The Association between Exposure to Fine Particulate Matter and MRI-Assessed Locus Coeruleus Integrity in the Vietnam Era Twin Study of Aging (VETSA)

Abstract

Background: Increased exposure to ambient air pollution, especially fine particulate matter =2:5 lm (PM2:5) is associated with poorer brain health and increased risk for Alzheimer's disease (AD) and related dementias. The locus coeruleus (LC), located in the brainstem, is one of the earliest regions affected by tau pathology seen in AD. Its diffuse projections throughout the brain include afferents to olfactory areas that are hypothesized conduits of cerebral particle deposition. Additionally, extensive contact of the LC with the cerebrovascular system may present an additional route of exposure to environmental toxicants.Objective: Our aim was to investigate if exposure to PM2:5 was associated with LC integrity in a nationwide sample of men in early old age, poten-tially representing one pathway through which air pollution can contribute to increased risk for AD dementia.Methods: We examined the relationship between PM2:5 and in vivo magnetic resonance imaging (MRI) estimates of LC structural integrity indexed by contrast to noise ratio (LCCNR) in 381 men [mean age = 67:3; standard deviation ðSDÞ = 2:6] from the Vietnam Era Twin Study of Aging (VETSA). Exposure to PM2:5 was taken as a 3-year average over the most recent period for which data were available (average of 5.6 years prior to the MRI scan). We focused on LCCNR in the rostral-middle portion of LC due to its stronger associations with aging and AD than the caudal LC. Associations between PM2:5 exposures and LC integrity were tested using linear mixed effects models adjusted for age, scanner, education, household income, and interval between exposure and MRI. A co-twin control analysis was also performed to investigate whether associations remained after controlling for genetic confounding and rearing environment.Results: Multiple linear regressions revealed a significant association between PM2:5 and rostral-middle LCCNR (b = - 0:16; p = 0:02), whereby higher exposure to PM2:5 was associated with lower LCCNR. A co-twin control analysis found that, within monozygotic pairs, individuals with higher PM2:5 exposure showed lower LCCNR (b = - 0:11; p = 0:02), indicating associations were not driven by genetic or shared environmental confounds. There were no associations between PM2:5 and caudal LCCNR or hippocampal volume, suggesting a degree of specificity to the rostral-middle portion of the LC.Discussion: Given previous findings that loss of LC integrity is associated with increased accumulation of AD-related amyloid and tau pathology, impacts on LC integrity may represent a potential pathway through which exposure to air pollution increases AD risk.

Headnote

Background: Increased exposure to ambient air pollution, especially fine particulate matter =2:5 lm (PM2:5) is associated with poorer brain health and increased risk for Alzheimer's disease (AD) and related dementias. The locus coeruleus (LC), located in the brainstem, is one of the earliest regions affected by tau pathology seen in AD. Its diffuse projections throughout the brain include afferents to olfactory areas that are hypothesized conduits of cerebral particle deposition. Additionally, extensive contact of the LC with the cerebrovascular system may present an additional route of exposure to environmental toxicants.

Objective: Our aim was to investigate if exposure to PM2:5 was associated with LC integrity in a nationwide sample of men in early old age, poten-tially representing one pathway through which air pollution can contribute to increased risk for AD dementia.

Methods: We examined the relationship between PM2:5 and in vivo magnetic resonance imaging (MRI) estimates of LC structural integrity indexed by contrast to noise ratio (LCCNR) in 381 men [mean age = 67:3; standard deviation ðSDÞ = 2:6] from the Vietnam Era Twin Study of Aging (VETSA). Exposure to PM2:5 was taken as a 3-year average over the most recent period for which data were available (average of 5.6 years prior to the MRI scan). We focused on LCCNR in the rostral-middle portion of LC due to its stronger associations with aging and AD than the caudal LC. Associations between PM2:5 exposures and LC integrity were tested using linear mixed effects models adjusted for age, scanner, education, household income, and interval between exposure and MRI. A co-twin control analysis was also performed to investigate whether associations remained after controlling for genetic confounding and rearing environment.

Results: Multiple linear regressions revealed a significant association between PM2:5 and rostral-middle LCCNR (b = - 0:16; p = 0:02), whereby higher exposure to PM2:5 was associated with lower LCCNR. A co-twin control analysis found that, within monozygotic pairs, individuals with higher PM2:5 exposure showed lower LCCNR (b = - 0:11; p = 0:02), indicating associations were not driven by genetic or shared environmental confounds. There were no associations between PM2:5 and caudal LCCNR or hippocampal volume, suggesting a degree of specificity to the rostral-middle portion of the LC.

Discussion: Given previous findings that loss of LC integrity is associated with increased accumulation of AD-related amyloid and tau pathology, impacts on LC integrity may represent a potential pathway through which exposure to air pollution increases AD risk. https://doi.org/10.1289/ EHP14344

Introduction

Neuropathological studies indicate that abnormal tau may first appear in the locus coeruleus (LC) starting early in life,1,2 with substantial accumulation occurring during the course of Alzheimer's disease (AD).3 LC neuron counts decline as the dis-ease progresses, with a 50-80% reduction in LC neuron number in the later stages of AD.3,4 Furthermore, antemortem cognitive decline, mild cognitive impairment, and dementia5,6 as well as measures of pathological accumulation7 are associated with post-mortem histological measures of LC degeneration. The LC is the primary source of norepinephrine (NE) for the brain,8 and experi-mental evidence demonstrates that damage to or dysfunction of the LC can result in depleted levels of norepinephrine, leading to neu-roinflammation9 and increased accumulation of amyloid10-12 and tau pathology.13 In a transgenic rat model of AD, pathologic altera-tion of the LC resulted in cognitive deficits even prior to the forma-tion of amyloid plaques.14 Despite evidence that LC damage or dysfunction may be a key contributor to disease progression, less is known about factors that negatively impact LC integrity. Given the early involvement of the LC in the AD pathophysiological processes and broad downstream effects, identifying modifiable factors that exacerbate LC degeneration will be important for efforts to reduce risk for AD.

The 2020 Lancet Commission report on dementia prevention now includes late-life exposure to ambient air pollution as one potentially modifiable risk factor for AD and related dementias.15 Fine particulate matter =2:5 lm (PM2:5) is a particularly toxic component of ambient air pollution.16 Although the mecha-nisms of air pollution neurotoxicity are poorly understood, prior work has linked PM2:5 exposure to cognitive decline,17 higher levels of b-amyloid accumulation,18,19 and increased risk for dementia.20,21 The LC may be particularly vulnerable to PM2:5 for several reasons. First, the LC has high exposure to the circulatory system and any toxicants contained within, as each LC neuron innervates long stretches of capillaries.22 Second, LC neurons contain large, unmyelinated axons that leave them particularly vulnerable to neurotoxic insult.23 Third, additional studies suggest a direct nose-to-brain expo-sure pathway where PM2:5 and its ultrafine constituents are inhaled through the nose and enter the brain through the olfac-tory bulb,24-26 which has noradrenergic projections to the LC. Studies have observed negative effects of PM2:5 on autonomic function,27-29 which may also be disrupted in AD.30,31 This represents another piece of evidence implicating LC involve-ment given its role in autonomic regulation.32,33 Fourth, and more directly, alterations in peripheral norepinephrine uptake have been observed following controlled exposure to PM2:5 in both younger34 and older adults,35 which may impact LC integ-rity. These observations coupled with greater exposure of LC neurons to environmental toxicants lend support to the idea that LC integrity could be augmented by exposure to PM2:5.

The LC has traditionally been difficult to measure in vivo because it is not visible on standard T1- or T2-weighted magnetic resonance imaging (MRI) sequences. However, the development of specialized MRI sequences has now made MRI-based assess-ment of LC integrity possible. For example, one technique utilizes a fast spin-echo T1-weighted imaging sequence and compares signal intensity in left and right LC, visible as hyperintense foci, relative to a nearby reference region in the pontine tegmentum to control for variations in overall image intensity.36-38 This comparison results in a measure of LC contrast-to-noise ratio (LCCNR), in which higher values are interpreted to reflect better integrity. Lower LCCNR has been associated with poorer cognitive performance,39 cognitive impairment38 (MCI), reduced cortical thickness,40 lower cortical microstructural integrity,41 chronic pain,42 and increased AD pathology.43 Consistent with autopsy studies, LC imaging studies find that there is a rostral-caudal gradient such that aging and AD-related associations are more pronounced in the rostral and middle LC, with caudal regions being relatively spared.36,38,44-46

PM2:5 xhas been associated with various MRI indices of poorer brain health.47-49 A limitation of these studies is that their measures of brain integrity often focus on macrostructural estimates of corti-cal thickness and hippocampal volume, which may contribute to some of the mixed findings in the literature.50 That is, no studies have explored the putatively adverse effects of PM2:5 exposure on LC integrity, which may be more sensitive to subtle effects. Previous work suggests that indices of LC integrity are more strongly associated with other cognitive38 and brain41 outcomes than cortical thickness and hippocampal volume in a sample that is dementia-free. Moreover, declines in LC integrity occur early in the process of AD, even prior to substantial amyloid and tau accu-mulation,14,51 and may therefore be useful for detecting the earliest impacts of environmental exposures on brain health. We hypothe-sized that higher exposure to PM2:5 would be associated with lower LC integrity in a nationwide sample of men in early old age. Together with previous findings, this would suggest that LC vul-nerability to toxic insults may represent one pathway through which environmental exposure can contribute to accumulation of AD pathology and increased risk for dementia.

Methods

Participants

Participants were from wave 3 of the longitudinal Vietnam Era Twin Study of Aging (VETSA),52 which took place from 2016 to 2019. VETSA participants were recruited from a previous study53 of members of the Vietnam Era Twin Registry, which comprises a national, community-dwelling study of American male twin pairs from all 50 states who served in the military at some point between 1965 and 1975, though most (? 80%) were not combat-exposed. VETSA participants are similar to the general popula-tion of American men in their age range with respect to health, education, and lifestyle characteristics based on Center for Disease Control and Prevention data.54 Participants included in the present study were limited to those that lived in the contigu-ous United States, as these were the states for which air pollution estimates were available. The MRI scans were acquired as part of a substudy of the parent VETSA. The MRI study began partway through the first wave of VETSA data collection, and the partici-pants receiving scans were then included in subsequent waves of the MRI study. Thus, MRI participants were not selected based on any characteristics other than time of enrollment and are rep-resentative of the overall VETSA sample.

Classification of MCI was applied at the time of LC imaging (2016-2019) and determined using the Jak-Bondi approach55 as described previously in more detail.56 Briefly, individuals com-pleted 18 cognitive tests across the domains of memory, executive function, attention, language, visuospatial ability, and processing speed. Criteria for impairment required performance on 2 or more tests within a domain falling more than 1.5 standard deviations (SDs) below age- and education-adjusted norms. Individuals with dementia were excluded from the current analysis, and dementia diagnosis was determined based on self-reported diagnosis from a doctor. Participants were included in the current analysis if they completed both a three-dimensional (3D) T1-weighted MRI sequence and an LC-sensitive MRI sequence described below, had air pollution exposure data available during the most recent period for which these data were collected on the VETSA cohort (an average of 5.6 years prior to the MRI scan), and had complete data for all covariates in the primary analysis of PM2:5 exposure and LCCNR. This resulted in a final analytic sample of 381 individuals with complete data.

The study was approved by the institutional review board at the University of California, San Diego (UCSD), and written informed consent was obtained from all study participants.

Air Pollution Measures

For the current analyses, we included 3-year average estimates of PM2:5, nitrogen dioxide (NO2), and ozone (O3) exposure during the most recent period that air pollution data was available for the VETSA cohort. This time period was defined as the 3 years prior to each individual's wave 2 VETSA assessment, which occurred between 2009 and 2013. Our primary focus is on PM2:5, but other exposure to NO2 and O3 are included to assess specificity of effects. The methods used to obtain ambient air pollution exposure for the VETSA cohort have been described in detail previously57 and are briefly summarized here. Address histories were used to geocode residential locations at the level of five decimal places. All geocodes were provided to the University of Washington Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air) group for estima-tion of time-resolved participant-specific air pollution exposures. The MESA Air statistical prediction of air pollution concentrations has been described in detail previously.58 Briefly, the fine scale national spatiotemporal model combines land-use regression (LUR) techniques in a universal kriging framework accounting for temporal trends to provide estimates of the value of a given pollu-tant over a continuous spatial field based on data from a limited sampling of geographic data (i.e., individual air pollution monitors). The model makes use of up to 400 geographic covariates based on geographic information systems (GIS) data as well as seasonal and systematic trends in emission sources, population, land use, and near-source concentrations that vary over space and time. The initial MESA Air models were developed for specific communities but have been extended to obtain predictions across the continental United States. The country was split into nine climatic/topographic regions (for PM2:5 model) or three regions (for O3 and NO2) to capture subnational region-specific pollution processes contributing to national estimates and ensure each region contained supplemen-tal monitors. The nationwide model includes data from ~940 investigator-deployed monitors and 1,500 regulatory monitors to produce estimates for PM2:5, NO2, and O3 at 2-week timescales. Each participant's exposure to PM2:5, NO2, and O3 was calculated as the average across the three-year period immediately prior to their wave 2 VETSA assessment. National estimates have been cross-validated with an R2 of 0.89 for PM2:5 and 0.87 for NO2.

MRIAcquisition

LC imaging was introduced during wave 3 of VETSA. MRI scans were acquired from 2016 to 2019 during the VETSA wave 3 assessments, which occurred an average of 5.64 years (SD = 0:54) after the wave 2 assessments that constituted the end point of air pollution exposure periods. Images were collected using two GE 3T Discovery 750× scanners (GE Health care) with eight-channel phased array head coils. The imaging protocol included a sagittal 3D fast spoiled gradient echo (FSPGR) T1-weighted (T1w) volume optimized for maximum gray/white contrast (Echo Time = 3:164 ms, Repetition Time = 8:084 ms, TI = 600 ms, flip angle = 8°, matrix = 256 × 192, in-plane resolution = 1 × 1 mm, slice thickness = 1:2 mm, and slices=172). The LC was imaged with an axial fast spin-echo T1-weighted image59 (Repetition Time = 600 ms, Echo Time = 14 ms, flip angle = 90°, matrix = 512 × 320, Field of view = 220 mm, pixel size = 0:42 × 0:68 mm, 8 slices, slice thickness = 2:5 mm, and interslice gap = 1 mm). Slices were acquired perpendicular to the lateral floor of the fourth ventricle, with the rostral-most slice positioned on the in-ferior third of the inferior colliculus.38

MRIProcessing

Our measure of LC integrity was calculated as described previ-ously.37,38 Briefly, two raters blind to PM2:5 exposure data manually marked regions of interest (ROIs) on three axially oriented slices using anatomical landmark-based rules to derive signal intensities corresponding to the rostral, middle, and caudal LC. To control overall signal intensity variability across slices, a 10-mm2 reference ROI was placed in the pontine tegmentum (PT). LC contrast-to-noise ratio (LCcnr) values were calculated for each slice as

LCcnr = (LCintensity - PTintensity)=PTintensity Higher LCcnr values

are thought to reflect better LC structural integrity due to correspon-dence between postmortem MRI-assessed LCcnr and histological examination of LC integrity at the neuronal level.60 We focus on a measure of rostral-middle LCcnr, defined as the average LCcnr from the rostral and middle slices, because extant literature suggests that the rostral and middle portions of the LC are more prone to age- and AD-related degeneration than more caudal, cerebellar-projecting portions.36,44-46 We also examined LCcnr from the caudal slice in separate sensitivity analyses to investigate regional specificity of associations within the LC. Final values for rostral-middle and caudal LCcnr were calculated by averaging corresponding LCcnr values across the two raters. We previously reported high interrater reliability on these markings.38

In addition to using caudal LCCNR as a comparison structure, hippocampal volume was examinedto further assess regional spec-ificity of the link between PM2:5 and brain structural integrity. T1-weighted structural MR images were processed with the FreeSurfer version 6.0 software package61-63 to obtain measures of left and right hippocampal volume and estimated intracranial volume (ICV). Segmentations for each hemisphere were visually reviewed for quality, and segmentation failure (i.e., highly inaccurate delineation of the hippocampus) were excluded from analyses. A bilateral average of hippocampal volume was used for analysis.

Statistical Analysis

Associations between PM2:5 exposures and LC integrity were tested using linear mixed effects models implemented with the lme4 package64 in R version 4.1.3 (R Development Core Team). LCCNR was used as the dependent variable with PM2:5 as the independent variable. Covariates were chosen based on prior literature and reflect factors that may be correlated with air pollution expo-sure and/or LCCNR.38,41,48,57 Covariates included fixed effects of scanner (one level for each of the two scanners) and age at time of MRI scan, as well as self-reported years of formal education completed and household income (coded as 1 = < $30,000, 2 = $30,000-49,000, 3 =$50,000-79,000, 4=$80,000-99,000, 5 = > $100,000) during the period that air pollution was measured. The interval (in years) between the end point of the air pollution ex-posure period and the MRI scan was also included as a covariate. Age, education, and interval between air pollution assessment and MRI scan were modeled as continuous covariates, household income was modeled as an ordinal factor, and scanner was modeled as a two-level factor. Family ID was included as a random intercept to account for the nonindependence of twin pairs. Continuous variables were standardized by the sample mean and standard deviation (standardized to mean = 0 and SD = 1), and standardized beta coeffi-cients are presented in the results. Degrees of freedom, t-statistics, and significance were calculated using the Satterthwaite method as implemented in the R package lmerTest.65 All analyses included only participants with complete data for all relevant measures.

We then conducted several sensitivity analyses. First, we investi-gated whether associations differed by region. The place of residence at which air pollution was estimated for each participant was catego-rized into one of four regions: Midwest, Northeast, South, and West. We used a likelihood ratio test to determine whether including an interaction between PM2:5 exposure and region improved the model fit. Second, we examined the association between PM2:5 and LCCNR after excluding individuals with MCI to determine whether associa-tions were present in unimpaired individuals or driven by those with impairment. Third, we explored whether there were associations between the other assessed pollutants and LCCNR. In two separate models, we used either NO2 or O3 as the primary predictor in place of PM2:5. All other covariates remained the same. We then investi-gated the specificity of associations with the rostral-middle LC by testing the same models (including participants with MCI) but using caudal LCCNR as the outcome measure. To determine whether effects were apparent in other brain regions highly relevant to AD, we reran models with bilateral hippocampal volume as the outcome. Bilateral hippocampal volume was first residualized for estimated ICV to account for effects of head size.

Next, we leveraged the twin design of the VETSA sample using a co-twin control analysis of monozygotic (MZ) twins. These models can provide added specificity for associations of interest because MZ twins are matched for their age, genetic background, and rear-ing environment, ruling out many potential confounding factors. We can decompose the associations between a predictor and outcome variable into "within-pair" and "between-pair" effects using the following formula66:

where Yij is each individual ij's LCcnr value, Xi is the mean level of PM2:5 exposure for a twin pair i, (Xij -Xi) is the deviation of each individual j's exposure to PM2:5 from the pair i mean, and Covariatesij are the values of covariates for individual j. The coefficient b W represents "within-pair" effects and can be inter-preted as the association between individual-specific variation in PM2:5-that is, the variation that is unique to each individual of the twin pair-and LCcnr. The coefficient bB represents "between-pair" effects and can be interpreted as the association between variation in PM2:5 exposure common to both individuals in a pair-such as contributions from genetics or shared early childhood environmental influence-and LCcnr. This analysis included only complete MZ pairs in which each individual had complete data (n = 284 individuals, 142 pairs).

Results

Participants were predominantly non-Hispanic white (88.3%) with ages ranging from 62 to 72 years (mean = 67:3; SD = 2:6) at the time of the MRI scan. The interval between VETSA wave 2 (which represents the end of the 3-year time periods during which air pol-lution exposure was assessed) and wave 3 assessments (during which MRI scans were acquired) was a mean of 5.6 years (SD = 0:54 years; range: 3.29-7.59 years). The average education level of the sample was 14.0 (SD = 2:1) years. Fifty-four (13.7%) individuals were diagnosed with MCI at the time of the MRI scan. See Table 1 for a table of sample characteristics. The average level of PM2:5 exposure in the time period assessed was 8:99 ug=m3 (SD = 2:10; range: 2:55-12:28 ug=m3; 25th-75th percentile: 7:57-10:52 ug=m3). The average rostral-middle LCcnr was 0.11 (SD = 0:03; range: 0.03-0.22; 25th-75th percentile: 0.09-0.13). PM2:5 exposures were not significantly related to lifetime education (Pearson's r= -0:01; p = 0:69) or concurrent household income (Spearman's p = 0:03;p = 0:31).

There was a significant association between PM2:5 and rostral-middle LCcnr whereby higher exposure to PM2:5 was associated with lower LCcnr [b= -0:12; 95% confidence interval (CI): -0:22, -0:02; p = 0:021] (Table 2; Table S1; Figure 1). There were no associations of rostral-middle LCcnr with age at time of MRI, household income, or education. Adding an interaction between PM2:5 exposure and region did not improve model fit [%2(3) = 3:99; p = 0:263], indicating that the association did not differ by region. The effect size of this association remained of a similar magnitude when excluding the 54 individuals with MCI (b= -0:13; 95% CI: -0:24, -0:03; p = 0:014) (Table 2; Table S2). The associations of NO2 and O3 with rostral-middle LCCNR in the full sample were much weaker and nonsignificant [NO2: b = - 0:01; 95% CI: -0:11, 0.09; p = 0:896 (Table 2; Table S3); O3: b = - 0:01; 95% CI: -0:11, 0.09; p = 0:896 (Table 2; Table S4)]. There were no associations between PM2:5 and caudal LCCNR (b = - 0:01; 95% CI: -0:11, 0.09; p = 0:896) (Table 2; Table S5) or hippocampal volume (b= 0:10; 95% CI: 0.00, 0.20; p = 0:061) (Table 2; Table S6). The association between PM2:5 exposure and LCCNR did not differ by region (F1,282 = 1:286; p = 0:28) (Table S7).

In the co-twin control analysis, there was a significant within-pair (i.e., unique environmental) effect such that individuals within a pair who experienced higher PM2:5 exposure had lower LCCNR (b= -0:11; 95% CI: -0:19, -0:03; p= 0:007) (Table 2; Table S8), although the point estimate was somewhat smaller than that seen in the full sample (b= -0:11 vs. -0:16). The between-pair (i.e., familial) effect was not significant (b = - 0:08; 95% CI:

-0:25, -0:10; p= 0:383), with the point estimate being about half that seen in the full sample. Although the within-pair effect was stronger than the between-pair effect, we note that the confidence intervals were highly overlapping.

Discussion

PM2:5 is a major component of ambient air pollution that is associated with increased risk for AD.67,68 Here, we found that higher levels of exposure to PM2:5 were associated with lower integrity of the rostral-middle LC, which is a very early site of AD-related pathology. Although AD is uncommon in the age range of our participants, a growing body of literature shows that midlife bio-markers are robust predictors of late-life health outcomes.69-71 The finding of an association between air pollution and LC integ-rity in late midlife provides additional evidence of the deleterious effects of exposure-related brain damage in a region that is vul-nerable to early AD pathology1,72 and thus relevant to later life dementia risk. Additionally, we found that, among MZ twin pairs, individuals with higher exposure to PM2:5 showed lower LCCNR, providing evidence that this association is not driven by confounders such as genetic background or environmental influ-ences shared between siblings.

In this study, the effects of PM2:5 exposure were specific to the rostral-middle LC. This is consistent with prior imaging38,42,45,46 and autopsy43,46 studies showing that the rostral-middle LC was preferentially associated with age- or disease-related brain changes and cognitive decline compared to the caudal LC, which shows rel-atively little or no such associations. We may detect associations between air pollution exposure and the rostral-middle LC prior to other brain structures due to its location, connectivity, and struc-tural characteristics. For example, inhaled toxicants may reach the LC through projections from the olfactory bulb or its extensive interface with the circulatory system.22,24,25 Damage to the LC promotes the accumulation of AD pathology and exacerbates the inflammatory response,9,10,13 which may represent one pathway through which exposure to air pollution contributes to AD risk. Consistent with prior studies, we did not find an association between PM2:5 and hippocampal volume,73-75 although other studies have.47 Studies finding associations of PM2:5 with macrostruc-tural measures of brain integrity (e.g., brain volume or cortical thickness) in adults have typically included older participants than those in the VETSA sample. In vivo studies of humans and experi-mental rat models provide evidence that reduction in LC integrity may occur relatively early in the pathophysiological process.14,51 Therefore, it is possible that associations with hippocampal volume may become apparent in this sample as individuals reach older ages. Although we did find an association between PM2:5 and rostral-middle LCCNR, we did not find associations with NO2 or O3. A prior review of the literature found that PM2:5 was more often associated with brain and cognitive outcomes than other pollutants, but it was noted that this may be because PM2:5 is more frequently assessed.20

There are several limitations to note about the current study. First, with only one timepoint of LC data, we cannot assess whether variation in LC integrity reflects an actual decline from previous levels or determine when the association between PM2:5 and LC integrity manifested (e.g., this association may only be apparent in later life, or it may be apparent at younger ages). However, our co-twin control analyses rule out genetic confound-ing or shared environmental influences between twins (e.g., rear-ing environment or common familial influences) driving the association. Second, we cannot determine which components of PM2:5 are driving these associations, and it may be highly vari-able across geographical regions. Obtaining levels of the constituent components will be an important future direction of study. Third, the current study includes only men, so we cannot speak to any putatively adverse effect of exposures on LC integrity in women. However, other studies have found associations between air pollution effects and increased AD risk in samples of older women.17,48,76,77 Fourth, these results may not generalize outside the levels of PM2:5 observed in this sample. The US has relatively clean air, and it is possible that the association of LCCNR and PM2:5 is nonlinear such that effects could be weaker or stronger at different ranges of PM2:5 exposure.

Our study also has several strengths. There is no published work using this combination of geographically specific air pollu-tion exposure data and LC-sensitive imaging measures. LCCNR may provide a more sensitive measure of subtle variation in brain structural integrity than more common macrostructural measures such as hippocampal volume or cortical thickness.41 The VETSA includes individuals from various urban and rural geographic regions across the entire US. Furthermore, VETSA participants are representative of American men in their age cohort with respect to health, education, and lifestyle characteristics.54 Finally, the twin design of the study allowed us to leverage MZ pairs as natural controls to reduce the impact of potentially con-founding genetic or environmental factors.

Conclusions

In conclusion, the current study provides evidence that exposure to PM2:5 may impact integrity of the LC. Prior work in animal models has demonstrated that damage to or dysfunction of the LC has wide-ranging effects on inflammation and AD pathologi-cal processes that contribute to disease progression,9-12,78,79 and human studies indicate that loss of LC integrity predicts later pathological accumulation.51 Thus, loss of LC integrity may represent a potential pathway through which exposure to air pollu-tion may increase AD risk. Experimental studies are needed to validate these findings and determine the causal mechanisms underlying this relationship.

Acknowledgments

Numerous organizations provided invaluable assistance in the conduct of the VET Registry, including the following: US Department of Veterans Affairs, Department of Defense; National Personnel Records Center, National Archives and Records Administration; Internal Revenue Service; National Opinion Research Center; National Research Council, National Academy of Sciences; the Institute for Survey Research, Temple University. The authors gratefully acknowledge the continued cooperation of the twins and the efforts of many staff members.

This work was supported by the National Institute on Aging at the National Institutes of Health (grant numbers P01 AG055367; R01s AG076838, AG022381, AG050595, and AG064955; and K01 AG063805).

The content of this manuscript is the responsibility of the authors and does not represent official views of NIA/NIH or the Veterans' Administration.

References

References

1. Braak H, Del Tredici K. 2011. The pathological process underlying Alzheimer's disease in individuals under thirty. Acta Neuropathol 121(2):171-181, PMID: 21170538, https://doi.org/10.1007/s00401-010-0789-4.

2. Braak H, Thal DR, Ghebremedhin E, Del Tredici K. 2011. Stages of the patho-logic process in Alzheimer disease: age categories from 1 to 100 years. J Neuropathol Exp Neurol 70(11):960-969, PMID: 22002422, https://doi.org/10. 1097/NEN.0b013e318232a379.

3. Kelly SC, He B, Perez SE, Ginsberg SD, Mufson EJ, Counts SE. 2017. Locus coeru-leus cellular and molecular pathology during the progression of Alzheimer's dis-ease. Acta Neuropathol Commun 5(1):8, PMID: 28109312, https://doi.org/10.1186/ s40478-017-0411-2.

4. Bondareff W, Mountjoy CQ, Roth M, Rossor MN, Iversen LL, Reynolds GP, et al. 1987. Neuronal degeneration in locus ceruleus and cortical correlates of Alzheimer disease. Alzheimer Dis Assoc Disord 1(4):256-262, PMID: 3453748, https://doi.org/10.1097/00002093-198701040-00005.

5. Wilson RS, Nag S, Boyle PA, Hizel LP, Yu L, Buchman AS, et al. 2013. Neural reserve, neuronal density in the locus ceruleus, and cognitive decline. Neurology 80(13):1202-1208, PMID: 23486878, https://doi.org/10.1212/WNL.0b013e3182897103.

6. Theofilas P, Ehrenberg AJ, Dunlop S, Di Lorenzo Alho AT, Nguy A, Leite REP, et al. 2017. Locus coeruleus volume and cell population changes during Alzheimer's disease progression: a stereological study in human postmortem brains with potential implication for early-stage biomarker discovery. Alzheimers Dement 13(3):236-246, PMID: 27513978, https://doi.org/10.1016/j.jalz.2016.06.2362.

7. Murray ME, Moloney CM, Kouri N, Syrjanen JA, Matchett BJ, Rothberg DM, et al. 2022. Global neuropathologic severity of Alzheimer's disease and locus coeruleus vulnerability influences plasma phosphorylated tau levels. Mol Neurodegener 17(1):85, PMID: 36575455, https://doi.org/10.1186/s13024-022-00578-0.

8. Berridge CW, Waterhouse BD. 2003. The locus coeruleus-noradrenergic system: modulation of behavioral state and state-dependent cognitive processes. Brain Res Brain Res Rev 42(1):33-84, PMID: 12668290, https://doi.org/10.1016/ s0165-0173(03)00143-7.

9. Heneka MT, Galea E, Gavriluyk V, Dumitrescu-Ozimek L, Daeschner J, O'Banion MK, et al. 2002. Noradrenergic depletion potentiates beta-amyloid-induced cortical inflammation: implications for Alzheimer's disease. J Neurosci 22(7):2434-2442, PMID: 11923407, https://doi.org/10.1523/JNEUROSCI. 22-07-02434.2002.

10. Heneka MT, Nadrigny F, Regen T, Martinez-Hernandez A, Dumitrescu-Ozimek L, Terwel D, et al. 2010. Locus ceruleus controls Alzheimer's disease pathology by modulating microglial functions through norepinephrine. Proc Natl Acad Sci USA 107(13):6058-6063, PMID: 20231476, https://doi.org/10.1073/pnas.0909586107.

11. Jardanhazi-Kurutz D, Kummer MP, Terwel D, Vogel K, Dyrks T, Thiele A, et al. 2010. Induced LC degeneration in APP/PS1 transgenic mice accelerates early cerebral amyloidosis and cognitive deficits. Neurochem Int 57(4):375-382, PMID: 20144675, https://doi.org/10.1016/j.neuint.2010.02.001.

12. Jardanhazi-Kurutz D, Kummer MP, Terwel D, Vogel K, Thiele A, Heneka MT. 2011. Distinct adrenergic system changes and neuroinflammation in response to induced locus ceruleus degeneration in APP/PS1 transgenic mice. Neuroscience 176:396-407, PMID: 21129451, https://doi.org/10.1016/j.neuroscience.2010.11.052.

13. Weinshenker D. 2018. Long road to ruin: noradrenergic dysfunction in neurodege-nerative disease. Trends Neurosci 41(4):211-223, PMID: 29475564, https://doi.org/ 10.1016/j.tins.2018.01.010.

14. Flores-Aguilar L, Hall H, Orciani C, Foret MK, Kovecses O, Ducatenzeiler A, et al. 2022. Early loss of locus coeruleus innervation promotes cognitive and neu-ropathological changes before amyloid plaque deposition in a transgenic rat model of Alzheimer's disease. Neuropathol Appl Neurobiol 48(6):e12835, PMID: 35822518, https://doi.org/10.1111/nan.12835.

15. Livingston G, Huntley J, Sommerlad A, Ames D, Ballard C, Banerjee S, et al. 2020. Dementia prevention, intervention, and care: 2020 report of the Lancet commission. Lancet 396(10248):413-446, PMID: 32738937, https://doi.org/10. 1016/S0140-6736(20)30367-6.

16. Thangavel P, Park D, Lee YC. 2022. Recent insights into particulate matter (PM2.5)-mediated toxicity in humans: an overview. Int J Environ Res Public Health 19(12):7511, PMID: 35742761, https://doi.org/10.3390/ijerph19127511.

17. Wang X, Younan D, Petkus AJ, Beavers DP, Espeland MA, Chui HC, et al. 2021. Ambient air pollution and long-term trajectories of episodic memory decline among older women in the WHIMS-ECHO cohort. Environ Health Perspect 129(9):097009, PMID: 34516296, https://doi.org/10.1289/EHP7668.

18. Alemany S, Crous-Bou M, Vilor-Tejedor N, Milà-Alomà M, Suárez-Calvet M, Salvadó G, et al. 2021. Associations between air pollution and biomarkers of Alzheimer's disease in cognitively unimpaired individuals. Environ Int 157:106864, PMID: 34537521, https://doi.org/10.1016/j.envint.2021.106864.

19. Iaccarino L, La Joie R, Lesman-Segev OH, Lee E, Hanna L, Allen IE, et al. 2020. Association between ambient air pollution and amyloid positron emission to-mography positivity in older adults with cognitive impairment. JAMA Neurol 78(2):197-207, PMID: 33252608, https://doi.org/10.1001/jamaneurol.2020.3962.

20. Power MC, Adar SD, Yanosky JD, Weuve J. 2016. Exposure to air pollution as a potential contributor to cognitive function, cognitive decline, brain imaging, and dementia: a systematic review of epidemiologic research. Neurotoxicology 56:235-253, PMID: 27328897, https://doi.org/10.1016/j.neuro.2016.06.004.

21. Ma Y-H, Chen H-S, Liu C, Feng Q-S, Feng L, Zhang Y-R, et al. 2023. Association of long-term exposure to ambient air pollution with cognitive decline and Alzheimer's disease-related amyloidosis. Biol Psychiatry 93(9):780-789, PMID: 35953319, https://doi.org/10.1016/j.biopsych.2022.05.017.

22. Pamphlett R. 2014. Uptake of environmental toxicants by the locus ceruleus: a potential trigger for neurodegenerative, demyelinating and psychiatric disorders. Med Hypotheses 82(1):97-104, PMID: 24315447, https://doi.org/10.1016/j. mehy.2013.11.016.

23. Matchett BJ, Grinberg LT, Theofilas P, Murray ME. 2021. The mechanistic link between selective vulnerability of the locus coeruleus and neurodegeneration in Alzheimer's disease. Acta Neuropathol 141(5):631-650, PMID: 33427939, https://doi.org/10.1007/s00401-020-02248-1.

24. Chen X, Guo J, Huang Y, Liu S, Huang Y, Zhang Z, et al. 2020. Urban airborne PM(2.5)-activated microglia mediate neurotoxicity through glutaminase-containing extracellular vesicles in olfactory bulb. Environ Pollut 264:114716, PMID: 32559876, https://doi.org/10.1016/j.envpol.2020.114716.

25. Calderón-Garcidueñas L, Franco-Lira M, Henríquez-Roldán C, Osnaya N, González-Maciel A, Reynoso-Robles R, et al. 2010. Urban air pollution: influen-ces on olfactory function and pathology in exposed children and young adults. Exp Toxicol Pathol 62(1):91-102, PMID: 19297138, https://doi.org/10.1016/j.etp. 2009.02.117.

26. Cheng H, Saffari A, Sioutas C, Forman HJ, Morgan TE, Finch CE. 2016. Nanoscale particulate matter from urban traffic rapidly induces oxidative stress and inflam-mation in olfactory epithelium with concomitant effects on brain. Environ Health Perspect 124(10):1537-1546, PMID: 27187980, https://doi.org/10.1289/EHP134.

27. Chang LT, Tang CS, Pan YZ, Chan CC. 2007. Association of heart rate variability of the elderly with personal exposure to PM1, PM1-2.5, and PM2.5-10. Bull Environ Contam Toxicol 79(5):552-556, PMID: 17639313, https://doi.org/10.1007/ s00128-007-9233-4.

28. Riojas-Rodríguez H, Escamilla-Cejudo JA, González-Hermosillo JA, Téllez-Rojo MM, Vallejo M, Santos-Burgoa C, et al. 2006. Personal PM2.5 and CO expo-sures and heart rate variability in subjects with known ischemic heart disease in Mexico city. J Expo Sci Environ Epidemiol 16(2):131-137, PMID: 16175197, https://doi.org/10.1038/sj.jea.7500453.

29. Bind MA, Peters A, Koutrakis P, Coull B, Vokonas P, Schwartz J. 2016. Quantile regression analysis of the distributional effects of air pollution on blood pres-sure, heart rate variability, blood lipids, and biomarkers of inflammation in el-derly American men: the normative aging study. Environ Health Perspect 124(8):1189-1198, PMID: 26967543, https://doi.org/10.1289/ehp.1510044.

30. Femminella GD, Rengo G, Komici K, Iacotucci P, Petraglia L, Pagano G, et al. 2014. Autonomic dysfunction in Alzheimer's disease: tools for assessment and review of the literature. J Alzheimers Dis 42(2):369-377, PMID: 24898649, https://doi.org/10.3233/JAD-140513.

31. Weinstein G, Davis-Plourde K, Beiser AS, Seshadri S. 2021. Autonomic imbalance and risk of dementia and stroke: the Framingham study. Stroke

52(6):2068-2076, PMID: 33874747, https://doi.org/10.1161/STROKEAHA.120. 030601.

32. Samuels ER, Szabadi E. 2008. Functional neuroanatomy of the noradrenergic locus coeruleus: its roles in the regulation of arousal and autonomic function part I: principles of functional organisation. Curr Neuropharmacol 6(3):235-253, PMID: 19506723, https://doi.org/10.2174/157015908785777229.

33. Mather M, Joo Yoo H, Clewett DV, Lee T-H, Greening SG, Ponzio A, et al. 2017. Higher locus coeruleus MRI contrast is associated with lower parasympa-thetic influence over heart rate variability. Neuroimage 150:329-335, PMID: 28215623, https://doi.org/10.1016/j.neuroimage.2017.02.025.

34. Li H, Cai J, Chen R, Zhao Z, Ying Z, Wang L, et al. 2017. Particulate matter expo-sure and stress hormone levels: a randomized, double-blind, crossover trial of air purification. Circulation 136(7):618-627, PMID: 28808144, https://doi.org/10. 1161/CIRCULATIONAHA.116.026796.

35. Heusser K, Tank J, Holz O, May M, Brinkmann J, Engeli S, et al. 2019. Ultrafine particles and ozone perturb norepinephrine clearance rather than centrally generated sympathetic activity in humans. Sci Rep 9(1):3641, PMID: 30842540, https://doi.org/10.1038/s41598-019-40343-w.

36. Betts MJ, Kirilina E, Otaduy MCG, Ivanov D, Acosta-Cabronero J, Callaghan MF, et al. 2019. Locus coeruleus imaging as a biomarker for noradrenergic dysfunc-tion in neurodegenerative diseases. Brain 142(9):2558-2571, PMID: 31327002, https://doi.org/10.1093/brain/awz193.

37. Clewett DV, Lee TH, Greening S, Ponzio A, Margalit E, Mather M. 2016. Neuromelanin marks the spot: identifying a locus coeruleus biomarker of cog-nitive reserve in healthy aging. Neurobiol Aging 37:117-126, PMID: 26521135, https://doi.org/10.1016/j.neurobiolaging.2015.09.019.

38. Elman JA, Puckett OK, Beck A, Fennema-Notestine C, Cross LK, Dale AM, et al. 2021. MRI-assessed locus coeruleus integrity is heritable and associated with multiple cognitive domains, mild cognitive impairment, and daytime dysfunc-tion. Alzheimers Dement 17(6):1017-1025, PMID: 33580733, https://doi.org/10. 1002/alz.12261.

39. Dahl MJ, Mather M, Düzel S, Bodammer NC, Lindenberger U, Kühn S, et al.

2019. Rostral locus coeruleus integrity is associated with better memory performance in older adults. Nat Hum Behav 3(11):1203-1214, PMID: 31501542, https://doi.org/10.1038/s41562-019-0715-2.

40. Bachman SL, Dahl MJ, Werkle-Bergner M, Düzel S, Forlim CG, Lindenberger U, et al. 2021. Locus coeruleus MRI contrast is associated with cortical thickness in older adults. Neurobiol Aging 100:72-82, PMID: 33508564, https://doi.org/10. 1016/j.neurobiolaging.2020.12.019.

41. Elman JA, Puckett OK, Hagler DJ, Pearce RC, Fennema-Notestine C, Hatton SN, et al. 2022. Associations between MRI-assessed locus coeruleus integrity and cortical gray matter microstructure. Cereb Cortex 32(19):4191-4203, PMID: 34969072, https://doi.org/10.1093/cercor/bhab475.

42. Bell TR, Elman JA, Beck A, Fennema-Notestine C, Gustavson DE, Hagler DJ, et al. 2022. Rostral-Middle locus coeruleus integrity and subjective cognitive decline in early old age. J Int Neuropsychol Soc 29(8):763-774, https://doi.org/ 10.1017/S1355617722000881.

43. Jacobs HIL, Becker JA, Kwong K, Engels-Domínguez N, Prokopiou PC, Papp KV, et al. 2021. In vivo and neuropathology data support locus coeruleus integrity as indicator of Alzheimer's disease pathology and cognitive decline. Sci Transl Med 13(612):eabj2511, PMID: 34550726, https://doi.org/10.1126/scitranslmed.abj2511.

44. German DC, Manaye KF, White CL, Woodward DJ, McIntire DD, Smith WK, et al. 1992. Disease-specific patterns of locus coeruleus cell loss. Ann Neurol 32(5):667-676, PMID: 1449247, https://doi.org/10.1002/ana.410320510.

45. Liu KY, Acosta-Cabronero J, Cardenas-Blanco A, Loane C, Berry AJ, Betts MJ, et al. 2019. In vivo visualization of age-related differences in the locus coeru-leus. Neurobiol Aging 74:101-111, PMID: 30447418, https://doi.org/10.1016/j. neurobiolaging.2018.10.014.

46. Van Egroo M, Riphagen JM, Ashton NJ, Janelidze S, Sperling RA, Johnson KA, et al. 2023. Ultra-high field imaging, plasma markers and autopsy data uncover a specific rostral locus coeruleus vulnerability to hyperphosphorylated tau. Mol Psychiatry 28(6):2412-2422, PMID: 37020050, https://doi.org/10.1038/s41380-023-02041-y.

47. Balboni E, Filippini T, Crous-Bou M, Guxens M, Erickson LD, Vinceti M. 2022. The association between air pollutants and hippocampal volume from mag-netic resonance imaging: a systematic review and meta-analysis. Environ Res 204(pt A):111976, PMID: 34478724, https://doi.org/10.1016/j.envres.2021.111976.

48. Younan D, Wang X, Casanova R, Barnard R, Gaussoin SA, Saldana S, et al.

2020. PM2.5 associated with gray matter atrophy reflecting increased Alzheimers risk in older women. Neurology 96(8):e1190-e1201, PMID: 33208540, https://doi.org/10.1212/WNL.0000000000011149.

49. Calderón-Garcidueñas L, Hernández-Luna J, Mukherjee PS, Styner M, Chávez-Franco DA, Luévano-Castro SC, et al. 2022. Hemispheric cortical, cerebellar and caudate atrophy associated to cognitive impairment in metropolitan Mexico city young adults exposed to fine particulate matter air pollution. Toxics 10(4):156, PMID: 35448417, https://doi.org/10.3390/toxics10040156.

50. de Prado Bert P, Mercader EMH, Pujol J, Sunyer J, Mortamais M. 2018. The effects of air pollution on the brain: a review of studies interfacing environmen-tal epidemiology and neuroimaging. Curr Environ Health Rep 5(3):351-364, PMID: 30008171, https://doi.org/10.1007/s40572-018-0209-9.

51. Jacobs HIL, Becker JA, Kwong K, Munera D, Ramirez-Gomez L, Engels-Domínguez N, et al. 2023. Waning locus coeruleus integrity precedes cortical tau accrual in preclinical autosomal dominant Alzheimer's disease. Alzheimers Dement 19(1):169-180, PMID: 35298083, https://doi.org/10.1002/alz.12656.

52. Kremen WS, Franz CE, Lyons MJ. 2019. Current status of the Vietnam Era Twin Study of Aging (VETSA). Twin Res Hum Genet 22(6):783-787, PMID: 31933447, https://doi.org/10.1017/thg.2019.125.

53. Tsuang MT, Bar JL, Harley RM, Lyons MJ. 2001. The Harvard Twin Study of Substance Abuse: what we have learned. Harv Rev Psychiatry 9(6):267-279, PMID: 11600486.

54. Schoenborn CA, Heyman KM. 2009. Health characteristics of adults aged 55 years and over: United States, 2004-2007. Natl Health Stat Report 16(16):1-31, PMID: 19697804.

55. Jak AJ, Bondi MW, Delano-Wood L, Wierenga C, Corey-Bloom J, Salmon DP, et al. 2009. Quantification of five neuropsychological approaches to defining mild cognitive impairment. Am J Geriatr Psychiatry 17(5):368-375, PMID: 19390294, https://doi.org/10.1097/JGP.0b013e31819431d5.

56. Kremen WS, Jak AJ, Panizzon MS, Spoon KM, Franz CE, Thompson WK, et al. 2014. Early identification and heritability of mild cognitive impairment. Int J Epidemiol 43(2):600-610, PMID: 24370560, https://doi.org/10.1093/ije/dyt242.

57. Franz CE, Gustavson DE, Elman JA, Fennema-Notestine C, Hagler DJ, Baraff A, et al. 2023. Associations between ambient air pollution and cognitive abilities from midlife to early old age: modification by APOE genotype. J Alzheimers Dis 93(1):193-209, PMID: 36970897, https://doi.org/10.3233/JAD-221054.

58. Kirwa K, Szpiro AA, Sheppard L, Sampson PD, Wang M, Keller JP, et al. 2021. Fine-scale air pollution models for epidemiologic research: insights from approaches developed in the multi-ethnic study of atherosclerosis and air pol-lution (MESA air). Curr Environ Health Rep 8(2):113-126, PMID: 34086258, https://doi.org/10.1007/s40572-021-00310-y.

59. Sasaki M, Shibata E, Tohyama K, Takahashi J, Otsuka K, Tsuchiya K, et al. 2006. Neuromelanin magnetic resonance imaging of locus ceruleus and sub-stantia nigra in Parkinson's disease. Neuroreport 17(11):1215-1218, PMID: 16837857, https://doi.org/10.1097/01.wnr.0000227984.84927.a7.

60. Keren NI, Taheri S, Vazey EM, Morgan PS, Granholm A-CE, Aston-Jones GS, et al. 2015. Histologic validation of locus coeruleus MRI contrast in post-mortem tissue. Neuroimage 113:235-245, PMID: 25791783, https://doi.org/10.1016/j. neuroimage.2015.03.020.

61. Fischl B. 2012. FreeSurfer. Neuroimage 62(2):774-781, PMID: 22248573, https://doi.org/10.1016/j.neuroimage.2012.01.021.

62. Fischl B, Salat DH, Busa E, Albert M, Dieterich M, Haselgrove C, et al. 2002. Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron 33(3):341-355, PMID: 11832223, https://doi.org/10. 1016/S0896-6273(02)00569-X.

63. Fischl B, Salat DH, van der Kouwe AJW, Makris N, Ségonne F, Quinn BT, et al. 2004. Sequence-independent segmentation of magnetic resonance images. Neuroimage 23(suppl 1):S69-S84, PMID: 15501102, https://doi.org/10.1016/j. neuroimage.2004.07.016.

64. Bates D, Machler M, Bolker BM, Walker SC. 2015. Fitting linear mixed-effects models using lme4. J Stat Softw 67(1):1-48, https://doi.org/10.18637/jss.v067.i01.

65. Kuznetsova A, Brockhoff PB, Christensen RHB. 2017. lmerTest package: tests in linear mixed effects models. J Stat Soft 82(13):1-26, https://doi.org/10.18637/ jss.v082.i13.

66. Carlin JB, Gurrin LC, Sterne JA, Morley R, Dwyer T. 2005. Regression models for twin studies: a critical review. Int J Epidemiol 34(5):1089-1099, PMID: 16087687, https://doi.org/10.1093/ije/dyi153.

67. Patten KT, Valenzuela AE, Wallis C, Berg EL, Silverman JL, Bein KJ, et al. 2021. The effects of chronic exposure to ambient traffic-related air pollution on Alzheimer's disease phenotypes in wildtype and genetically predisposed male and female rats. Environ Health Perspect 129(5):057005, PMID: 33971107, https://doi.org/10.1289/EHP8905.

68. Kilian J, Kitazawa M. 2018. The emerging risk of exposure to air pollution on cognitive decline and Alzheimer's disease - evidence from epidemiological and animal studies. Biomed J 41(3):141-162, PMID: 30080655, https://doi.org/10. 1016/j.bj.2018.06.001.

69. Malik R, Georgakis MK, Neitzel J, Rannikmäe K, Ewers M, Seshadri S, et al. 2021. Midlife vascular risk factors and risk of incident dementia: longitudinal cohort and mendelian randomization analyses in the UK biobank. Alzheimers Dement 17(9):1422-1431, PMID: 33749976, https://doi.org/10.1002/alz.12320.

70. Kivipelto M, Ngandu T, Fratiglioni L, Viitanen M, Kåreholt I, Winblad B, et al. 2005. Obesity and vascular risk factors at midlife and the risk of dementia and Alzheimer disease. Arch Neurol 62(10):1556-1560, PMID: 16216938, https://doi.org/ 10.1001/archneur.62.10.1556.

71. Kivipelto M, Ngandu T, Laatikainen T, Winblad B, Soininen H, Tuomilehto J. 2006. Risk score for the prediction of dementia risk in 20 years among middle aged people: a longitudinal, population-based study. Lancet Neurol 5(9):735-741, PMID: 16914401, https://doi.org/10.1016/S1474-4422 (06)70537-3.

72. Braak H, Del Tredici K. 2015. The preclinical phase of the pathological process underlying sporadic Alzheimer's disease. Brain 138(10):2814-2833, PMID: 26283673, https://doi.org/10.1093/brain/awv236.

73. Casanova R, Wang X, Reyes J, Akita Y, Serre ML, Vizuete W, et al. 2016. A voxel-based morphometry study reveals local brain structural alterations associated with ambient fine particles in older women. Front Hum Neurosci 10:495, PMID: 27790103, https://doi.org/10.3389/fnhum.2016.00495.

74. Chen J-C, Wang X, Wellenius GA, Serre ML, Driscoll I, Casanova R, et al. 2015. Ambient air pollution and neurotoxicity on brain structure: evidence from wom-en's health initiative memory study. Ann Neurol 78(3):466-476, PMID: 26075655, https://doi.org/10.1002/ana.24460.

75. Power MC, Lamichhane AP, Liao D, Xu X, Jack CR, Gottesman RF, et al. 2018. The association of long-term exposure to particulate matter air pollution with

brain MRI findings: the ARIC study. Environ Health Perspect 126(2):027009, PMID: 29467108, https://doi.org/10.1289/EHP2152.

76. Wang X, Younan D, Millstein J, Petkus AJ, Garcia E, Beavers DP, et al. 2022. Association of improved air quality with lower dementia risk in older women. Proc Natl Acad Sci USA 119(2):e2107833119, PMID: 34983871, https://doi.org/10. 1073/pnas.2107833119.

77. Younan D, Wang X, Millstein J, Petkus AJ, Beavers DP, Espeland MA, et al. 2022. Air quality improvement and cognitive decline in community-dwelling older women in the United States: a longitudinal cohort study. PLoS Med 19(2): e1003893, PMID: 35113870, https://doi.org/10.1371/journal.pmed.1003893.

78. Heneka MT, Ramanathan M, Jacobs AH, Dumitrescu-Ozimek L, Bilkei-Gorzo A, Debeir T, et al. 2006. Locus ceruleus degeneration promotes Alzheimer pathoge-nesis in amyloid precursor protein 23 transgenic mice. J Neurosci 26(5):1343- 1354, PMID: 16452658, https://doi.org/10.1523/JNEUROSCI.4236-05.2006.

79. Kalinin S, Gavrilyuk V, Polak PE, Vasser R, Zhao J, Heneka MT, et al. 2007. Noradrenaline deficiency in brain increases beta-amyloid plaque burden in an animal model of Alzheimer's disease. Neurobiol Aging 28(8):1206-1214, PMID: 16837104, https://doi.org/10.1016/j.neurobiolaging.2006.06.003.