Skip to main content
NCBI home page
Springer logoLink to Springer
. 2026 Apr 7;44(2):19. doi: 10.1007/s12640-026-00786-2

Associations between air pollution and markers of neuroinflammation, synaptic dysfunction and core Alzheimer’s disease pathology vary by APOE genotype

Kana Kimura 1, Ira Driscoll 2,4,✉, Noah Cook 2,3, Sonia Shahzad 2, Tobey J Betthauser 2,4, Sterling C Johnson 2,4, Sanjay Asthana 2,4, Catherine L Gallagher 5,6, Bruce P Hermann 2,4,6, Mark A Sager 4, Kaj Blennow 7,8,9,10, Henrik Zetterberg 7,8,11,12,13,14,15, Cynthia M Carlsson 2,4,5, Gwendlyn Kollmorgen 16, Ozioma C Okonkwo 2,4,✉
PMCID: PMC13056749  PMID: 41944915

To determine whether long-term residential air pollution [AP; ozone (O₃) and fine particulate matter (PM₂.₅)] is associated with (1) incident mild cognitive impairment (MCI) or Alzheimer’s disease (AD), (2) biomarkers of core and AD-relevant pathology, and (3) whether these relationships are moderated by APOE4+/- (carrier/non-carrier of one or both ε4 alleles) status or mediated by neuroinflammation. Sample included 795 participants (Mage 68.7 ± 7.9; 68% female) from the Wisconsin Alzheimer’s Disease Research Center and Wisconsin Registry for Alzheimer’s Prevention parent studies, both enriched for AD risk at enrollment based on parental AD history. Residential zip code and 2009–2021 EPA-based annual AP reports were used to estimate individual exposure. Cox proportional hazards models assessed MCI/AD risk. Linear regressions examined the relationships between AP exposure and biomarkers of core and AD-relevant pathology, with and without APOE4 + stratification. Causal mediation analysis examined whether markers of inflammation mediated the AP-AD pathology relationships. Neither O₃ nor PM₂.₅ exposure predicted MCI/AD incidence nor core AD pathology (Ps > 0.05). Higher PM₂.₅ was associated with higher CSF GFAP levels (P = 0.003). APOE4 + with higher levels of PM₂.₅ exposure had higher CSF levels of tTau (P = 0.01), pTau₁₈₁ (P = 0.01) and neurogranin (P = 0.02). These relationships were not mediated by neuroinflammation (Ps > 0.05). In this AD-risk enriched cohort, AP was not associated with MCI/AD incidence. However, higher PM₂.₅ exposure was associated with astrocytic activation, and in APOE4+, AD pathology, neurodegeneration, and synaptic dysfunction. Our findings suggest AP as an environmental risk factor contributing to AD-relevant pathology, particularly among genetically at-risk individuals.

Keywords: Air pollution, PM₂.₅, APOE ε4 allele, Alzheimer’s disease, AD biomarkers, Neuroinflammation, Gene-environment interaction

Highlights

Long-term O₃ and PM₂.₅ exposures are not associated with MCI/AD risk in a cognitively unimpaired, middle-aged and older cohort enriched for AD risk based on parental history.

Higher long-term PM₂.₅ exposure is associated with higher levels of CSF GFAP, a marker of astrocyte reactivity.

In APOE ε4 allele carriers, higher long-term PM₂.₅ exposure is associated with higher levels of CSF pTau₁₈₁, tTau and neurogranin.

CSFbiomarkers of neuroinflammation did not mediate the relationships between PM₂.₅ exposure and AD-relevant pathology.

Results highlight gene–environment interplay in early AD pathophysiology.

Introduction

Alzheimer’s disease (AD) is a progressive neurodegenerative disease of complex etiology that likely involves multiple risk genes and environmental factors (Migliore & Coppede, 2022). To date, age is the strongest known risk factor for AD (Alzheimer 2024) while the ε4 allele of the apolipoprotein E(APOE) gene is the strongest known genetic risk factor for late-onset AD (Corder et al. 1993; Farrer et al. 1997; Saunders et al. 1993). Carrying one ε4 allele increases the risk by 20%, whereas carrying two ε4 alleles increases the risk by 90% (Corder et al. 1993). Given that AD is a multifactorial disease, efforts are focused on interactions between genetic and potentially modifiable risk factors in hopes of forestalling or mitigating disease incidence or progression, and societal burden associated with it.

Air pollution (AP), a potentially modifiable environmental risk factor, is linked to general health and 7 million (11.6%) premature deaths globally every year (World Health Organization, 2002). AP is also associated with mild cognitive impairment (MCI;Krishnamoorthy et al. 2018; Peters et al. 2019), which is considered a prodrome of AD (Petersen et al., 2009). Although the literature is largely nascent we do know that reductions in AP are associated with proportional reductions in AD and dementia incidence (Carey et al. 2018, Chen et al. 2017; Letellier et al., 2021, Peters et al. 2019, Wang et al. 2022).

Ozone (O3) is one of the common air pollutants that has been investigated in relation to AD. Literature suggests that repeated and long-term exposure to O3 leads to development of neuroinflammation, which is in turn associated with AD (Singh et al., 2023). Postmortem studies in people of all ages, including children, link severe O3exposure to neuroinflammation cerebral atrophy and tau accumulation all of which are events associated with AD (Croze & Luc., 2018).

Another air pollutant that has been investigated in relation to AD is particulate matter (PM). PM is emitted from various sources and can be categorized based on its aerodynamic diameter normally ranging between 2.5 and 10 micrometers (Tao et al., 2003; Zhang et al., 2015). Acute exposure to PM2.5is associated with inflammation (Arias-Pérez et al., 2020; Dubowsky et al., 2006) and synaptic dysfunction (Li et al., 2022). Long-term exposure to PM2.5is associated with neurodegenerative diseases (Shi et al., 2021) and increased mortality (Hao et al., 2023) in adults 65 years of age or older. Given the commonality of their presence in both urban and rural areas and their associations with neurodegenerative disease, both O3 and PM2.5 warrant investigation as possible risk factors for AD.

In the present study, we examine (1) the relationships between long-term residential AP (O3 and PM2.5) exposure and a combined risk for MCI or AD incidence, and (2) whether long-term AP exposure is associated with biomarkers of core AD neuropathology in cerebrospinal fluid (CSF) and via PET, and (3) CSF biomarkers of relevance to AD (neuroinflammation, synaptic dysfunction and neurodegeneration), and (4) whether the aforementioned relationships are moderated by APOE ε4 status or mediated by inflammation. Based on existent literature, we hypothesize that higher long-term AP exposure is associated with (1) increased risk for MCI or AD, and (2) the deleterious biomolecular changes of significance to AD reflected in CSF biomarkers or via positron emission tomography (PET), and that (3) the deleterious relationships between higher long-term AP exposure and PET or CSF biomarkers of relevance to AD are moderated by APOE ε4 status, while (4) the relationships between AP exposure and AD-relevant pathology are mediated by inflammation.

Methods

Participants

The sample consisted of cognitively unimpaired, late middle-aged and older adults (N = 795; MeanAGE(SD) = 68.7 ± 8; 68% female) from the Wisconsin Alzheimer’s Disease Research Center and the Wisconsin Registry for Alzheimer’s Prevention (WRAP; Johnsonet al., 2018) with available data of interest, with the exception of the survival analysis sample which in addition to cognitively unimpaired participants also included individuals with both MCI (N = 39) or dementia-AD (N= 11) consensus diagnoses. The enrollment criteria have been previously published in detail (Johnson et al., 2018). Briefly participants were 40–65 years of age at study entry fluent in English with adequate visual and auditory faculties for neuropsychological testing in general good health and cognitively unimpaired (Johnson et al., 2018). The sample was enriched at enrollment for AD risk based on parental history of AD (52%) and subsequently APOE ε4 allele carriage (APOE4+; 40%) compared to the general population, with the idea that AD pathogenesis, its predictors and the tempo of progression would be illuminated compared to their counterparts devoid of the same risks. Participants return for their second visit approximately 4 years after baseline and subsequent visits occur every 2 years thereafter. Medical history is collected by self-report.

Residential address was recorded for each participant starting with the baseline visit and only those who reported living at the same address for the duration of their participation were included in the current analyses. All procedures were approved by the University of Wisconsin School of Medicine and Public Health Institutional Review Board. Written informed consent was provided by all participants prior to study participation.

Consensus Diagnosis

Cognitive status of participants was assessed in accordance with the National Institute on Aging Alzheimer’s Association workgroup diagnostic criteria (Dubois et al., 2021). Participants were classified as “Cognitively Unimpaired” (CU), “MCI,” or “dementia-AD”. Majority of the participants (94%, N = 745) were CU; 5% of the participants had a diagnosis of MCI (N = 39) and 1% received dementia-AD (AD) diagnosis (N= 11). Those diagnosed as “Impaired-Other” were excluded from current analyses. Cognitive performance and functional status were evaluated at each visit; the procedures including the cognitive battery are published elsewhere (Johnson et al., 2018). Diagnosis at each visit was used for survival analyses, and only cognitively unimpaired participants were included in all other analyses.

Air Pollution

AP levels were determined using U.S. Environmental Protection Agency’s (EPA) annual pollution measurements based on participants’ zip codes, which should be identical between first and last visit for inclusion in this study. First and last year of participants’ recorded data for the current sample was 2009 and 2021 respectively. Thus, the AP exposure reflects the average pollutant values for each participant across follow-up (up to 13 years; Mean(SD) = 11.1(Balachandar et al. 2025) years). Figure1 depicts the AP level fluctuations per quarter year and of yearly average across the study follow-up along with the number of participants/observations for each period. The averages include all the measuring sites in Wisconsin, where 91% of our study participants reside. O3 is monitored hourly at each of the air quality measurement sites. From these hourly measurements, 17 different 8-hour rolling averages are calculated for each day (e.g., 7 AM-3 PM, 8 AM-4 PM,., 10 PM-6 AM; EPA, 2025). Then, the highest of these 17 averages becomes the “daily maximum 8-hour average O3” for that particular day. Ground-level O3is formed when nitrogen oxides and volatile organic compounds react in the presence of strong sunlight and high temperatures (Jiang et al., 2025). This means ozone concentrations are generally at their lowest in Q4 (fall/winter) and highest in Q2 (spring/summer). For PM2.5, averages are calculated over the entire day (24-hours; US. Environmental Protection Agency, 2025). PM2.5tends to be higher in the summer and winter months in many areas of the U.S. (Sun & Valachovic, 2025 ;Zhao et al., 2018). This is due to the use of fireplaces and increased energy consumption for heating in the winter (Q1/Q4) and also due to gases like sulfur dioxide and nitrogen oxide which undergo chemical reactions in the atmosphere to form fine particles. This process is accelerated during summertime (Q3) due to the increased sunlight and higher temperatures (Sun & Valachovic, 2025). Nonetheless, the annual averages are rather consistent across 13 years of follow-up (Fig. 1; dashed line).

Fig. 1.

Fig. 1

Open in a new tab

A O3 and (B) PM2.5 quarterly (solid line) and superimposed yearly (dashed line) average along with participant counts from 2009 to 2021. A 8-hour max daily average concentrations of the ground-level O3 (parts per million; ppm) per quarter (solid line) and per year (dashed line)

APOE Genotyping

DNA was isolated from whole-blood samples using the PUREGENE DNA Isolation Kit (Gentra Systems, Inc., Minneapolis, MN) and DNA concentrations were measured with Ultraviolet spectrophotometry (DU 530 Spectrophotometer, Beckman Coulter, Fullerton, CA). Genotyping for APOE (rs429358 and rs7412) was performed by LGC Genomics (Beverly, MA) via competitive allele-specific PCR-based KASP genotyping assays. For analyses purposes, participants were split into two groups based on genotype: carriers of at least one APOE ε4 allele carriers (APOE4+; ε2/ε4, ε3/ε4, ε4/ε4) and non-carriers (APOE4-; ε2/ε2, ε2/ε3, ε3/ε3).

CSF Collection and Assays

CSF samples were collected after a 12-hour fasting period at L3-4 or L4-5 using a drip method and/or gentle extraction technique into polypropylene syringes using a Sprotte 24- or 25-gauge spinal needle. A thin needle was used to inject 1% lidocaine as a local anesthetic prior to the insertion of the Sprotte spinal needle. Approximately 22 mL of CSF was pooled, gently mixed and centrifuged at 2,000 g for 10 min. Supernatants were frozen in 0.5 milliliters aliquots in polypropylene tubes, and kept at −80 °C.

CSF samples were analyzed for phosphorylated tau (pTau181), Aβ40, and Aβ42 to obtain biomarkers of core AD pathology (pTau181 and Aβ42/Aβ40), neurodegeneration [total tau (tTau)], synaptic dysfunction (neurogranin), and neuroinflammation[glial fibrillary acidic protein (GFAP) interleukin 6 (IL-6) S100 calcium-binding protein B (S100B) soluble triggering receptor expressed on myeloid cells 2 (sTREM2) and chitinase-3-like protein 1 (YKL-40)] using the NeuroToolKit (NTK; Roche Diagnostics International Ltd Rotkruez Switzerland). The NTK is a panel of exploratory prototype assays designed to robustly evaluate established AD biomarkers (Aβ and tau) as well as emerging markers. This toolkit enables a comprehensive characterization of AD pathology as well as a panel of synaptic axonal and glial biomarkers providing enhanced insights into the disease’s pathophysiological processes (Van Hulle et al., 2021).

Positron Emission Tomography (PET)

Details of acquisition protocols image reconstruction processing and quantification of PET images have been previously published (Betthauser et al., 2019). Briefly, participants underwent PET imaging on either a Siemens EXACT HR + or a Siemens Biograph Horizon scanner. Parametric PET images were co-registered to 3D T1-weighted MR images [inversion time (TI)/echo time (TE)/repetition time (TR) = 450ms/3.2ms/8.2ms, flip angle = 12°, slice thickness = 1 mm no gap, field of view (FOV) = 256, matrix size = 256 × 256; voxel resolution: 1 mm×1 mm×1 mm] obtained on a 3 T GE X750 Discovery scanner with an 8- or 32‐channel phased array head coil. Aβ burden was quantified from 11C-Pittsburg Compound B (PiB) 70-minute dynamic PET acquisition scans by generating a global cortical distribution volume ratio (DVR) using Logan graphical analysis with cerebellum grey matter as a reference (Johnson et al., 2014). To quantify tau, standard uptake value ratios (SUVRs) using the inferior cerebellar grey matter as reference were calculated from a 20-min dynamic 18F-MK6240 PET acquisition 70 min post bolus injection (Betthauser et al., 2019). MK6240 SUVR was the average of all regional SUVR values corresponding toBraak stages 1 –6 (Braak & Braak, 1991).

Statistical Analyses

Statistical analyses were performed in R Studio, version 4.2.3. Participant’s demographic characteristics were compared between three diagnostic groups (MCI, AD, and CU) via ANOVAs for parametric variables and Chi-square tests for categorical variables. We also calculated the means and standard deviations (M, SD) for each pollutant to assess average exposure levels to O₃ and PM₂.₅ across the entire sample and summarized the proportion of participants living in areas exceeding the corresponding National Ambient Air Quality Standards (NAAQS).

All models covaried for sex, age, education, race, APOE4 + status and parental history of AD, given their established relationships with AD risk (Alzheimer’s Association, 2024), and for years at a current residence. Age at the lumbar puncture visit served as a covariate in analyses with CSF biomarkers as outcomes and age at the PET visit date served as a covariate in analyses with PET-PiB and -MK6240 as outcomes. If multiple CSF and PET visits were recorded for a participant, we used the data for two observations that occurred closest in time. The age at the lumbar puncture visit differed from age at the PET visit for several participants; we controlled for this difference in age in our analyses.

Cox proportional hazard models examined the relationships between long-term AP exposure and combined MCI + AD risk. Regression models investigated the associations between biomarkers of core AD pathology (PET and CSF), and CSF biomarkers of neuroinflammation, neurodegeneration and synaptic dysfunction with long-term residential AP exposure. We also conducted sensitivity analyses with a sub-sample of CU participants (N = 48) matched to MCI + AD group based on age and sex, given that the CU group was substantially larger than the MCI + AD group. The regression models were repeated after stratifying by APOE4+/- status.

To test the hypothesis that inflammation mediates the relationship between AP and AD pathology, we conducted a parallel mediation analysis in which five inflammatory markers (GFAP, IL-6, S100B, sTREM2, and YKL-40) were modeled as simultaneous mediators with four core AD pathology markers (global PiB DVR, whole brain MK6240 SUVR, CSF pTau181, and CSF Aβ42/Aβ40ratio). This approach allowed us to examine whether each core AD biomarker uniquely contributes to the indirect effect of AP on AD pathology while controlling for potential confounders. Age sex parental history of AD and years of education were included in all models as covariates. Models were fitted using the ‘lavaan’ package (Rosseel, 2012). Maximum likelihood estimation with bias-corrected bootstrap standard errors (1,000 resamples) was applied to derive 95% confidence intervals and p‐values for all parameters. Given that multiple mediation pathways were being evaluated, p-values for the indirect and direct effects were corrected for multiple comparisons using the false discovery rate (FDR) procedure (Benjamini & Hochberg, 1995). Mediation analyses were repeated after stratifying by APOE4+/- status.

Results

Sample Characteristics

Characteristics of the entire sample and each diagnostic group separately are detailed in Table 1. The sample was on average 68.7 ± 8 years of age, predominately white (93%), female (68%) and college educated (Mean = 16.2 years). MCI + AD were significantly older than the CU (P < 0.001). There were no significant differences between the diagnostic groups with regard to sex, APOE4+ proportion, parental AD history, race, years of education, and years at residence (all Ps > 0.05).

Table 1.

Demographic characteristics of the entire sample and by diagnostic group

Baseline Characteristic Entire Sample
(N = 795)		 MCI
(N = 39)		 AD
(N = 11)		 CU
(N = 745)		 P
Sex N (%) 0.78
 Female 544 (68) 25 (64) 7 (64) 512 (69)
 Male 251 (32) 14 (36) 4 (36) 233 (31)
Age M (SD) 68.7 (8.0) 77.5 (7.6) 75.3 (10.2) 68.1 (8.0) < 0.001*
APOE N (%) 0.06
 APOE4+ 315 (40) 18 (46) 8 (73) 289 (39)
 APOE4- 478 (60) 21 (54) 3 (27) 454 (61)
 Unknown 2 (< 1) 0 (0) 0 (0) 2 (< 1)
Parental History of AD N (%) 0.70
 Yes 412 (52) 21 (54) 7 (64) 384 (52)
 No 203 (26) 10 (26) 1(9) 192 (26)
 Unknown 180 (23) 8 (21) 3 (27) 169 (23)
Race N (%) 0.47
 Caucasian/white 740 (93) 35 (90) 11 (100) 700 (93)
 African American/Black 42 (5) 4 (10) 0(0) 38 (5)
 Native American 7 (1) 0 (0) 0 (0) 7 (0.9)
 Asian 3 (0) 0 (0) 0 (0) 3 (0.4)
 Unknown 3 (0) 0 (0) 0 (0) 3 (0.4)
Education (years) M (SD) 16.2 (2) 15.7 (3) 15.9 (2) 16.1 (2) 0.60
Years at Residency M (SD) 11.1 (4) 10.9 (3) 10.7 (2) 11.2 (4) 0.84
Residency in Wisconsin N (%) 0.08
 Yes 725 (9) 39 (100) 11 (100) 675 (91)
 No 70 (10) 0 (0) 0 (0) 70 (9)
Open in a new tab

* - result significant at P < 0.001

Abbreviations: AD Alzheimer’s disease, APOE4 + careers of at least one ε4 allele, APOE4- non-careers of any ε4 allele, CU cognitively unimpaired, M mean, MCI mild cognitive impairment, SD standard deviation

Air Pollution (AP) Levels

Table 2 displays the mean concentration of each pollutant of interest across the entire sample along with the associated National Ambient Air Quality Standards (NAAQS) and the number/percentage of participants living in an area where pollutant concentrations exceed the NAASQ. Mean O3 concentration for the entire sample is 0.07 ppm (SD = 0.004) with 11% of the participants living in areas exceeding the NAAQS. The average PM2.5 concentration for the entire sample is 9.07 µg/m3 (SD = 0.06), with 58% of the participants living in areas with PM2.5 concentrations exceeding the NAAQS. We only included O3 and PM2.5 as our target pollutants given that less than 11% of the sample had available AP data for other pollutants [including carbon monoxide (CO), lead, nitrogen dioxide (NO2), sulfur dioxide (SO2), and particulate matter ≤ 10 μm (PM10)].

Table 2.

Average air pollution exposure and corresponding national ambient air quality standards for the entire sample

Pollutant Entire Sample
M (SD)		 Total # of participants with available data NAAQS Total # of participants above NAAQS
O3 (ppm) 0.07 (0.004) 749 (94%) 0.07 ppm 88 (12%)
PM2.5 (µg/m3) 9.07 (0.63) 494 (62%) 9 µg/m3 462 (94%)
Open in a new tab

Abbreviations: M mean, NAAQS National Ambient Air Quality Standards, O3 Ozone, PM2.5 particulate matter ≤ 2.5 μm, SD standard deviation

Associations between AP Exposure and Combined MCI + AD Risk

The survival curves for O3 and PM2.5 are presented in Fig. 2A. There were no significant associations between either long-term O3 or PM2.5 exposures and MCI + AD risk (all Ps > 0.05; Table 3). Figure 2B depicts the survival curves from sensitivity analyses, where we assessed the relationship between long-term AP exposure and MCI + AD risk in a sub-sample of CU that were matched to the MCI + AD group based on sex and age. The results remained unchanged (all Ps > 0.05).

Fig. 2.

Fig. 2

Open in a new tab

Probability of combined MCI or AD (MCI + AD) Incidence Over Time Based on AP Exposure. A Results for the Entire Sample (N = 795): no significant associations were observed between MCI + AD risk and long-term residential exposure to either O3 (red; P = 0.57) or PM2.5 (blue; P = 0.18). B Results from the “Matched Sample” sensitivity analyses of a subset of cognitively unimpaired participants matched to those with consensus MCI or AD diagnoses by age and sex (N = 96): no significant associations were observed between long-term residential exposure to either O3 (red; P = 0.91) or PM2.5 (blue; P = 0.94) and MCI + AD risk

Table 3.

Relative risk for MCI + AD based on AP exposure. (A) The results of survival analyses for the entire samples. Neither O3 nor PM2.5 were significantly associated with MCI + AD risks. (B) The results of survival analyses for the cognitively unimpaired sample (n = 48) matched to MCI + AD (n = 48) based on age and sex. Relationships between O3 or PM2.5 with MCI + AD risk remained non-significant

A. Entire Sample (n = 795)
Pollutant HR CI P-value
O3 (ppm) 3.95e-13 4.23e−56, 3.68e+30 0.57
PM2.5 (mg/m3) 0.71 0.43, 1.17 0.18
B. Matched Sample (n = 96)
Pollutant HR CI P-value
O3 (ppm) 4.26 1.25e−31, 3.99e+34 0.91
PM2.5 (mg/m3) 0.01 0.70, 1.48 0.94
Open in a new tab

All models covary for years at current address, APOE4 status, parental AD history, education, race, sex, and age

Abbreviations: CI confidence interval, HR hazard ratio, O3 ozone, PM2.5 particulate matter ≤ 2.5 μm

Associations between AP Exposure and AD-relevant Biomarkers

There were no significant associations between long-term AP exposure and core AD neuropathology [PET (Aβ or tau); CSF (Aβ42/Aβ 40, pTau181); all Ps > 0.05]. Higher CSF GFAP concentrations were significantly associated with higher levels of PM2.5 (P = 0.003), but not O3 (P > 0.05). No other significant associations were observed between AP exposure and any of the biomarkers investigated (all Ps > 0.05).

Associations between AP Exposure and AD-relevant Biomarkers after APOE4+/- Stratification

Results of APOE-stratified analyses for both PET and CSF biomarkers of significance to AD are presented in Figs. 3 and 4. The relationships with O3 exposure (Fig. 3) remain non-significant for all variables of interest after stratifying by APOE4 + status (all Ps > 0.05).

Fig. 3.

Fig. 3

Open in a new tab

Results of regression analyses assessing the relationships between AD-relevant PET and CSF biomarkers with O3, stratified by APOE4+/- status. The relationships between O3 exposure and all biomarkers of interest remained non-significant after stratifying by APOE4 + status (all Ps > 0.05). * - result significant at P < 0.05. Note. All models include age, sex, parental AD history APOE4 status, and years of education as covariates. Abbreviations. Aβ = beta amyloid; AD = Alzheimer’s disease; CI = confidence interval; CSF = cerebral spinal fluid; DVR = distribution volume ratio; GFAP = glial fibrillary acidic protein; IL-6 = interleukin-6; PET = positron emission tomography; PiB = Pittsburg compound B; S100B = S100 calcium-binding protein B; SE = standard error; sTREM2 = soluble triggering receptor expressed on myeloid cells 2; SUVR = standardized uptake value ratio; pTau181 = phosphorylated tau at threonine 181; tTau = total tau; YKL-40 = chitinase-3-like protein 1

Fig. 4.

Fig. 4

Open in a new tab

Results of regressions analyses assessing relationships between AD-relevant PET and CSF biomarkers with PM2.5, stratified by APOE4+/- status. APOE4 + compared to APOE4- with high long-term exposure to PM2.5 had significantly higher levels of pTau181 (P = 0.01), tTau (P = 0.01) and neurogranin (P = 0.02). There was a non-significant trend for the association between higher PM2.5 exposure and higher GFAP levels in APOE4+ (P = 0.07) and a significant association between higher PM2.5 exposure and GFAP levels in APOE4- (P = 0.04). * - result significant at P < 0.05. Note. All models include age, sex, parental AD history APOE4 status, and years of education as covariates. Abbreviations. Aβ = beta amyloid; AD = Alzheimer’s disease; CI = confidence interval; CSF = cerebral spinal fluid; DVR = distribution volume ratio; GFAP = glial fibrillary acidic protein; IL-6 = interleukin-6; PET = positron emission tomography; PiB = Pittsburg compound B; S100B = S100 calcium-binding protein B; SE = standard error; sTREM2 = soluble triggering receptor expressed on myeloid cells 2; SUVR = standardized uptake value ratio; pTau181 = phosphorylated tau at threonine 181; tTau = total tau; YKL-40 = chitinase-3-like protein 1

APOE4 + compared to APOE4- with high long-term exposure to PM2.5 (Fig. 4) had significantly higher levels of pTau181 (P = 0.01), tTau (P = 0.01) and neurogranin (P = 0.02). There was a non-significant trend for the association between higher PM2.5 exposure and higher GFAP levels in APOE4+ (P = 0.07) and a significant association between higher PM2.5 exposure and GFAP levels in APOE4− (P = 0.04).

Inflammation Does Not Mediate the Relationship between PM2.5 Exposure and Core AD Pathology

The mediation analyses assessed whether the association between long-term PM2.5 exposure and core AD pathology is mediated by inflammation (Table 4). The direct pathways quantify the association between PM2.5 and each AD pathology marker separately, holding inflammatory mediators constant. The direct effects of PM2.5 on each inflammatory marker were also tested separately and non were significant (all Ps > 0.05). Indirect effects represent the product of (a) the effect of PM2.5 on each inflammatory marker, and (b) the effect of inflammation on AD pathology. None of the pathways significantly mediated PM2.5’s effect on core AD pathology (all Ps > 0.05). Total effects are the combination of direct and indirect pathways, reflecting the overall association between PM2.5 and each core AD pathology with the mediators and none were significant (all Ps > 0.05). Table 5 shows the results of mediation models stratified by APOE4+/- status. The direct, indirect, and total effects remained non-significant (all Ps > 0.05).

Table 4.

Results of mediation analyses assessing if the relationship between PM2.5 and core AD pathology is mediated by inflammation

Pathway b (SE) P
Direct Effects PM2.5 → PiB global DVR 0.004 (0.03) 0.99
PM2.5 → MK6240 SUVR 0.04 (0.04) 0.66
PM2.5 → CSF pTau181 0.10 (1.43) 0.99
PM2.5 → CSF Ab42/Ab40 −0.004 (0.004) 0.63
PM2.5 → GFAP 0.7 (0.5) 0.60
PM2.5 → IL-6 0.11 (0.47) 0.98
PM2.5 → S100B 0.003 (0.07) 0.99
PM2.5 → sTREM2 0.01 (0.50) 0.99
PM2.5 → YKL-40 −3.30 (11.75) 0.95
Indirect Effects Total indirect effects < 0.001 (0.001) 0.38
PM2.5 → GFAP → core AD pathology 0.001 (0.001) 0.38
PM2.5 → IL-6 → core AD pathology < 0.001 (< 0.001) 0.93
PM2.5 → S100B → core AD pathology <−0.001 (0.001) 0.97
PM2.5 → sTREM2 → core AD pathology <−0.001 (0.001) 0.99
PM2.5 → YKL-40 → core AD pathology <−0.001 (< 0.001) 0.88
Total Effects PiB global DVR (Direct + Indirect Effects) 0.01 (0.03) 0.89
MK6240 SUVR (Direct + Indirect Effects) 0.04 (0.04) 0.37
CSF pTau181(Direct + Indirect Effects) 0.10 (1.44) 0.95
CSF Ab42/Ab40(Direct + Indirect Effects) < 0.001 (< 0.001) 0.40
Open in a new tab

P-values are false discovery rate (FDR) adjusted

Abbreviations: Aβ beta amyloid, AD Alzheimer’s disease, CI confidence interval, CSF cerebral spinal fluid, DVR distribution volume ratio, GFAP glial fibrillary acidic protein, IL-6 interleukin-6, PET positron emission tomography, PiB Pittsburg compound B, S100B S100 calcium-binding protein B, SE standard error, sTREM2 soluble triggering receptor expressed on myeloid cells 2, SUVR standardized uptake value ratio, pTau181 phosphorylated tau at threonine 181, tTau total tau, YKL-40 chitinase-3-like protein 1

Table 5.

Results of mediation analyses assessing whether the relationship between PM2.5 and core AD pathology is mediated by inflammation; stratified by APOE4+/- status

APOE4+ APOE4-
Paths β (SE) P β (SE) P
Direct Effects
PM2.5 → PiB global DVR −0.01 (0.09) 0.97 0.02 (0.15) 0.99
PM2.5 → MK6240 SUVR 0.02 (0.14) 0.99 0.03 (0.07) 0.99
PM2.5 → CSF pTau181 1.83 (2.74) 0.96 −1.21 (2.56) 0.99
PM2.5 → CSF Aβ42/Aβ40 −0.01 (0.01) 0.96 −0.003 (0.01) 0.99
PM2.5 → GFAP −0.04 (1.97) 0.99 1.12 (1.00) 0.87
PM2.5 → IL-6 0.33 (0.60) 0.96 −0.28 (0.86) 0.99
PM2.5 → S100B −0.09 (0.11) 0.96 0.08 (0.07) 0.85
PM2.5 → sTREM2 −0.29 (0.79) 0.96 0.47 (0.94) 0.99
PM2.5 → YKL-40 −4.18 (14.74) 0.96 −2.91 (22.91) 0.99
Indirect Effects
Total indirect effects 0.003 (0.01) 0.70 −0.001 (0.003) 0.83
PM2.5 → GFAP → core AD pathology <−0.0001 (0.01) 0.99 < 0.0001 (0.001) 0.71
PM2.5 → IL-6 → core AD pathology 0.001 (0.003) 0.65 < 0.0001 (0.001) 0.98
PM2.5 → S100B → core AD pathology 0.002 (0.003) 0.57 <−0.0001 (0.001) 0.71
PM2.5 → sTREM2 → core AD pathology < 0.0001 (0.003) 0.90 <−0.0001 (0.001) 0.92
PM2.5 → YKL-40 → core AD pathology < 0.0001 (0.003) 0.96 <−0.0001 (0.002) 0.95
Total Effects
PiB global DVR (Direct + Indirect Effects) −0.01 (0.09) 0.94 0.02 (0.08) 0.81
MK6240 SUVR (Direct + Indirect Effects) 0.03 (0.14) 0.84 0.03 (0.07) 0.61
CSF pTau181 (Direct + Indirect Effects) 1.84 (2.74) 0.50 −1.21 (2.57) 0.64
CSF Aβ42/Aβ40 (Direct + Indirect Effects) −0.01 (0.01) 0.54 −0.004 (0.01) 0.66
Open in a new tab

P-values are false discovery rate (FDR) adjusted

Abbreviations: Aβ beta amyloid, AD Alzheimer’s disease, CI confidence interval, CSF cerebral spinal fluid, DVR distribution volume ratio, GFAP glial fibrillary acidic protein, IL-6 interleukin-6, PET positron emission tomography, PiB Pittsburg compound B, S100B S100 calcium-binding protein B, SE standard error, sTREM2 soluble triggering receptor expressed on myeloid cells 2, SUVR standardized uptake value ratio, pTau181 phosphorylated tau at threonine 181, tTau total tau, YKL-40 chitinase-3-like protein 1

Discussion

We report no significant associations between long-term AP exposure and MCI + AD risk in our late middle-aged and older sample enriched for AD risk at enrollment. Specifically, neither O3 nor PM2.5were associated with risk for combined MCI and AD incidence and the relationships remained non-significant when the sensitivity analyses were performed with an age- and sex-matched sample of CU participants. Our findings are in contrast with both our hypothesis and the existing literature which generally reports a relationship between long-term AP exposure and AD risk (Chen et al., 2017; He et al., 2022; Shi et al., 2021; Sullivan et al., 2021).

The absence of significant findings in the present study may be due to a relatively low number of participants with MCI and AD and subsequently low statistical power to detect associations between disease incidence or AD-relevant biomarkers and AP. Sensitivity analyses which restricted the much larger CU sample to match those with MCI or AD diagnosis in age and sex confirmed the lack of significant findings. Still it is of note that the existing studies included 1000 + participants with AD diagnosis (Chen et al., Shi et al., 2021; Sullivan et al., 2021) compared to a total of 48 participants with a combined MCI and AD diagnoses in our current sample. Moreover, participants in the current study had similar levels of AP exposure - most reside in the state of Wisconsin where AP appears rather uniform across the state. This lack of variability in long-term residential AP exposure across our sample is likely to have contributed to the null findings. Limitations notwithstanding, WRAP remains a valuable longitudinal dataset for investigating AD risk, as it recruits individuals with parental history of AD and higher genetic risk of developing AD than the general population. Future studies should also consider integrating higher resolution of AP data for close areas in Wisconsin. Integrating more fine-grained analysis of AP with WRAP’s rich prospective biomarker and cognitive data would allow for examination of the role of the duration and intensity of AP exposure on AD trajectories, particularly among at-risk individuals.

We also report no significant associations between long-term AP exposure and core AD biomarkers (PET or CSF) which again is in opposition with the existing literature reporting an association between long-term AP exposure and higher levels of Aβ deposition in ALFA+ cohort of cognitively unimpaired adults at risk for AD (Alemany et al., 2021). There are several notable differences between this and our current study that may have contributed to disparate results. We used participant’s current residential zip code and 2009–2021 EPA-based annual air pollution while Alemany and colleagues (2021) applied Land Use Regression model from 2009 to estimate residential exposure to air pollutants. Moreover, some of the significant findings in their study were in relation to nitrogen dioxide (NO2) and particulate matter PM10, which we did not investigate. Furthermore, Alemany and colleagues (2021) required participants to have resided in the same residence for at least previous 3 years and all resided in the city of Barcelona, while we required that the participants have lived in the same residence over the period of up to 13 years since their enrollment in WRAP. It is also of note that Europe has pollutant profiles defined differently from the United States, that is a problem with using a proxy that does not represent accurately exposures to the smallest PM: UFPM and nanoparticles (Europeans are now doing such measurements).

When we extend the analyses to biomarkers of neuroinflammation, neurodegeneration, and synaptic dysfunction, however, we find that higher CSF glial fibrillary acidic protein (GFAP) levels are significantly associated with higher PM2.5exposure which is in line with the existing literature (Balachandar et al., 2025 ;Song et al., 2022). GFAP is predominantly expressed in astrocytes (Bongcam-Rudloff et al., 1991 ) and thus elevated GFAP expression in CSF is widely used as a biomarker of astrocyte reactivity and glial injury (Nielsen et al., 2002 ). It is postulated that PM2.5would first induce inflammation which in turn would activate astrocytes (Balachandar et al., 2025). Several mechanisms underlying glial activation in response to PM2.5 exposure have been identified; they primarily involve oxidative stress direct particle contact systemic inflammation and the release of inflammatory mediators (Kang et al., 2021; Li et al., 2022). These pathways lead to the activation of microglia and astrocytes causing neuroinflammation and neurotoxicity that can contribute to neurodegenerative disorders (Nunez et al., 2022; Thiankhaw et al., 2022).

After stratifying the analyses based on APOE4+/- status, we did observe the hypothesized relationships between CSF biomarkers of core AD neuropathology (pTau181), neurodegeneration (tTau) and synaptic dysfunction (neurogranin) in relation to PM2.5 exposure, with higher CSF concentrations present in those at genetic risk for AD (APOE4+) who are also experiencing higher levels of AP exposure. CSF GFAP levels were positively correlated with PM2.5, but only in APOE-. APOE4 + individuals often have higher baseline levels of CSF GFAP (Stocker et al., 2025 ), which can show up as a more robust relationships with AP in APOE- individuals. Indeed, we observed numerically higher CSF GFAP levels in APOE4 + with higher PM2.5 exposure (p = 0.07). Thus, although statistically non-significant, APOE4 + individuals seem to have a positive relationship between PM2.5 and CSF GFAP, despite its possible ceiling effect.

In animal models, E4FAD transgenic mice exposed to nanoscale urban particulate matter show 2.8-fold greater increase in cortical amyloid plaque load compared to E3FAD controls despite identical exposure, suggesting that being an ε4 carrier may drastically lower the threshold for PM 2.5-induced neurodegeneration (Cacciottolo et al., 2017). Moreover, APOE4 knock-in mice treated with repeated low-dose Lipopolysaccharide injections which model environmental inflammation had significant dendritic spine loss even in the absence of significant changes in astrocyte or microglia numbers indicating that ε4 allele may heighten synaptic vulnerability (Ganesan et al., 2024). Human epidemiology studies also suggest that APOE4 + individuals exposed to urban PM and ultrafine particles exhibit early AD pathology and neuroinflammatory responses (Calderón‑Garcidueñas et al. 2023) worse cognition (Hsiao et al., 2023 ) and lower hippocampal volumes (Popov et al., 2024). We report a similar pattern of results with PM2.5 classified based on U.S. standards, while all the aforementioned studies were conducted outside of the United States and have different standards for classifying air pollutants. Overall, our findings suggest that being a carrier of at least one ε4 allele of the APOE gene is associate with greater vulnerability to not only genetic but also environmental risk factors known to exacerbate core AD neuropathology, neurodegeneration and neuroinflammation.

No prior study to date has investigated whether the association between AP exposure and AD pathology is mediated by neuroinflammatory mechanisms. The results of our mediation analyses suggest that there is a direct relationship between inflammatory markers and core AD pathology but no significant mediation effects of neuroinflammation on the relationship between AP exposure and core AD pathology. Still the direct relationship between neuroinflammation and core AD pathology is worthy of further investigation given that the literature generally supports a significant relationship between inflammation and AP (Block & Calderón-Garcidueñas, 2009 ; Jayaraj et al., 2017 ) and the limitations of the current study, namely the small number of participants with MCI or AD diagnoses and the lack variability in AP exposure.

Also, AP values were based on participant’s home address and data related to workplace or proximity to highways were not available. Nonetheless, the AP exposure in our study spanned up to 13 years, from enrollment to the most recent available WRAP visit and study records were carefully examined for a potential address change. Moreover, though the current study did not limit where in the United States participants lived, 91% of the participants resided in Wisconsin where O3 exposure levels were relatively low across the state. The O3 and PM2.5 were monitored and collected according to the U.S. EPA standards, potentially affecting the generalizability of our findings to those of studies conducted outside of the U.S., yet most report a similar pattern of results Third, the current study has a small sample size of participants with MCI and AD diagnosis. Although the sensitivity analyses showed similar results, careful interpretation of our results within the context of its limitations is required.

Overall, our results suggest that being a carrier of one or both ε4 alleles of the APOE gene may uniquely increase sensitivity to environmental factors, such as long-term AP exposure, which is known to exacerbate core AD pathology, neurodegeneration, and synaptic dysfunction. Moreover, studies with larger samples are needed to better understand the risk associated not with carrying one vs. both risk (ε4) but also the protective (ε2) alleles. Nonetheless, our findings highlight the need for consideration of air quality when assessing AD risk, especially for those with genetic predisposition for the disease. Further, our findings suggest that reducing long-term AP exposure could be a potentially viable practical intervention aimed at slowing the disease onset or progression in APOE4+. As urbanization intensifies and percentage of older population increases globally, management of AP levels may become a crucial step in both risk assessment and as preventive strategy for AD.

Acknowledgements

We would like to acknowledge and thank the staff and study participants of the Wisconsin Registry for Alzheimer’s Prevention and the Wisconsin Alzheimer’s Disease Research Center and the laboratory technicians at the Clinical Neurochemistry Laboratory at the Mölndal campus, University of Gothenburg, Sweden, without whom this work would not be possible. This work was supported by the National Institute on Aging grants R01AG062167 and R01AG085592 (O.C.O.), R01AG077507 (I.D.), R01AG027161 (S.C.J), R01AG021155 (S.C.J.), R01AG054059 (C.E.G.) and P30AG062715 (S.A.), and a Clinical and Translational Science Award (UL1RR025011) to the University of Wisconsin, Madison. Portions of this research were supported by the Veterans Administration, including facilities and resources at the Geriatric Research Education and Clinical Center of the William S. Middleton Memorial Veterans Hospital, Madison, WI, a core grant to the Waisman Center from the National Institute of Child Health and Human Development (P50 HD105353), NIH High-End Instrumentation grant (S10 OD030415), European Research Council (#101053962), the Swedish Research Council (#2023-00356, #2022-01018 and #2019-02397), the Swedish Brain Foundation (#FO2017-0243), the Swedish Alzheimer Foundation (#AF-742881); the Swedish state under the agreement between the Swedish government and the County Councils, the ALF-agreement (#ALFGBG-715986; #ALFGBG-71320), and the Knut and Alice Wallenberg Foundation. H.Z. is a Wallenberg Scholar and a Distinguished Professor at the Swedish Research Council.

Author Contributions

K.K. was the project manager, visualized data, and did all the analyses.K.K., I.D., and S.S wrote the main manuscript.I.D., S.C.J., S.A., and O.C.O. provided funding.N.C. performed data curation.T.J.B., S.C.J., S.A., C.L.G., B.P.H., M.A.S., K.B., H.Z., C.M.C., G.K provided resources.I.D., and O.C.O. were K.K.‘s supervisors on this project.All authors reviewed the manuscript.

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Ira Driscoll, Email: idriscoll@medicine.wisc.edu.

Ozioma C. Okonkwo, Email: ozioma@medicine.wisc.edu

References

  1. Alzheimer’s disease facts and figures (2024) Alzheimer’s & dementia: the journal of the Alzheimer’s Association, 20(5), 3708–3821. 10.1002/alz.13809 [DOI] [PMC free article] [PubMed]
  2. Alemany S, Crous-Bou M, Vilor-Tejedor N, Milà-Alomà M, Suárez-Calvet M, Salvadó G, Cirach M, Arenaza-Urquijo EM, Sanchez-Benavides G, Grau-Rivera O, Minguillon C, Fauria K, Kollmorgen G, Domingo Gispert J, Gascón M, Nieuwenhuijsen M, Zetterberg H, Blennow K, Sunyer J, Molinuevo L, J (2021) Associations between air pollution and biomarkers of Alzheimer’s disease in cognitively unimpaired individuals. Environ Int 157:106864. 10.1016/j.envint.2021.106864 [DOI] [PubMed] [Google Scholar]
  3. Arias-Pérez RD, Taborda NA, Gómez DM, Narvaez JF, Porras J, Hernandez JC (2020) Inflammatory effects of particulate matter air pollution. Environ Sci Pollut Res 27(34):42390–42404. 10.1007/s11356-020-10574-w [DOI] [PubMed] [Google Scholar]
  4. Balachandar R, Viramgami A, Singh DP, Kulkarni N, Chudasama B, Sivaperumal P, Upadhyay K (2025) Association between chronic PM2.5 exposure and neurodegenerative biomarkers in adults from critically polluted area. BMC Public Health 25:1413. 10.1186/s12889-025-22641-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Benjamini Y, Hochberg Y (1995) Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J Royal Stat Soc Ser B (Methodological) 57(1):289–300. http://www.jstor.org/stable/2346101 [Google Scholar]
  6. Betthauser TJ, Cody KA, Zammit MD, Murali D, Converse AK, Barnhart TE, Stone CK, Rowley HA, Johnson SC, Christian BT (2019) In Vivo Characterization and Quantification of Neurofibrillary Tau PET Radioligand18 F-MK-6240 in Humans from Alzheimer Disease Dementia to Young Controls. J Nucl Med 60(1):93–99. 10.2967/jnumed.118.209650 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Block ML, Calderón-Garcidueñas L (2009) Air pollution: mechanisms of neuroinflammation and CNS disease. Trends Neurosci 32(9):506–516. 10.1016/j.tins.2009.05.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bongcam-Rudloff E, Nistér M, Betsholtz C, Wang JL, Stenman G, Huebner K, Croce CM, Westermark B (1991) Human glial fibrillary acidic protein: complementary DNA cloning, chromosome localization, and messenger RNA expression in human glioma cell lines of various phenotypes. Cancer Res 51(5):1553–1560 [PubMed] [Google Scholar]
  9. Braak H, Braak E (1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol 82(4):239–259. 10.1007/BF00308809 [DOI] [PubMed] [Google Scholar]
  10. Cacciottolo M, Wang X, Driscoll I, Woodward N, Saffari A, Reyes J, Serre ML, Vizuete W, Sioutas C, Morgan TE, Gatz M, Chui HC, Shumaker SA, Resnick SM, Espeland MA, Finch CE, Chen JC (2017) Particulate air pollutants, APOE alleles and their contributions to cognitive impairment in older women and to amyloidogenesis in experimental models. Transl Psychiatry 7(1):e1022–e1022. 10.1038/tp.2016.280 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Calderón-Garcidueñas L, Hernández-Luna J, Aiello-Mora M, Brito-Aguilar R, Evelson PA, Villarreal-Ríos R, Torres-Jardón R, Ayala A, Mukherjee PS (2023) APOE peripheral and brain impact: APOE4 carriers accelerate their Alzheimer continuum and have a high risk of suicide in PM2.5 polluted cities. Biomolecules 13(6):927–927. 10.3390/biom13060927 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Carey IM, Anderson HR, Atkinson RW, Beevers SD, Cook DG, Strachan DP, Dajnak D, Gulliver J, Kelly FJ (2018) Are noise and air pollution related to the incidence of dementia? A cohort study in London, England. BMJ Open 8(9):e022404. 10.1136/bmjopen-2018-022404 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Chen H, Kwong JC, Copes R, Hystad P, Van Donkelaar A, Tu K, Brook JR, Goldberg MS, Martin RV, Murray BJ, Wilton AS, Kopp A, Burnett RT (2017) Exposure to ambient air pollution and the incidence of dementia: a population-based cohort study. Environ Int 108:271–277. 10.1016/j.envint.2017.08.020 [DOI] [PubMed] [Google Scholar]
  14. Chen W, Ge P, Lu Z, Liu X, Cao M, Yan Z, Chen M (2024) Acute exposure to seasonal PM2.5 induces toxicological responses in A549 cells cultured at the air-liquid interface mediated by oxidative stress and endoplasmic reticulum stress. Environ Res 248:118283. 10.1016/j.envres.2024.118283 [DOI] [PubMed] [Google Scholar]
  15. Corder EH, Saunders AM, Strittmatter WJ, Schmechel DE, Gaskell PC, Small GW, Roses AD, Haines JL, Pericak-Vance MA (1993) Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families. Sci (New York N Y) 261(5123):921–923. 10.1126/science.8346443 [DOI] [PubMed] [Google Scholar]
  16. Croze ML, Zimmer L (2018) Ozone atmospheric pollution and Alzheimer’s disease: from epidemiological facts to molecular mechanisms. J Alzheimer’s disease: JAD 62(2):503–522. 10.3233/JAD-170857 [DOI] [PubMed] [Google Scholar]
  17. Dubois B, Villain N, Frisoni GB, Rabinovici GD, Sabbagh M, Cappa S, Bejanin A, Bombois S, Epelbaum S, Teichmann M, Habert M-O, Nordberg A, Blennow K, Galasko D, Stern Y, Rowe CC, Salloway S, Schneider LS, Cummings JL, Feldman HH (2021) Clinical diagnosis of Alzheimer’s disease: recommendations of the International Working Group. Lancet Neurol 20(6):484–496. 10.1016/S14744422(21)00066-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Dubowsky SD, Suh H, Schwartz J, Coull BA, Gold DR (2006) Diabetes, obesity, and hypertension may enhance associations between airpollution and markers of systemic inflammation. Environ Health Perspect 114(7):992–998. 10.1289/ehp.8469 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Farrer LA, Cupples LA, Haines JL, Hyman B, Kukull WA, Mayeux R, Van Duijn CM (1997) Effects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer disease: a meta-analysis. JAMA 278(16):1349–1356. 10.1001/jama.278.16.1349 [PubMed] [Google Scholar]
  20. Ganesan K, Rentsch P, Langdon A, Milham LT, Vissel B (2024) Modeling sporadic Alzheimer’s disease in mice by combining Apolipoprotein E4 risk gene with environmental risk factors. Front Aging Neurosci 16:1357405. 10.3389/fnagi.2024.1357405 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hao H, Wang Y, Zhu Q, Zhang H, Rosenberg A, Schwartz J, Amini H, Van Donkelaar A, Martin R, Liu P, Weber R, Russel A, Yitshak-sade M, Chang H, Shi L (2023) National cohort study of long-term exposure to PM2.5 components and mortality in Medicare American older adults. Environ Sci Technol 57(17):6835–6843. 10.1021/acs.est.2c07064 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. He F, Tang J, Zhang T, Lin J, Li F, Gu X, Chen R (2022) Impact of air pollution exposure on the risk of Alzheimer’s disease in China: a community-based cohort study. Environ Res 205:112318. 10.1016/j.envres.2021.112318 [DOI] [PubMed] [Google Scholar]
  23. Hsiao P-J, Wu C-D, Wong P-Y, Chung M-C, Yang Y-W, Wu L-Y, Hsiao K-Y, Chung C-J (2023) APOE-ε4 alleles modify the decline of MMSE scores associated with time-dependent PM2.5 exposure: findings from a community-based longitudinal cohort study. Am J Geriatr Psychiatry 32(9):1080–1092. 10.1016/j.jagp.2023.12.021 [DOI] [PubMed] [Google Scholar]
  24. Jayaraj RL, Rodriguez EA, Wang Y, Block ML (2017) Outdoor ambient air pollution and neurodegenerative diseases: the neuroinflammation hypothesis. Curr Environ Health Rep 4(2):166–179. 10.1007/s40572-017-0142-3 [DOI] [PubMed] [Google Scholar]
  25. Jiang L, Peng H, Zhou Y et al (2025) Current progress on tropospheric ozone sources, biological effects and trends. Int J Biometeorol 69:2915–2940. 10.1007/s00484-025-03010-6 [DOI] [PubMed] [Google Scholar]
  26. Johnson SC, Christian BT, Okonkwo OC, Oh JM, Harding S, Xu G, Hillmer AT, Wooten DW, Murali D, Barnhart TE, Hall LT, Racine AM, Klunk WE, Mathis CA, Bendlin BB, Gallagher CL, Carlsson CM, Rowley HA, Hermann BP, Sager MA (2014) Amyloid burden and neural function in people at risk for Alzheimer’s Disease. Neurobiol Aging 35(3):576–584. 10.1016/j.neurobiolaging.2013.09.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Johnson SC, Koscik RL, Jonaitis EM, Clark LR, Mueller KD, Berman SE, Bendlin BB, Engelman CD, Okonkwo OC, Hogan KJ, Asthana S, Carlsson CM, Hermann BP, Sager MA (2018) The Wisconsin Registry for Alzheimer’s Prevention: a review of findings and current directions. Alzheimer’s Dementia: Diagnosis Assess Disease Monit 10(1):130–142. 10.1016/j.dadm.2017.11.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kang YJ, Tan HY, Lee CY, Cho H (2021) An air particulate pollutant induces neuroinflammation and neurodegeneration in human brain models. Advanced science (Weinheim, Baden-Wurttemberg, Germany) 8(21):e2101251. 10.1002/advs.202101251 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Krishnamoorthy Y, Sarveswaran G, Sivaranjini K, Sakthivel M, Majella MG, Kumar SG (2018) Association between indoor air pollution and cognitive impairment among adults in rural Puducherry, South India. J Neurosci Rural Pract 09(04):529–534. 10.4103/jnrp.jnrp_123_18 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Li W, Lin G, Xiao Z, Zhang Y, Li B, Zhou Y, Ma Y, Chai E (2022) A review of respirable fine particulate matter (PM2.5)-induced brain damage. Front Mol Neurosci 15:967174. 10.3389/fnmol.2022.967174 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Migliore L, Coppedè F (2022) Gene–environment interactions in Alzheimer disease: the emerging role of epigenetics. Nat Reviews Neurol 18(11):643–660. 10.1038/s41582-022-00714-w [DOI] [PubMed] [Google Scholar]
  32. Nielsen AL, Holm IE, Johansen M, Bonven B, Jørgensen P, Jørgensen AL (2002) A new splice variant of glial fibrillary acidic protein, GFAP epsilon, interacts with the presenilin proteins. J Biol Chem 277(33):29983–29991. 10.1074/jbc.M112121200 [DOI] [PubMed] [Google Scholar]
  33. Nunez Y, Boehme AK, Goldsmith J, Li M, van Donkelaar A, Weisskopf MG, Re DB, Martin RV, Kioumourtzoglou MA (2022) PM2.5 composition and disease aggravation in amyotrophic lateral sclerosis: An analysis of long-term exposure to components of fine particulate matter in New York State. Environ Epidemiol 6(2):e204. 10.1097/EE9.0000000000000204 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Peters R, Ee N, Peters J, Booth A, Mudway I, Anstey KJ (2019) Air Pollution and Dementia: A Systematic Review. J Alzheimer’s Disease 70(s1):S145–S163. 10.3233/JAD-180631 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Petersen RC, Roberts RO, Knopman DS, Boeve BF, Geda YE, Ivnik RJ, Smith GE, Jack CR Jr (2009) Mild cognitive impairment: ten years later. Arch Neurol 66(12):1447–1455. 10.1001/archneurol.2009.266 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Popov VA, Ukraintseva SV, Duan H, Yashin AI, Arbeev KG (2024) Traffic-related air pollution and APOE4 can synergistically affect hippocampal volume in older women: new findings from UK Biobank. Frontiers in Dementia. 10.3389/frdem.2024.1402091 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Rosseel Y (2012) Lavaan: an R package for structural equation modeling. J Stat Softw 48(2):1–36. 10.18637/jss.v048.i02 [Google Scholar]
  38. Saunders AM, Strittmatter WJ, Schmechel D, St. George-Hyslop PH, Pericak-Vance MA, Joo SH, Rosi BL, Gusella JF, Crapper-MacLachlan DR, Alberts MJ, Hulette C, Crain B, Goldgaber D, Roses AD (1993) Association of apolipoprotein E allele Ε4 with late-onset familial and sporadic alzheimer’s disease. Neurology 43(8):1467–1467. 10.1212/wnl.43.8.1467 [DOI] [PubMed] [Google Scholar]
  39. Shi L, Steenland K, Li H, Liu P, Zhang Y, Lyles RH, Requia WJ, Ilango SD, Chang HH, Wingo T, Weber RJ, Schwartz J (2021) A national cohort study (2000–2018) of long-term air pollution exposure and incident dementia in older adults in the United States. Nat Commun 12(1):6754. 10.1038/s41467-021-27049-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Shin J, Han S-H, Choi J (2019) Exposure to ambient air pollution and cognitive impairment in community-dwelling older adults: the Korean frailty and aging cohort study. Int J Environ Res Public Health 16(19):3767. 10.3390/ijerph16193767 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Singh S. A, Suresh S, Vellapandian C (2023) Ozone-induced neurotoxicity: in vitro and in vivo evidence. Ageing Res Rev 91:102045. 10.1016/j.arr.2023.102045 [DOI] [PubMed] [Google Scholar]
  42. Song J, Qu R, Sun B, Chen R, Kan H, An Z, Jiang J, Li J, Zhang Y, Wu W (2022) Associations of short-term exposure to fine particulate matter with neural damage biomarkers: a panel study of healthy retired adults. Environ Sci Technol 56(11):7203–7213. 10.1021/acs.est.1c03754 [DOI] [PubMed] [Google Scholar]
  43. Stocker H, Beyer L, Trares K, Stevenson-Hoare J, Rujescu D, Holleczek B, Beyreuther K, Schoettker B, Gerwert K, Brenner H (2025) Association of nonmodifiable risk factors with Alzheimer disease blood biomarkers in community-dwelling adults in the ESTHER study. Neurology 104(9):e213500. 10.1212/WNL.0000000000213500 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Sullivan, K. J., Ran, X., Wu, F., Chang, C. C. H., Sharma, R., Jacobsen, E., … Ganguli,M. (2021). Ambient fine particulate matter exposure and incident mild cognitive impairment and dementia. Journal of the American Geriatrics Society, 69(8), 2185–2194. 10.1111/jgs.17188 [DOI] [PMC free article] [PubMed]
  45. Sun Y, Valachovic EL (2025) Seasonal and periodic patterns of PM2.5 in Manhattan using the variable bandpass periodic block bootstrap. PLoS One 20(6):e0326767. 10.1371/journal.pone.0326767 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Thiankhaw K, Chattipakorn N, Chattipakorn SC (2022) PM2.5 exposure in association with AD-related neuropathology and cognitive outcomes. Environmental pollution (Barking, Essex : 1987) 292(Pt A):118320. 10.1016/j.envpol.2021.118320 [DOI] [PubMed] [Google Scholar]
  47. Tao F, Gonzalez-Flecha B, Kobzik L (2003) Reactive oxygen species in pulmonary inflammation by ambient particulates. Free Radic Biol Med 35(4):327–340. 10.1016/S0891-5849(03)00280-6 [DOI] [PubMed] [Google Scholar]
  48. US Environmental Protection Agency (2025) National Ambient Air Quality Standards (NAAQS) for PM. https://www.epa.gov/pm-pollution/national-ambient-air-quality-standards-naaqs-pm
  49. Valavanidis A, Fiotakis K, Vlachogianni T (2008) Airborne particulate matter and human health: toxicological assessment and importance of size and composition of particles for oxidative damage and carcinogenic mechanisms. J Environ Sci Health Part C 26(4):339–362. 10.1080/10590500802494538 [DOI] [PubMed] [Google Scholar]
  50. Van Hulle C, Jonaitis EM, Betthauser TJ, Batrla R, Wild N, Kollmorgen G, Andreasson U, Okonkwo O, Bendlin BB, Asthana S, Carlsson CM, Johnson SC, Zetterberg H, Blennow K (2021) An examination of a novel multipanel of CSF biomarkers in the Alzheimer’s disease clinical and pathological continuum. Alzheimers Dement 17(3):431–445. 10.1002/alz.12204 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Wang X, Younan D, Millstein J, Petkus AJ, Garcia E, Beavers DP, Espeland MA, Chui HC, Resnick SM, Gatz M, Kaufman JD, Wellenius GA, Whitsel EA, Manson JE, Rapp SR, Chen J-C (2022) Association of improved air quality with lower dementia risk in older women. Proc Natl Acad Sci U S A 119(2):e2107833119. 10.1073/pnas.2107833119 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. World Health Organization (2002) Air Pollution. https://www.who.int/health-topics/air-pollution
  53. Xing Y-F, Xu Y-H, Shi M-H, Lian Y-X (2016) The impact of PM2.5 on the human respiratory system. J Thorac Dis. 10.3978/j.issn.2072-1439.2016.01.19 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Younan, D., Wang, X., Casanova, R., Barnard, R., Gaussoin, S. A., Saldana, S., … Women’s Health Initiative. (2021). PM2. 5 associated with gray matter atrophy reflecting increased Alzheimer risk in older women. Neurology, 96(8), e1190-e1201. 10.1212/WNL.0000000000011149 [DOI] [PMC free article] [PubMed]
  55. Zhang Y-L, Cao F (2015) Fine particulate matter (PM2.5) in China at a city level. Sci Rep 5(1):14884. 10.1038/srep14884 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Zhao N, Liu Y, Vanos JK, Cao G (2018) Day-of-week and seasonal patterns of PM2.5 concentrations over the United States: time-series analyses using the prophet procedure. Atmos Environ 192:116–127. 10.1016/j.atmosenv.2018.08.050 [Google Scholar]

Associated Data

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

Data Availability Statement

No datasets were generated or analysed during the current study.

Articles from Neurotoxicity Research are provided here courtesy of Springer

RESOURCES