Posttraumatic stress disorder (PTSD) is associated with elevated risk for cognitive decline and neurodegeneration in later life, yet the underlying biological mechanisms remain poorly understood. One hypothesized pathway involves dysfunction of the blood–brain barrier (BBB) via altered expression of claudin-5 (CLDN5), a tight junction protein key to BBB integrity. Our prior cross-sectional study found that DNA methylation (DNAm) in blood at three CLDN5 loci was associated with trauma exposure, PTSD symptom severity, and plasma neurofilament light (NFL). Building on these results, this study examined whether CLDN5 DNAm in blood predicted future PTSD symptom severity, NFL, neuropsychological performance, and CLDN5 RNA expression in blood, a possible proxy for CLDN5 in the brain. Trauma-exposed Veterans (N = 221; 174 returned for follow-up) were assessed twice over 5.6 years on average. Participants completed diagnostic interviews, provided blood samples at both timepoints, and completed a neuropsychological battery at follow-up (California Verbal Learning Test-II, Wechsler Adult Intelligence Scale and Delis-Kaplan Executive Function System subtests). Baseline CLDN5 DNAm at cg21872764 predicted increased PTSD symptom severity at follow-up, accounting for baseline PTSD symptoms (β = .128, p = .035). This probe was associated with future CLDN5 RNA expression (β= −.186, p-adj = .035) as was cg00804504 (β = .185, p-adj = .035). CLDN5 DNAm at cg16773741, linked to cognitive decline, predicted working memory performance at follow-up (β = .269, p-adj = .015). No loci were associated with NFL levels after corrections for multiple testing. Results suggest that CLDN5 DNAm is associated with increased PTSD severity over time among Veterans with substantial trauma histories and is associated with neuropsychological functioning. This raises the possibility that BBB disruption is part of PTSD pathology, which may help identify the biology underlying comorbidity between PTSD and neurodegenerative disorders.

Posttraumatic stress disorder (PTSD) is a psychiatric condition that may develop after trauma exposure. PTSD is characterized by considerable clinical heterogeneity. The amygdala's key role in fear conditioning makes it an important focus for investigating the neurobiology of PTSD. However, associations between amygdala volume and PTSD have been inconsistent. The amygdala consists of functionally distinct nuclei. Specific associations between amygdala nuclei volumes and PTSD may account for previous discrepancies between PTSD and whole amygdala volume. This study investigates the associations between amygdala nuclei volumes, PTSD diagnosis, severity, symptom cluster scores, age of onset and childhood trauma. Individuals with a PTSD diagnosis (n = 771) and controls (n = 1 081, 72% trauma-exposed) were sourced from the Enhancing Neuro-Imaging Genetics through Meta-Analysis and Psychiatric Genomics Consortium (mean age = 32.4 years, (SD = 13 years), 60% male). Nine amygdala nuclei volumes were compared to PTSD diagnosis, age of onset, overall severity, symptom cluster scores (re-experiencing, arousal, and avoidance/emotional numbing), and childhood trauma subscales. Analyses were performed using ordinary least-squares regression, corrected for age, sex, intracranial volume, and whole amygdala volume. PTSD diagnosis was not significantly associated with amygdala nuclei volumes. PTSD severity scores were associated with smaller right lateral nucleus volume (β = -0.26, pBON = 0.01). Smaller right lateral nucleus volume was also associated with re-experiencing (β = -1.01, pBON = 0.04) and arousal (β = -0.9, pBON = 0.04), smaller left paralaminar nucleus volume was associated with re-experiencing (β = -0.1, pBON = 0.04), smaller left corticoamygdaloid transition area volume was associated with avoidance (β = -0.31, pBON = 0.02). Larger left and right central nucleus volumes were significantly associated with childhood physical abuse (β = 0.24, pBON = 9 × 10-3) and neglect (β = 0.29, pBON = 0.04), respectively. Differences in select amygdala nuclei volumes among adults are associated with PTSD severity, symptom cluster scores, and childhood physical abuse and neglect. These findings demonstrate nuclei-specific patterns consistent with their functional roles in fear learning and expression.

Objective:Alzheimer's disease (AD) is a leading cause of death and disability, and treatment options for Alzheimer's disease and related dementias (ADRD) remain limited. We applied a data-driven, mechanism-agnostic Medication-Wide Association Study Plus (MWAS+) framework to identify candidate medications associated with ADRD using longitudinal electronic health record data and explainable artificial intelligence (AI). Methods:We used Veterans Health Administration electronic health record data from January 1999 to May 2022. The initial study population comprised 8,424,715 Veterans aged 65 years or older. Cases were defined by ADRD-related diagnosis codes or ADRD-related medication prescriptions, and controls were free of ADRD diagnosis and ADRD-related medication use. After exclusions and matching on sex, race, age at first encounter, and duration of follow-up, the primary analytic cohort included 505,817 matched case-control pairs (1:1; 1,011,634 Veterans). Longitudinal features were extracted from historical data up to 1 year before the index date and aggregated into 1-year intervals. We developed an upgraded Hybrid Value-Aware Transformer (HVAT 2.0) to jointly learn from longitudinal and nonlongitudinal clinical data while incorporating numerical values associated with clinical concepts, including cumulative medication dose. To enhance interpretability, we applied a medication-specific impact score method to estimate model-derived associations between medication exposure and ADRD risk. Findings:The model demonstrated stable performance across data partitions, with area under the receiver operating characteristic curve values of 0.791 in the training set, 0.772 in the validation set, and 0.775 in the testing set. Metolazone and varenicline were identified as the top 2 candidate medications with negative impact scores, suggesting potentially protective associations with new-onset ADRD. The impact score was -0.196 per unit of cumulative dose for metolazone (1800 mg) and -0.134 per unit for varenicline (280 mg). Although individual-level impact scores varied, most exposed patients had negative scores, including 12,020 of 12,480 metolazone users (96%) and 8,341 of 8,786 varenicline users (95%). Implications:This study demonstrates the feasibility of combining a medication-wide association framework, longitudinal dose-aware modeling, and explainable AI to identify candidate medications for ADRD from real-world electronic health record data. The findings should be interpreted as signals for hypothesis generation rather than evidence of causality. This framework may support prioritization of repurposing candidates for expert review, follow-up cohort validation, and future clinical investigation.

ABSTRACT Introduction Posttraumatic stress disorder (PTSD) is associated with an increased risk for metabolic syndrome (MetS). PTSD is also associated with low physical activity, an important determinant of cardiovascular and metabolic health. This study examined cross-sectional associations between MetS, PTSD, and self-reported global physical activity and moderate-to-vigorous physical activity (MVPA) to determine the extent to which PTSD and global physical activity and MVPA independently contribute to MetS. Methods Participants ( n = 191) were recruited from an ongoing longitudinal cohort study of veterans at the Translational Research Center for TBI and Stress Disorders (TRACTS). Logistic regression analyses were conducted to examine associations among MetS, PTSD symptoms, global physical activity, and MVPA. MetS and contributing risk factors (i.e., obesity, dyslipidemia, hypertension, and high blood glucose) were modeled as the dependent variables; PTSD symptoms, global physical activity, and MVPA as the independent variables (step 1); and age, sex, race, ethnicity, and nicotine and alcohol as the covariates (step 2). Results A significant interaction between PTSD symptoms and MVPA was observed for MetS (odds ratio (OR) = 0.92, 95% confidence interval (CI) = 0.86–0.99, P = 0.02), such that greater PTSD severity was associated with higher odds of MetS in insufficiently active participants, but not active individuals. A similar significant interaction between PTSD symptoms and MVPA was observed for obesity (OR = 0.95, 95% CI = 0.90–1.00, P = 0.05). Greater levels of global physical activity were associated with significantly lower odds of meeting the criteria for obesity (OR = 0.55, 95% CI = 0.35–0.85, P < 0.01) and high triglycerides (OR = 0.64, 95% CI = 0.39–1.00, P < 0.05). Greater MVPA was associated with significantly lower odds of having low-level high-density lipoproteins (OR = 0.42, 95% CI = 0.20–0.89, P < 0.05). Conclusions Higher levels of both global physical activity and MVPA were associated with a lower risk of MetS, obesity, and dyslipidemia in military veterans. Additionally, this study provides preliminary evidence that MVPA moderates the relationship between PTSD, MetS, and obesity.

Posttraumatic stress disorder (PTSD) is associated with early onset of neurological conditions, but the mechanism by which PTSD relates to diseases of the central nervous system is unclear. One possibility is that PTSD perpetuates breakdown of the blood brain barrier (BBB), allowing for bidirectional passage of molecules across the periphery and central nervous system that promote neuropathology. Preclinical studies have implicated claudin-5 (CLDN5), a protein integral to the integrity of the BBB tight junctions, in the pathogenesis of depression. Based on this, we evaluated if trauma exposure and PTSD related to CLDN5 epigenetics in blood among 1,311 trauma-exposed individuals (primarily Veterans) and in the brain tissue from 100 decedents. Three (out of 19) CLDN5 DNA methylation (DNAm) probes, cg00804504, cg17411190, and cg21872764, were significantly associated with trauma exposure or PTSD severity after multiple testing correction in blood. The latter two probes also showed association with PTSD diagnosis in ventromedial prefrontal cortex. The most strongly associated DNAm probe, cg21872764, also evidenced associations with the neuropathology biomarker neurofilament light in plasma. CLDN5 expression was strongly associated with estimated proportion of brain endothelial cells. The cross-sectional associations observed in this study highlight the importance of studying the link between traumatic stress and early onset of neuropathology. Future research is needed to test the mechanistic hypothesis that trauma exposure and chronic PTSD alter CLDN5 DNAm, lead to increased BBB permeability and allow for bidirectional passage of neuroinflammatory molecules across the BBB.

OBJECTIVE:The claudin-5 ( CLDN5) gene is critical for blood-brain barrier integrity and may link traumatic stress, accelerated aging, and neurological disease. Building on prior research showing associations between trauma exposure and PTSD with CLDN5 DNA methylation (DNAm), we tested whether candidate CLDN5 DNAm loci were associated with advanced epigenetic aging in blood and brain tissue and with hippocampal volume. METHODS:A total of 1302 trauma-exposed individuals (M age = 44.23, SD = 13.71; 76% males) underwent psychiatric diagnostic interviews and blood draws for obtaining epi/genetic information; 473 underwent magnetic resonance imaging of the brain. Data from 109 PTSD brain bank decedents with DNAm from ventromedial prefrontal cortex (vmPFC) were also examined (M age-at-death = 45.20, SD = 14.21; 62% males). RESULTS:All candidate loci were associated with metrics of epigenetic age in blood ( p-adj range: .0396 to 4.5e-05), and these associations largely extended to postmortem vmPFC. There was an indirect association between PTSD severity and CLDN5 DNAm in blood at cg21872764 through GrimAge residuals (indirect β = 0.033, p = .040) that was diminished when the direct PTSD association was modeled. The CLDN5 probe cg17411190 in blood was negatively related to left and right hippocampal volume ( p-adj = .042) and with volume of multiple hippocampal substructures. The association between PTSD severity and hippocampal volume was indirect through blood DNAm at cg17411190 (indirect β = -0.011, p = .045). CONCLUSIONS:PTSD severity-related accelerated aging may be associated with altered CLDN5 DNAm, which may signal neurodegeneration, such as reduced hippocampal volume. CLDN5 DNAm in blood may serve as a useful proxy for brain CLDN5 DNAm. Given that prior environmental enrichment and antidepressant studies show initial efficacy in altering CLDN5 expression, future studies could evaluate if PTSD treatment alters CLDN5 epigenetics and reduces risk for neurodegeneration.

BACKGROUND AND OBJECTIVES:Approximately 450,000 Veterans are living with Alzheimer disease and related dementias (ADRD), and the high prevalence of ADRD represents a major public health challenge for the Veterans Health Administration. While advancing age and genetic predisposition are well-established ADRD risk factors, growing evidence suggests that additional modifiable factors may also play an important role. This study leveraged data from the VA Million Veteran Program (MVP) to (1) estimate 10-year incidence of ADRD and (2) evaluate associations between a broad range of individual-level risk and resilience factors and incident ADRD in a large, nationally representative sample of Veterans. METHODS:This retrospective cohort study included Veterans aged ≥65 years at MVP enrollment who completed the MVP Baseline Survey and had VA electronic health record (EHR) data available. Individual-level variables including sociodemographic factors, military-specific characteristics, military environmental exposures (MEEs), health conditions, and health behaviors were characterized using MVP Baseline Survey data and supplemented with EHR data as available. The primary outcome was ADRD, which was determined using a validated algorithm based on International Classification of Diseases diagnosis codes extracted from the EHR. Associations between each risk/resilience factor and incident ADRD were examined using separate Cox regression models adjusted for age, sex, and education. RESULTS:The sample included 245,949 Veterans (age: mean 73.16, SD 6.84 years; 2.59% female). Approximately 4.56% (n = 11,216) of the sample developed ADRD over 10 years. History of traumatic brain injury (TBI; hazard ratio [HR] 2.96, 95% CI 2.76-3.17), depression (HR 2.93, 95% CI 2.82-3.04), and alcohol use disorder (AUD; HR 2.35, 95% CI 2.19-2.53) were the health factors most strongly associated with ADRD. ADRD risk was also elevated among Veterans with a history of exposure to Agent Orange (HR 1.09, 95% CI 1.03-1.14), chemical/biological warfare agents (HR 1.31, 95% CI 1.23-1.39), and pyridostigmine bromide tablets (HR 1.67, 95% CI 1.44-1.93). DISCUSSION:Findings identified TBI, depression, AUD, and MEEs as key variables associated with ADRD in Veterans. These factors may represent important targets for prevention and intervention efforts aimed at improving the long-term health of aging Veterans. Additional work is needed to clarify the mechanisms through which these factors influence ADRD risk and to establish whether observed associations are causal.

Introduction:Biobank-scale cohorts of individuals with genetic data and diagnoses of Alzheimer's disease and related dementias (ADRD) have facilitated the discovery of additional risk loci via meta-analysis, with existing cohorts assembled specifically for ADRD genetic discovery. Cross-ancestry meta-analyses have further elucidated the overall genetic architecture of these dementias. Here, we include for the first time the European ancestry (EA) and Hispanic ancestry (HA) subset of the VA Million Veterans Program (MVP) along with the African ancestry (AA) MVP participants in a meta-analysis with a large-scale EA and AA meta-analysis. Methods:Independent genome-wide association studies (GWASs) were conducted in MVP participants using four phenotypes derived from electronic medical records and surveys: ADRD, prescriptions for common dementia medications, and self-reported maternal and paternal history of dementia (dementia by proxy). These GWASs were repeated in the EA, AA, and HA cohorts. MVP ancestry-specific and cross-ancestry meta-analyses were conducted. These were then meta-analyzed with existing GWAS results. Functionality of the peak variants was explored using brain-derived gene expression data and co-localization analysis. Results:Apart from the APOE region, 17, 4, and 3 genome-wide significant (GWS) loci were observed in the MVP EA, AA, and HA meta-analyses, respectively. When we meta-analyzed these with consortium results, we observed 72 loci in the EA GWAS, and 62 lead loci in the cross-ancestry meta-analysis. While most of these loci were known, 27 genes/regions were identified containing variants surpassing genome-wide significance for the first time: 7 EA specific, 12 in the cross-ancestry meta-analysis, and 8 driven by AA and HA cohorts. Several of these are members of pathways containing established ADRD risk genes, and several of the peak SNPs showed evidence for eQTL effects on their respective genes. Several of the novel SNPs showed significant eQTL effects in brain-derived mRNA-seq experiments. Additionally, there was a significant differential expression of the novel gene PAX7 in ADRD cases and controls. Discussion:MVP represents a large and unique primarily male cohort comprised of US Veterans from a range of backgrounds with a unique set of environmental exposures. The results generated here demonstrate the utility of biobank level cohorts for AD genetic discovery. Furthermore, our discovery of ADRD genes was enhanced by the inclusion of MVP data that provided an increase of underrepresented ancestry groups in contrast to prior cross ancestry GWASs. The new AD risk loci identified present potential new targets for dementia treatment confirmed that future large-scale analyses of AD genetic risk and prediction will be enhanced by the inclusion of MVP data.

Abstract Background Although traumatic brain injury (TBI) has been identified as a risk factor for Alzheimer’s disease and related dementias (ADRD), not all studies have shown a clear link between TBI and ADRD, suggesting that the relationship between TBI and ADRD is complex, nuanced, and likely influenced by a multitude of factors. The purpose of this retrospective cohort study was to examine interactions between TBI history and co-occurring health conditions and health behaviors on 10-year incidence of ADRD among Veterans enrolled in the VA Million Veteran Program (MVP). Methods Participants ( N = 245,949) included Veterans aged ≥ 65 years at study enrollment who completed the MVP Baseline Survey and had VA electronic health record (EHR) data. Participants were followed from the date of MVP enrollment until the earliest ADRD diagnosis, death, or last visit date before the end of the observation period (January 2011 through September 2021). TBI status (TBI + vs. TBI-) and health conditions/health behaviors were characterized using a combination of MVP survey and EHR data. ADRD status (ADRD + vs. ADRD-) was based on a validated algorithm using EHR-extracted ICD codes. Cox proportional hazards regression analyses adjusted for age, sex, and education were used to assess the association between each health condition/health behavior and the hazard of ADRD in the TBI + and TBI- groups. Additive interactions between TBI and health conditions/health behaviors were tested using the relative excess risk due to interaction (RERI) statistic. Results Among Veterans with a TBI history ( n = 7,613), 12.16% developed ADRD over 10 years of follow-up; among those without TBI ( n = 238,336), 4.32% developed ADRD. RERI analyses showed significant additive TBI by health condition/health behavior interactions for depression (RERI = 1.55, 95% CI = 0.96–2.15), heart attack/coronary artery disease (RERI = 0.76, 95% CI = 0.29–1.24), and physical inactivity (RERI = 0.58; 95% CI = 0.10–1.05), such that ADRD risk in Veterans with TBI was increased for those with these health conditions/health behaviors compared to those without. Conclusions Findings suggest that ADRD risk following TBI may be heightened in the presence of certain health conditions/health behaviors, highlighting targeted areas of intervention for potentially mitigating adverse late-life outcomes in Veterans with a history of TBI.

Background:Electronic health record (EHR)-linked biorepositories provide opportunities to advance epidemiological research in Alzheimer's disease (AD) and related dementias. Objective:Evaluate the extraction, curation, and associative validity of Mini Mental State Examination (MMSE) scores from the VA EHR for participants in the VA Million Veteran Program (MVP). Methods:The sample (N = 49,555; 7.4% women) included a multiethnic cohort (European [68.3%], African [20.4%], Hispanic [9.0%]) with EHR-extracted MMSE scores; 30.7% were apolipoprotein E (APOE) e4 carriers, and 25.8% had multiple scores. Linear regressions examined cross-sectional associations between e4 dosage (0, 1, 2) and first and lowest MMSE scores. MMSE scores were also evaluated against MVP dementia diagnostic algorithms in participants aged ≥65 years. Results:Among participants of European ancestry, there was a significant e4 dose-response relationship (ps < .001) with MMSE scores. Homozygote carriers scored lower than heterozygote carriers (Mdiff: first = -0.5; lowest = -0.9), who scored lower than non-carriers (Mdiff: first = -0.4; lowest = -0.6). Among Veterans of African and Hispanic ancestry, no dose-response relationship was observed, although e4 carriers had lower scores than non-carriers (ps ≤ .04). MMSE scores corresponded strongly with dementia case/control status across phenotypes: mild impairment on the MMSE was strongly associated with AD (odds ratio [OR] = 11.48), with more severe MMSE impairment showing stronger associations (moderate OR = 17.95; severe OR = 27.83). Conclusion:This study demonstrated MMSE scores can be systematically extracted and curated from the VA EHR. Findings offer a scalable framework for future studies on risk stratification, highlighting the potential for harnessing MVP to explore genetic and clinical factors contributing to cognitive and dementia outcomes in diverse samples.

Alzheimer's Disease and related dementias (ADRD) are under diagnosed and that under diagnosis if detrimental to patients, their families and the VA Health Care System. This crisis of under diagnosis exacerbates existing disparities in healthcare, disproportionately affect Black Americans (BAs) compared to White Americans (WAs). Using a machine learning (ML) model to assign ADRD-risk scores may help to identify Veterans with undiagnosed ADRD and their risk of developing dementia in the near future. We previously developed a ML model based on 850+ variables extracted from structured and unstructured data from the VHA's vast electronic health records, using a training sample of 20,000 BA and 20,000 WA Veterans. Kaplan-Meier curves were calculated separately for race and ML score quintile group. A Cox-proportions hazards model was used to obtain an estimate of the hazard ratio (HR) for the risk of developing ADRD. The HR comparing the 75th to the 25th percentile of scores is estimated to be 2.20 (95% CI, 1.93-2.51, p  < 0.01). The Kaplan-Meier plot showed differences in risk for converting to ADRD between races; for BAs, the curves of the highest two quintiles have visibly steeper slopes than those for other quintiles; for WAs, only the highest quintile have visibly steeper slopes than the other quintiles. Higher ML scores were associated with higher risk of developing ADRD in the years following the index date. BAs in our sample were at a higher risk for developing ADRD compared to their WA counterparts with the same score.

Alzheimer’s disease (AD) is influenced by genetic and modifiable risk factors, including sex-specific hormones and vascular comorbidities that may contribute to disparities in disease risk. We developed an APOE-independent multi-ancestry polygenic risk score (PRS) for AD using a PRS-CS weighted summation approach, integrating summary statistics from European ancestry (EA), African American, and East Asian cohorts from the Alzheimer’s Disease Genetics Consortium. PRS performance was evaluated in the Alzheimer’s Disease Sequencing Project (ADSP) and validated in All of Us, Framingham Heart Study, and Minority Aging Research Study cohorts. We evaluated associations of the PRS with AD risk, cognitive performance and brain MRI measures among individuals with and without several vascular comorbid diseases and with hormone replacement therapy (HRT) use among women who had natural menopause. The AD PRS was associated with AD in the ADSP sample, with odds ratios per standard deviation ranging from 1.14 to 1.52 across ancestry groups and robust validation in independent cohorts (Figure 1). Stratified analyses showed that the association between PRS and AD risk was stronger among individuals with vascular comorbidities (Figure 2), suggesting that long-term poor vascular health may amplify polygenic risk for AD. The AD PRS was also associated with memory, executive function, and language performance, and sex-specific differences in memory score observed among EA individuals. Among individuals with a high PRS (top tertile of the PRS distribution), ever-users of HRT exhibited better memory trajectories over time compared to never-users. Timing played a crucial role: HRT initiation within five years of menopause was associated with higher memory scores (β = 0.24, p = 3.05×10⁻³; Figure 3A). Trajectory analyses indicated that the association between HRT and memory was most prominent in midlife and attenuated in later years (Figure 3B). AD PRS was also associated with AD-related brain MRI measures, suggesting its contribution to early structural changes linked to AD. Our findings demonstrate the utility of an ancestry-aware PRS for advancing understanding of AD in diverse populations. Future research incorporating AD PRS in studies of sex-specific, vascular, and hormonal risk factors may offer promise for personalized interventions and prevention strategies for AD

Cardiovascular diseases (CVDs) such as peripheral artery disease (PAD) and coronary artery disease (CAD) are risk factors for Alzheimer's disease (AD) and related dementias (ADRD). The APOE -ε4 variant, which codes for a cholesterol transporter protein, is the largest AD genetic risk factor, increases LDL cholesterol and triglycerides, and augments the risk of cardiovascular disease. In this study of participants in the US Department of Veterans Affairs’ Million Veteran Program (MVP), we examined the interactive effects of APOE -ε4 status with CVDs (PAD, CAD, myocardial infarction, hypertension, and hyperlipidemia) on ADRD prevalence. Our cohort included MVP participants of European ancestry age 65 and older with available genotype data ( n  = 11,112 ADRD cases and 170,361 controls). Cross-sectional logistic regression analyses were performed using the GEM (Gene–Environment interaction analysis in Millions of samples) software package and included fitting an omnibus test for gene by environment (GxE) interactions between additively-coded ε4 and the CVDs as a group, followed by GxE analysis of individual CVDs. Additive-scale interactions were measured using the Relative Excess Risk due to Interaction (RERI) statistic. ADRD was derived from International Classification of Diseases (ICD) codes using our validated algorithm. We used validated algorithms for MI and PAD identified in the VA's Centralized Interactive Phenomics Resource (CIPHER). CAD, hypertension, and hyperlipidemia cases were identified using Phecodes. CVDs showed both strong main-effect associations with ADRD (ORs 1.55 to 1.82, all p  < 10 99 ; see Table). Both the omnibus test ( p  = 5x10 -12 ) and the individual CVD interaction terms were significant ( p from 7x10 -8 to 0.025). RERI estimates indicated significant positive additive-scale interactions (see example figure illustrating additive hypertension x ε4 interaction). These additive-scale interactions are more directly interpretable than multiplicative-scale interactions. They indicate that the prevalence of ADRD associated with cardiovascular disease increases with the number of inherited APOE -ε4 alleles (e.g. from 3.3% greater ADRD frequency associated with hypertension at age 80 for those with 0 ε4 copies to 5.6% for those with 2 copies; see Figure). Combining genetic testing with information about health comorbidities could contribute to more accurate dementia risk assessment within the Veteran population, and likely within other populations as well.

A pilot medication-wide association study (MWAS) was conducted to identify candidate drugs for repurposing in the prevention of Alzheimer’s disease and related disorders (ADRD). This MWAS is a hypothesis-free, agnostic exploration of the Million Veteran Project (MVP) dataset, which contains extensive information on genotypes, drug exposure, and incident ADRD. Using an age-, sex-, and race-matched cohort ( n = 263,256) from the MVP, we trained a Histogram-Based Gradient Boosting (HGB) model incorporating PTSD status, Social Determinants of Health (SDOH) (represented by ADI scores based on patients' most recent zip codes), APOE ε4 status, and Polygenic Risk Scores (PRS) for ADRD. Other predictors included age, gender, ancestry (derived from genomic data), and scaled cumulative doses of vitamin D3, atorvastatin, gabapentin, and other medications. To assess the contribution of individual features to population-wide ADRD risk, we applied novel explainable AI methods that we developed and validated, calculating an impact score that quantifies each feature’s contribution to ADRD risk, comparable to odds ratio. Additionally, we calculated interaction scores for atorvastatin in relation to other model features, enabling a more nuanced analysis of drug-drug and drug-risk factor interactions. Our analysis confirmed previous findings and aligned with existing literature, suggesting both positive and negative associations between certain medications and ADRD risk. For example, atorvastatin had a negative impact score, indicating a decreased risk of ADRD. Additionally, atorvastatin showed a positive interaction with age, suggesting that its beneficial effect diminishes with increasing age—consistent with prior literature. We also observed a novel negative interaction between atorvastatin and escitalopram, indicating that their concurrent use may reduce ADRD risk beyond the additive effect. This finding is novel and plausible, as both drugs have been reported to lower ADRD risk. The impact and interaction scores for atorvastatin are presented in Tables 1 and 2. These findings demonstrate the robustness of the MWAS approach while highlighting new insights through the integration of genomic and SDOH data. It also underscores the potential of leveraging real-world data and explainable AI to identify candidate drugs for ADRD prevention.

INTRODUCTION:Few African American (AA) donors have been included in post mortem Alzheimer's disease (AD) studies compared to European-ancestry (EA) individuals. METHODS:We generated transcriptome-wide bulk pre-frontal cortex (PFC) gene expression data from 125 AA donors with neuropathologically determined AD and 82 AA controls. RESULTS:Transcriptome-wide significant differential expression was observed with 482 genes. The most significant, ADAMTS2, showed 1.52 times higher expression in AD cases (p = 2.96x10-8). Comparison of findings with those from a recent gene expression study of EA brain donors revealed substantial concordance, including ADAMTS2. Other associations not observed in EA results may be especially relevant to AD risk in the AA population. Examination of AA AD GWAS-implicated variants identified several expression quantitative trait loci. CONCLUSION:This first large-scale AA brain AD gene expression study identified many differentially expressed genes, including ADAMTS2, and supports gene expression as a molecular pathway underlying the impact of several AA AD risk variants. HIGHLIGHTS:We performed the largest African American brain tissue Alzheimer's disease (AD) gene expression study. Expression differences for 482 genes, notably ADAMTS2, were study-wide significant. Many significant differentially expressed genes are involved in energy metabolism. Several previously known AD-associated variants in African Americans are eQTLs. These results advance knowledge of the genetic basis of AD in the AA population.

BACKGROUND:Posttraumatic stress disorder (PTSD) has been associated with advanced epigenetic age cross-sectionally, but the association between these variables over time is unclear. This study conducted meta-analyses to test whether new-onset PTSD diagnosis and changes in PTSD symptom severity over time were associated with changes in two metrics of epigenetic aging over two time points. METHODS:We conducted meta-analyses of the association between change in PTSD diagnosis and symptom severity and change in epigenetic age acceleration/deceleration (age-adjusted DNA methylation age residuals as per the Horvath and GrimAge metrics) using data from 7 military and civilian cohorts participating in the Psychiatric Genomics Consortium PTSD Epigenetics Workgroup (total N = 1,367). RESULTS:Meta-analysis revealed that the interaction between Time 1 (T1) Horvath age residuals and new-onset PTSD over time was significantly associated with Horvath age residuals at T2 (meta β = 0.16, meta p = 0.02, p-adj = 0.03). The interaction between T1 Horvath age residuals and changes in PTSD symptom severity over time was significantly related to Horvath age residuals at T2 (meta β = 0.24, meta p = 0.05). No associations were observed for GrimAge residuals. CONCLUSIONS:Results indicated that individuals who developed new-onset PTSD or showed increased PTSD symptom severity over time evidenced greater epigenetic age acceleration at follow-up than would be expected based on baseline age acceleration. This suggests that PTSD may accelerate biological aging over time and highlights the need for intervention studies to determine if PTSD treatment has a beneficial effect on the aging methylome.

BACKGROUND: Posttraumatic stress disorder (PTSD) is accompanied by disrupted cortical neuroanatomy. We investigated alteration in covariance of structural networks associated with PTSD in regions that demonstrate the case-control differences in cortical thickness (CT) and surface area (SA). METHODS: Neuroimaging and clinical data were aggregated from 29 research sites in >1300 PTSD cases and >2000 trauma-exposed control subjects (ages 6.2-85.2 years) by the ENIGMA-PGC (Enhancing Neuro Imaging Genetics through Meta Analysis-Psychiatric Genomics Consortium) PTSD working group. Cortical regions in the network were rank ordered by the effect size of PTSD-related cortical differences in CT and SA. The top-n (n = 2-148) regions with the largest effect size for PTSD > non-PTSD formed hypertrophic networks, the largest effect size for PTSD < non-PTSD formed atrophic networks, and the smallest effect size of between-group differences formed stable networks. The mean structural covariance (SC) of a given n-region network was the average of all positive pairwise correlations and was compared with the mean SC of 5000 randomly generated n-region networks. RESULTS: Patients with PTSD, relative to non-PTSD control subjects, exhibited lower mean SC in CT-based and SA-based atrophic networks. Comorbid depression, sex, and age modulated covariance differences of PTSD-related structural networks. CONCLUSIONS: Covariance of structural networks based on CT and cortical SA are affected by PTSD and further modulated by comorbid depression, sex, and age. The SC networks that are perturbed in PTSD comport with converging evidence from resting-state functional connectivity networks and networks affected by inflammatory processes and stress hormones in PTSD.

Glial fibrillary acidic protein (GFAP) is an astrocytic marker that can be assessed in blood using single molecule array technology. Recent studies suggest that individuals with posttraumatic stress disorder (PTSD) have suppressed circulating levels of this CNS biomarker. This study examined the hypothesis that PTSD and plasma GFAP levels share common genetic and epigenetic pathways. Using data from 1096 veterans and civilians, we computed a PTSD polygenic risk score (PRS) derived from a prior PTSD genomewide association study (GWAS) and found that PTSD severity and the PRS were each associated with reduced levels of GFAP. To clarify the basis of the PRS association, we performed a GWAS of GFAP which identified 20 genomewide-significant loci including genes implicated in independent GWASs of PTSD and neurodegenerative disease (e.g., PRKN, NFIA). Comparison of the PTSD and GFAP GWAS results showed that PTSD-associated genes were significantly enriched in the GFAP results with notable overlap involving NPSR1 and the protocadherin alpha (PCDHA) gene cluster. Similarly, we performed an epigenomewide association study (EWAS) of GFAP, which identified 4 genomewide-significant associations (including loci in MCT4 and SREBF1) and then compared those results to the findings of a PTSD EWAS. Results again showed significantly greater overlap than would be expected by chance and included loci implicated in prior studies of depression, dementia, and inflammation. This study clarifies the genetic and epigenetic basis of the association between PTSD and plasma GFAP levels and should encourage future research into the role of GFAP in the pathophysiology of PTSD.

Although the rate of Alzheimer’s disease (AD) in African-ancestry (AA) Americans is higher than that of persons from European-ancestry (EA) populations, AA participants have been underrepresented in AD neuropathological studies. Utilizing the AD Research Centers (ADRC) infrastructure, we obtained AA donor pre-frontal cortex (PFC) tissue from brain repositories of 12 ADRC and generated bulk RNA sequencing (RNA-seq) data for 179 samples that met QC and inclusion criteria. Previously generated PFC RNAseq data were obtained for 28 additional AA donors from the Columbia University ADRC. Differential gene expression was evaluated among 125 donors with a neuropathological diagnosis of AD (NIA-Reagan intermediate or high likelihood) and 82 neuropathologically confirmed controls using regression models including covariates for age at death, sex, cell-type frequencies, and RNA integrity number (RIN) calculated with Limma. FDR-corrected p-values (p adj ) were calculated to control for the 33,611 genes examined. A total of 482 genes surpassed the multiple-testing threshold. The most significant, ADAMTS2 (p = 2.96 × 10 -8 , p adj = 0.001), showed increased expression in AD cases (see Table/Figure). We note that ADAMTS2 was differentially expressed in a prior EA study of neuropathologically confirmed AD cases and controls (Panitch et al. Molecular Psychiatry 2021). Additionally, a recent analysis of cognitive resilience in EA neuropathological AD cases identified a strong association with ADAMTS2 (See Li et al. AAIC2024 abstract). Of the differentially expressed genes observed in the Panitch et al. EA study, 385 (35%) were nominally significant,65 (5.8%) were corrected significant, and most (89%) of these genes had the same effect direction in the AA cohort. Some of the observed associations appear to be AA specific (e.g., EFR3B , IRS4 , and CA12 ; see Table). Additionally, we found nominally significant (p<0.05) associations with expression of APOE (Log2 fold change [L2FC = -0.20, p = 0.012) and several other established AD genes including SORL1 (L2FC = -0.10, p = 0.014) and IGF1R (L2FC = 0.11, p = 0.0013). This largest-ever (to our knowledge) gene expression study of AD in postmortem brain tissue from AA donors implicates many more genes as having a role in AD in this population than previously identified in genome-wide association studies and provides insight into trans-ancestry differences risk for AD.

Alzheimer’s disease (AD) risk variants have been identified in European ancestry cohorts that have stronger effects at certain ages, in individuals with a specific sex, or in those with specific isoforms of APOE, the strongest AD risk locus. However, sample sizes in African ancestry (AA) cohorts have been underpowered to perform stratified analyses. We generated genome-wide association study datasets stratified by sex, age at onset (< 75 vs ≥ 75), and APOE-ε4 carrier status in AA cohorts from MVP and the Alzheimer’s Disease Genetics Consortium (ADGC). Outcomes in MVP were AD and related dementias (ADRD; n = 4073 cases and 19,648 controls) and proxy dementia (i.e., reported dementia in a parent, n = 6216 cases and 21,566 controls) while ADGC analyses examined AD (n = 2425 cases and 5069 controls). The proxy dementia GWASs were included in the sex-stratified meta-analysis corresponding to the sex of the affected parent. The top genes were tested for differential expression in AA brain tissue. In addition to the APOE region, genome-wide significant associations were observed in an intergenic region near the EPHA5 gene (rs141838133, p = 2.19 × 10–8) in individuals with onset < 75 years, in GRIN3B near the known AD risk gene ABCA7 (rs115882880, p = 3.83 × 10–8) in females, and near TSPEAR (rs139130053, p = 4.27 × 10–8) in APOE-ε4 non-carriers. EPHA5 regulates glucose homeostasis, and ephrin receptors modify the strength of existing synapses in the brain and in pancreatic islets. It is unclear whether GRIN3B represents a locus distinct from ABCA7. Rs115882880 was a significant eQTL for GRIN3B but not ABCA7 in AA brain samples. TSPEAR regulates Notch signaling but has not been linked to neuronal function. Age, sex, and APOE-stratified analyses of dementia in AA participants from two cohorts revealed potential new associations. Stratified analyses may yield critical information about the genetic heterogeneity underlying dementia risk and lead to advances in precision medicine.