided here courtesy of American Journal Experts

RESOURCES