Excessive Alcohol Use as a Risk Factor for Alzheimer’s Disease: Epidemiological and Preclinical Evidence

Abstract

Alcohol use has recently emerged as a modifiable risk factor for Alzheimer’s disease (AD). However, the neurobiological mechanisms by which alcohol interacts with AD pathogenesis remain poorly understood. In this chapter, we review the epidemiological and preclinical support for the interaction between alcohol use and AD. We hypothesize that alcohol use increases the rate of accumulation of specific AD-relevant pathologies during the prodromal phase and exacerbates dementia onset and progression. We find that alcohol consumption rates are increasing in adolescence, middle age, and aging populations. In tandem, rates of AD are also on the rise, potentially as a result of this increased alcohol use throughout the lifespan. We then review the biological processes in common between alcohol use disorder and AD as a means to uncover potential mechanisms by which they interact; these include oxidative stress, neuroimmune function, metabolism, pathogenic tauopathy development and spread, and neuronal excitatory/inhibitory balance (EIB). Finally, we provide some forward-thinking suggestions we believe this field should consider. In particular, the inclusion of alcohol use assessments in longitudinal studies of AD and more preclinical studies on alcohol’s impacts using better animal models of late-onset Alzheimer’s disease (LOAD).

10.1. Introduction

10.1.1. Alzheimer’s Disease

Alzheimer’s disease (AD), the most common type of dementia, is a chronic neurodegenerative disease predominantly defined by the accumulation of amyloid beta (Aβ) as extracellular plaques and pathologically modified protein tau as neurofibrillary tangles (NFTs; Knopman et al. 2021; Korczyn and Grinberg 2024). Clinically, AD is characterized by progressive memory impairments, difficulties in thinking, and neuropsychiatric symptoms (Apostolova et al. 2014; Li et al. 2014), including depression (Benoit et al. 2012; der Mussele et al. 2012), apathy (Ota et al. 2011), aggression (Guadagna et al. 2012), and psychosis (Gilley et al. 2004). AD is currently the seventh leading cause of death in the United States, but with global AD cases expected to reach 153 million by 2050 (GBD 2019 Dementia Forecasting Collaborators 2022; The Alzheimer’s Association 2024), it is beginning to outpace other causes of death and therefore remains a pressing public health concern.

Despite the high disease burden, preventative treatments for AD do not exist, partially due to an incomplete understanding of the complex etiology of the disease (Scheltens et al. 2021). Approximately 5% of AD cases are hereditary (familial AD) and primarily arise from mutations in genes leading to abnormal amyloid beta (Aβ) processing such as amyloid precursor protein (App), presenilin 1 (Ps1), and presenilin 2 (Ps2) (Lanoiselée et al. 2017). However, most AD diagnoses (>95%) occur later in life and are referred to as late-onset Alzheimer’s disease (LOAD) or sporadic AD. LOAD is characterized by an estimated 20-year preclinical phase of pathological aggregation (Vermunt et al. 2019), during which time changes in brain structure and function occur without any outward clinical symptoms (Fig. 10.1). Following the preclinical phase, the prodromal phase is characterized by progressive memory deficits in which mild cognitive impairment (MCI) is first diagnosed likely due to the now extensive pathology that has accumulated during the preclinical stage (Hampel and Lista 2012). Earlier AD prevalence studies used clinical symptoms to classify patients as having AD, but we have since learned that this approach led to a ~30% misdiagnosis rate compared to current blood/CSF and neuroimaging approaches (Hansson 2021; Hansson et al. 2022). Further advancements in these technologies have even allowed us to begin to mechanistically characterize each of the phases of AD: preclinical, prodromal (MCI), and AD (Jack et al. 2018; Janelidze et al. 2020; Moscoso et al. 2021; Leuzy et al. 2022), as well as to determine the clinical efficacy of AD therapeutic candidates (Angioni et al. 2022; Hansson et al. 2023). This work continues, but has thus far identified a host of pathological changes involved in LOAD progression, including disruptions in amyloid processing leading to amyloid beta (Aβ) plaque accumulation (Hampel et al. 2021), changes in cerebral vascularization (Klohs 2020; Fisher et al. 2022; Cai et al. 2023), neuroinflammation (Franceschi and Campisi 2014; Pinti et al. 2016), metabolism (Maghsudi et al. 2020), autophagy (Aman et al. 2021; Caponio et al. 2022), pathological tau aggregation (Guillozet et al. 2003; Binder et al. 2005; Chiu et al. 2017) and spread (Walsh and Selkoe 2016; Wegmann et al. 2019), and alterations in neuronal excitatory/inhibitory balance (EIB) (Kang et al. 2020; Radulescu et al. 2023). Thus, there are many potential neurobiological mechanisms regulating LOAD, each of which has the potential to influence LOAD progression.

Fig. 10.1.

Pathogenesis of late-onset Alzheimer’s disease (LOAD, AD) and the proposed impact of alcohol misuse. Alcohol misuse throughout key developmental periods could both accelerate onset and exacerbate progression of AD measured by the biomarkers in the ATN framework: the amyloid beta plaques (A), pathological tau (T), and neurodegeneration (N). Development of these pathologic factors often develops before the outward manifestations of clinically diagnosed AD dementia (i.e., cognitive decline). There is an approximate 10–20-year period of preclinical changes, where cognition is not impacted. The Prodromal phase (Mild Cognitive Impairment, MCI) is characterized by the first appearance of cognitive decline. Dementia of the Alzheimer’s type can exist as mild, moderate, and severe. Our hypothesis is that heavy alcohol use (solid line) can exacerbate and accelerate onset of AD (dotted line indicates “normal” AD pathogenesis), as indicated by the arrows

One approach used to identify mechanisms mediating LOAD is via genome-wide association studies (GWAS). These population genetic studies identify single-nucleotide polymorphisms (SNPs) statistically associated with certain diseases, like LOAD, to ferret out specific gene products (proteins) highly likely to be involved in the disease of interest (ADGC et al. 2019; Bellenguez et al. 2022). These studies have identified alleles associated with LOAD that involve neuroimmune function (Jansen et al. 2019; Andrews et al. 2020; Jorfi et al. 2023), particularly microglia (Thorlakur et al. 2013; ARUK Consortium et al. 2017; Efthymiou and Goate 2017), autophagy dysfunction (Barrachina et al. 2006; Ginsberg et al. 2010; Nixon 2013; Wolfe et al. 2013; Peric and Annaert 2015; Almeida et al. 2020; Lee et al. 2022), and lipid metabolism (Katzov et al. 2004; Mace et al. 2005; Liu et al. 2013; Huang and Mahley 2014), as well as MAPT (microtubule-associated protein tau) gene expression and circulating tau levels (ADGC et al. 2019; Sarnowski et al. 2022) and excitatory/inhibitory neuron populations (Gazestani et al. 2023). One GWAS for AD/dementia identified over 90 independent variants across 75 susceptibility loci, including genes enriched in amyloid plaque and NFT formation, cholesterol metabolism, endocytosis/phagocytosis, and innate immune function (Andrews et al. 2023). While these GWAS studies corroborate mechanisms discussed in this chapter, they also collectively emphasize the complexity of AD pathogenesis and progression.

In parallel to exploring the molecular genetic underpinnings of LOAD using GWAS, the field has leveraged advances in assessing brain function to explore the diseased brain directly. Positron emission tomography (PET), functional magnetic resonance imaging (MRI), and electroencephalogram (EEG) have been the tools of choice. Work done in this space has revealed that tau-PET alone, not Aβ-PET or anatomical MRI, accurately predicts the conversion from MCI to all-cause dementia (Groot et al. 2024). Additional evidence suggests that microglial-mediated neuroinflammation may play a role in the development and progression of AD. A translocator protein (TSPO) PET ligand associated with microglia was found to potentially act as a connective factor between the formation of Aβ plaques and the spread of pathogenic tau (Rossano et al. 2024). Furthermore, TSPO expression was linked to the severity of AD, brain volume, cognitive decline, and the accumulation of both Aβ plaques and pTau (Rossano et al. 2024). Long-term EEG monitoring studies have also observed an increase in the frequency of subclinical epileptiform discharge (SED) in patients compared with non-demented controls (Vossel et al. 2013, 2016, 2017; Beagle et al. 2017; Lam et al. 2020; Horvath et al. 2021). Interestingly, SEDs and other epileptiform activity are associated with earlier disease onset (Amatniek et al. 2006; Scarmeas et al. 2009; Irizarry et al. 2012; Vossel et al. 2013) as well as accelerated disease progression (Horvath et al. 2021; Yeh et al. 2022; Kamondi et al. 2024). Thus, the emergence of neuronal hyperexcitability observed in LOAD may be driven by the accumulation of Aβ plaques and/or pathological tau, or more insidiously, the instigator of pathological processes that facilitate the production, spread, and accumulation of these pathological proteins (Harris et al. 2020).

Using the revised ATN (Amyloid/Tau/Neurodegeneration) classification system (Jack et al. 2018; van der Flier and Scheltens 2022) (Fig. 10.1) and armed with new molecular tools (single molecule array, SiMoA) (Li and Mielke 2019; Chen et al. 2020; Emeršič et al. 2020; Hanes et al. 2020; Karikari et al. 2020; Kvartsberg et al. 2020; Meyer et al. 2020a; Thijssen et al. 2020; Ashton et al. 2021; Chatterjee et al. 2021; Fowler et al. 2022; Thomas et al. 2022), researchers have begun to evaluate the accuracy of blood-based biomarkers which may outperform traditional cerebrospinal fluid (CSF) metrics for tracking both disease progression and success of clinical therapeutics (Barthélemy et al. 2024). Notably, these sensitive clinical blood assessments have substantiated the traditional ATN framework by being able to robustly measure Aβ plaques (A) (Meyer et al. 2020b), multiple pathogenic species of tau (pT181 (Mielke et al. 2018; Qin et al. 2022; Tzartos et al. 2022), pT217 (Barthélemy et al. 2020, 2023), pT231 (Ashton et al. 2021; Wisch et al. 2024), T), as well as, total tau (Dage et al. 2016), and neurofilament light (NfL) (Bacioglu et al. 2016) to measure neurodegeneration (N). Notably, they have continued to single out pathologically modified tau (pTau) as the strongest predictor of conversion from MCI to AD (Barthélemy et al. 2020, 2023; Karikari et al. 2020; Ashton et al. 2021; Moscoso et al. 2021). Importantly, these blood-based biomarkers are beginning to work their way into clinical practice (Olsson et al. 2016) and clinical trials (Angioni et al. 2022)—a trend that will hopefully continue to shed light on LOAD progression.

Despite our limited current understanding of LOAD progression, there is an approximate 20-year preclinical phase prior to outward neurological symptom presentation, i.e., cognitive decline (Jack et al. 2018; Long and Holtzman 2019) (Fig. 10.1) where interventions and/or environmental and lifestyle factors could greatly influence AD risk (Eid et al. 2019; Frigerio et al. 2019; Yu et al. 2020; Wieckowska-Gacek et al. 2021). For example, smoking (Peters et al. 2008), heavy alcohol use (Schwarzinger et al. 2018; Rehm et al. 2019), diabetes (Athanasaki et al. 2022), hypertension (Tang et al. 2023), and obesity (Pedditzi et al. 2016) have all been demonstrated to potentially increase the risk of developing AD or accelerating the onset of AD (Fig. 10.1). On the other hand, engaging in a combination of protective activities, maintaining an active social life (Shafighi et al. 2023), and chronic disease maintenance have been shown to collectively reduce the risk of AD by up to 40–60% (Livingston et al. 2020). Unfortunately, by the time individuals are diagnosed with mild cognitive impairment (MCI) due to AD (Bradfield and Ames 2020), which can be defined as individuals performing at least 1.5 standard deviations below normal on memory tasks, they have most likely already accumulated significant AD-relevant pathology (Fig. 10.1). Therefore, developing a better understanding of how specific common lifestyle factors, for example alcohol use, contribute to AD pathogenesis and progression is critical to understanding disease etiology and uncovering novel therapeutic targets.

10.1.2. Alcohol Use Throughout the Lifespan

Alcohol is widely accessible and frequently used, with over 84% of adults aged 18 and older reporting lifetime use (NSDUH 2023). Alcohol misuse is defined as any drinking behavior that jeopardizes one’s well-being or the well-being of others and includes two primary patterns of excessive alcohol consumption: heavy alcohol use (>4 drinks on any day or >14 drinks per week for men, and >3 drinks on any day or >7 drinks per week for women) and binge drinking (4–5 drinks consumed within 2 hours) (NIAAA 2024). Alcohol use disorder (AUD), a serious medical condition resulting from patterns of excessive use, is characterized by an impaired ability to reduce or cease alcohol use despite negative social, occupational, or health consequences. In 2021, nearly 28.6 million adults aged 18 and older in the United States had an AUD (NSDUH 2021), yet only 1 in 10 individuals seek and receive treatment (Mintz et al. 2023). Further, harmful drinking behaviors are not exclusive to people suffering from AUD, as described in detail below.

While alcohol misuse is typically associated with younger adults (18–25 years old), as nearly one-third of individuals in this age group report past month heavy or binge drinking (Fig. 10.2b), current data suggest that heavy alcohol use and binge drinking is increasing among older demographics including middle-aged adults (35–50 years) (Fig. 10.2b). Similar trends are being observed in older populations. A meta-analysis that analyzed surveys from 2000 to 2016 revealed significant increases in alcohol use among adults aged 50 and above, with a notable rise in binge drinking among those 30 and older, particularly in the oldest age group (65+ years) (Grucza et al. 2018). The most recent NSDUH 2022 data also found that 43.4% of people 65+ consumed alcohol within the past month, with 9.7% reporting binge drinking, and 2.4% reporting heavy alcohol use (Fig. 10.2b) (NSDUH 2022/2012). Over the last decade, the percentage of individuals 65+ consuming alcohol in the past month increased from 41.2% in 2012 (Fig. 10.3a) to 43.4% in 2022 (Fig. 10.3b), with the most substantial rise occurring within the binge alcohol use category from 6.2% in 2012 to 7.3% in 2022 (Fig. 10.3a, b). Interestingly, although men have historically consumed more alcohol than women, this gender gap in alcohol misuse is narrowing across all age groups (Keyes et al. 2019). As but one example, from 1997 to 2014 women over 60 experienced the largest increase in binge drinking behavior (Breslow et al. 2017). To be clear, adults who engage in excessive drinking are increasing in age, and excessive drinking rates continue to increase among older adults. Coupled with the fact that fewer than 15% of individuals with an AUD seek and receive treatment (Degenhardt et al. 2017; Glantz et al. 2020; GBD 2021 Diseases and Injuries Collaborators et al. 2024), it is imperative that we develop a better understanding of the risks alcohol consumption poses to brain health, particularly neurodegenerative diseases such as LOAD, hereafter referred to as Alzheimer’s disease (AD).

Fig. 10.2.

Epidemiology of AD and alcohol use. (a) The percentage of people 65+ living with late-onset AD (LOAD) broken down by age, data from 2024. (b) Percentage of individuals reporting alcohol use throughout the lifespan, SAMHSA data from Table 2.9B, 2022

Fig. 10.3.

The percentage of 65+ years old and type of alcohol use throughout the last month demonstrates a substantial increase from 2012 to 2022 (NSDUH 2022/2012). (a) Percentage of individuals 65+ in 2012 shows that 58.8% did not consume alcohol within the last month, while 41.2% of people engaged in alcohol use. 8.2% of those 65 and older engaged in binge drinking or heavy alcohol use. (b) Percentage of individuals 65+ consuming alcohol (43.4%); 9.7% reported engaging in binge alcohol use or heavy alcohol use

10.2. Alcohol and Alzheimer’s Disease: Epidemiological Evidence for Alcohol Misuse as an AD Risk Factor

Unraveling the influence of alcohol on the development and progression of AD is challenging; in part because heavy alcohol use was previously used as exclusion criterion for AD diagnosis (Tyas 2001). Although chronic heavy drinking can cause unique forms of neurodegeneration, like alcohol-related dementia (Moriyama et al. 2006) or Wernicke’s-Korsakoff syndrome (Oudman et al. 2022), evidence from the past two decades supports that alcohol misuse throughout the lifespan is a modifiable risk factor for AD (Schwarzinger et al. 2018; Rehm et al. 2019). Adolescence is a key developmental period wherein high levels of alcohol consumption are often first experienced (Spear 2013). It is also a time of rapid brain development (Spear 2000, 2018). Yet, there is no data on how adolescent alcohol exposure impacts the risk of AD development later in life in clinical populations. However, teenagers who participate in alcohol misuse exhibit decreased brain volume in areas typically affected in AD, including the frontal and temporal lobes (Phillips et al. 2021), as well as impaired attention span and difficulties in memory (Lees et al. 2020). Because impaired cognitive performance in adolescence is associated with AD-related dementia later in life (Moceri et al. 2000; Huang et al. 2018), alcohol-related brain exposure and/or damage during adolescence may possibly promote AD pathogenesis.

The field has made more headway evaluating the impact of alcohol use on adults, in large part because many more adults drink alcohol, but also because there are fewer ethical considerations than working with minors. To this end, a landmark retrospective cohort study of adult men and women 20 years or older in France found that an alcohol use disorder (AUD) in early adulthood or middle age was the strongest modifiable risk factor for the development of dementia (hazard ratio/HR: ~3.3) (Schwarzinger et al. 2018). In a separate study, the presence of alcohol-related brain damage was associated with an earlier dementia onset (Zhao et al. 2024), with AUD being associated with an increased risk of both vascular (HR: 2.3) and other dementias, including AD (HR: 2.36). Other work suggests heavy alcohol use is associated with risk for AD independent of AUD, although AUD diagnosis was not defined in the study (Jeon et al. 2023). In this large retrospective cohort study assessing the baseline and change in alcohol consumption patterns over time, adults 40 years and older who engaged in sustained heavy drinking (i.e., >30 g/day) had an 8% increased risk for the development of subsequent AD 6 years earlier compared to participants who drank moderately or did not drink at all (Jeon et al. 2023). Furthermore, participants who increased their drinking from study-defined mild/moderate levels to heavy drinking also saw an increase in all-cause dementia (HR: 1.37) (Jeon et al. 2023). In studies with a longer follow-up period, such as the HUNT study, participants who reported drinking 5 or more times 2 weeks prior to the start of the study (during young adulthood-midlife) had an increased risk of AD diagnosis later in life compared to individuals who drank infrequently (Langballe et al. 2015). Together, these data suggest that heavy alcohol use, perhaps from late adolescence through midlife, promotes the development of AD and other related dementias (Fig. 10.4).

Fig. 10.4.

Overlap of alcohol use history within the population of patients with an Alzheimer’s disease diagnosis. Using the NSDUH 2022 percentages (Figs. 10.2a and 10.3b) we can speculate that within the percentages of US adults 65+ with AD, there are approximately 10% that are 65–74 years old with a history of alcohol use, 2% with a history of binge drinking, and more than 0.6% with a history of heavy alcohol use. In patients with an Alzheimer’s diagnosis that are 75–84 years old, that percentage is troublingly higher where 13% report alcohol consumption, nearly 3% report consuming binge alcohol, and nearly 1% engage in heavy alcohol use. 85+ years and older with a diagnosis also have astonishingly high rates of alcohol use (11% consume alcohol, 2.5% consume binge alcohol, and nearly 1% engage in heavy alcohol use)

A growing body of literature also indicates that heavy alcohol use in advanced age increases the risk for AD (Mukamal et al. 2003; Koch et al. 2019). Older adults aged 65 years and older who drank more than 14 alcoholic drinks per week were found to have increased odds for AD and all-cause dementia compared to older adults who had low or moderate drinking behaviors (Mukamal et al. 2003). These data were corroborated by findings in participants with MCI at baseline, which found that those reporting drinking 14 or more drinks/week had an increased hazard ratio (HR) for dementia (HR: 1.72) than those drinking less than 1.0/week (Koch et al. 2019). However, in addition to heavy drinking increasing the risk for disease development, it may also accelerate disease onset. A retrospective study in patients with an AD diagnosis suggests individuals who had 2 or more drinks a day exhibited disease onset 4 years earlier than those compared to individuals who had less than 2 drinks a day (Harwood et al. 2010). These data from adolescence to advanced age together support alcohol use at all developmental time periods as having the potential to impact AD diagnosis later in life.

Since age is the largest non-modifiable risk factor for AD, and by 2030 all members of the baby-boom generation will be at least 65+ (Guerreiro and Bras 2015), there is a high potential for a wave of new AD diagnosis. The current age breakdown of those 65 and older with an AD diagnosis is as follows: 1.83 million are 65–74 years old, 2.67 million are 75–84, and 2.42 are 85+ (Rajan et al. 2021) (Fig. 10.2a). However, the population of Americans aged 65 and older is estimated to rise from 58 million in 2022 to 83 million by 2050 (The Alzheimer’s Association 2024). Thus, taking the percentages of individuals reporting alcohol use within the last month (Fig. 10.3b data), and extrapolating to the percentage of individuals within each age group who have a diagnosis of AD (Fig. 10.2a), may provide some insight into our epidemiological future. To this end, it is likely that a growing number of individuals with an AD diagnosis will have a history of alcohol use that includes binge drinking and/or heavy alcohol use (Fig. 10.4).

Despite the negative outcomes alcohol has on brain health and likely AD outcomes in particular, a small set of studies have suggested decreased rates of AD with low levels of alcohol use (Sabia et al. 2018; Jeon et al. 2023). These results may be mediated by a J-shaped curve for alcohol as a risk factor for AD, with all but very low exposure increasing AD risk. However, interpretation of these studies is limited by several confounding factors. First, there is no consistent definition of “light,” “moderate,” and “heavy” across many studies, and the rate of drinking and time since last drink among participants are not regularly reported. Second, these studies do not account for the cause of abstinence, which may be in direct response to an AUD diagnosis, or other illness or confounding factors known to influence both AD development and drinking behavior, such as education, social activity, mental health, and diet. That even relatively low levels of alcohol consumption have recently been found to be associated with reduced brain volume is difficult to reconcile with the studies listed above (Topiwala et al. 2022). Regardless, the epidemiological observations that low alcohol-consuming individuals may be less prone to developing AD, while high alcohol-consuming individuals have an increased risk of developing AD, directly support the hypothesis that AD risk scales as a function of relative alcohol exposure.

10.3. Alcohol and Alzheimer’s Disease: Shared Disruption of Biological Processes

Alcohol-related brain damage and AD pathogenesis lead to similar pathological changes in oxidative stress (Gella and Durany 2009; Qin and Crews 2012a), neuroimmune function (Eikelenboom et al. 2010; Pascual et al. 2021), metabolism (Volkow et al. 2015; Butterfield and Halliwell 2019; Tomasi et al. 2019; Popova et al. 2024), pathogenic tauopathy development and spread (Liu et al. 2012; Mohamed et al. 2013; DeVos et al. 2018; Gibbons et al. 2018; Bell et al. 2020; Gu and Liu 2020; Annadurai et al. 2021; Downs et al. 2022; Tucker et al. 2022), and neuronal excitatory/inhibitory balance (EIB) (Born 2015; Anastacio et al. 2022; Huang et al. 2022; Alberto et al. 2023; Barbour et al. 2023). Importantly, all of these pathologies promote neurodegeneration (Andersen 2004; Pasantes-Morales and Tuz 2006; González et al. 2014; Jha et al. 2017; Strang et al. 2019). Thus, alcohol misuse may promote AD pathogenesis by initiating similar insults and/or by exacerbating and facilitating those caused directly by AD. Given recent evidence that the brain is more vulnerable to alcohol-related damage from adolescence and into aging than we had previously appreciated (Coleman et al. 2011; Vetreno and Crews 2012; Coleman et al. 2014; Vetreno et al. 2014; Salling et al. 2018; Tapia-Rojas et al. 2018; Barnett et al. 2022; Khan et al. 2023; Anton et al. 2024), exposure to alcohol at virtually all stages of life has the potential to impact these molecular mechanisms to influence AD outcomes. In the next section, we will discuss the shared mechanisms of pathogenesis in AD and alcohol misuse throughout the lifespan.

10.3.1. Oxidative Stress

Alcohol (i.e., ethanol) freely passes through the blood–brain barrier (BBB) and is toxic to neurons at high concentrations (Zimatkin and Deitrich 1997; Wilson and Matschinsky 2020). Although the liver is the primary site of ethanol metabolism, ethanol is also locally metabolized in the brain via catalase and CYP2E1 leading to the creation of acetaldehyde, a toxic intermediate (Aragon et al. 1992). Regardless of where it is produced, acetaldehydes are damaging to the neural microenvironment through the formation of protein and DNA adducts, and the production of oxidative species during its metabolism into acetate by acetaldehyde dehydrogenase (Nakamura et al. 2003). Oxidizing agents such as hydrogen peroxide (H₂O₂) and free radicals (superoxide and hydroxyl radical) are also generated as by-products of CYP2E1 and catalase activity (Zimatkin and Deitrich 1997). Free radicals are harmful to the neurons through disrupting the cell membranes and consequential lipid peroxidation (Hernández et al. 2016). Ethanol also induces permeability of mitochondrial membranes and leakage of superoxide, which further perpetuates oxidative stress and can initiate apoptosis (González et al. 2007). Thus, ethanol, which is mainly metabolized in the liver, can also be locally metabolized in the brain and quite toxic to neurons, as well as other cell types.

Although all CNS cell types are vulnerable to ethanol-induced oxidative stress, glial cells play an important role in the production of oxidative factors because of their specialized functions (Montoliu et al. 1995; Qin and Crews 2012a). Repeated binge ethanol exposure in mice activates NADPH oxidase (NOX) in microglia, the resident macrophages of the CNS, leading to superoxide production and subsequent neurodegeneration (Qin and Crews 2012a). Astrocytes are an important site of ethanol metabolism in the brain and thus are a potent source of superoxide, further elevating oxidative-induced damage (Montoliu et al. 1995). Further, the impact of ethanol-related oxidative stress is exaggerated by depletion of the brain’s natural antioxidant defense system by ethanol exposure (Reddy et al. 1999). Given the numerous pathways through which ethanol-induced oxidative stress can be amplified, its presence is unsurprising in rodent models of excessive ethanol exposure and in post-mortem human brain tissue from individuals with AUDs (Montoliu et al. 1995; Reddy et al. 1999; Qin and Crews 2012a).

Oxidative stress is also believed to be an early participant in AD pathogenesis (Rapoport 2003). Markers of lipid peroxidation, protein oxidation, and DNA oxidation are increased in the brains of individuals with AD compared to healthy controls (Gella and Durany 2009). Recent data have shown that this may be due in part to the dysregulation of a few key mitochondria proteins known to regulate oxidative stress, so much so that the degree of expression is able to predict LOAD pathogenesis (Yan et al. 2024). A mechanistic role for oxidative stress has also been demonstrated in vitro, where neuroblastoma cell lines treated with Aß under oxidative conditions were found to promote Aß aggregation (Zheng et al. 2006). On the other hand, antioxidant intervention in vivo mitigates neuropathology and cognitive dysfunction in aged 3x-Tg mice (Young and Franklin 2019). And finally, to make matters worse, the presence of Aß promotes oxidative stress itself directly through inhibition of mitochondria superoxide dismutase, creating a damaging positive feedback loop of oxidative stress in AD (Tamagno et al. 2021). Additional work in both a mouse model of tauopathy and antecedent AD brain tissue found that soluble tau pathology could also exacerbate oxidative stress through mitochondrial disruption (Kopeikina et al. 2011). Importantly, oxidative stress has been observed in many primary tauopathies, like Pick’s disease, corticobasal degeneration (Castellani et al. 1995), and frontal-temporal dementia (Martínez et al. 2008), as well as secondary tauopathies, like AD (Stamer et al. 2002; David et al. 2005). These data suggest that the presence of both pathologic species relevant to AD, namely Aß and pathological tau, can exacerbate and are mediated by oxidative stress.

Thus, both excessive alcohol use and AD prominently feature damage from oxidative stress. As oxidative stress is a central mechanism of ethanol-related neuronal injury in adolescence (Vetreno et al. 2014; Pelicao et al. 2016; Barnett et al. 2022), and adolescence is the time excessive alcohol consumption is most initiated, it remains possible that cumulative oxidative stress beginning in adolescence exhausts neural resources that would otherwise be used to mitigate AD-induced oxidative stress. Likewise, because impaired mitochondria function and antioxidant depletion are a hallmark of brain aging (Ionescu-Tucker and Cotman 2021), any avenue by which ethanol increases oxidative load would likely be exacerbated with increasing age. Despite the overlap in oxidative stress, to date clinical trials using antioxidants for AD have not been successful (Polidori and Nelles 2014). While more research in this area is needed, to the extent that AD outcomes are impacted by oxidative stress, ethanol-induced increases in oxidative stress remain an important potential contributor to AD risk.

10.3.2. Neuroimmune Dysregulation

Heightened neuroinflammation is a critical factor that facilitates pathogenesis of both AD and heavy alcohol use (Cribbs et al. 2012; Qin et al. 2021; Barnett et al. 2022). In response to alcohol, damaged neurons release danger-associated molecular patterns (DAMPs), such as high mobility group box 1 (HMGB1) (Crews et al. 2013; Coleman et al. 2017). DAMPs, like HMGB1, bind to pattern recognition receptors (PRRs) including toll like receptors (TLRs) and receptors for advanced glycation end products (RAGE) on neurons and glial cells. TLR and RAGE activation can initiate pro-inflammatory signaling cascades including nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κb) and inflammasomes (Montesinos et al. 2016). These pathways culminate in the production of several pro-inflammatory cytokines (TNF-α, IL-1β, IL-6) (Alfonso-Loeches et al. 2010; Coleman et al. 2017). Prolonged production of pro-inflammatory cytokines from repeated binge or chronic ethanol exposure also upregulates cytokine receptors (TNF-R, IL-1R) and TLRs to sustain these pro-inflammatory signaling loops (Qin and Crews 2012b; Vetreno and Crews 2012). Ultimately, the sustained production of pro-inflammatory cytokines can lead to neuronal death via apoptosis (Liu et al. 2021; Qin et al. 2021). The pro-inflammatory effects of ethanol are evident in mouse and rat models of chronic ethanol exposure and binge ethanol exposure (Crews and Vetreno 2014; Erickson et al. 2018; Pascual et al. 2021), as well as in post-mortem brain tissue of individuals with AUD, which specifically found increases in HMGB1, TLR, cytokines, and neurodegeneration (Qin and Crews 2012a; Crews et al. 2013; Coleman et al. 2017).

Similar innate immune pathways are implicated in AD. RAGE receptor expression levels in the hippocampus of AD patients correlates with more severe disease pathology (Lue et al. 2001), and microarray analysis of human brain tissue reveals an increase in innate immune signaling genes in the brains of people with AD compared to healthy controls (Cribbs et al. 2012). Increased expression of DAMPs, TLRs, and cytokines is also elevated in the brains of aged people compared to young controls and may suggest dysfunctional innate immune signaling precedes overt AD pathology (Cribbs et al. 2012). Heightened indicators of neuroinflammation are present in mouse models of AD as well, and suppression of pro-inflammatory cytokines slows Aβ accumulation and neurotoxicity (McAlpine et al. 2009; Heneka et al. 2013). Notably, the presence of pathologic tau is known to accelerate neuroinflammation through NLRP3 inflammasome signaling (Heneka et al. 2013; Ising et al. 2019; Jiang et al. 2021), and neuroinflammation, in turn, is known to influence and increase pathologic tau species (Bhaskar et al. 2010; Weston et al. 2021). HMGB1 promotes neurodegeneration in transgenic mouse models of AD without affecting plaque deposition, suggesting these immune pathways might also worsen neurodegeneration parallel to Aβ (Fujita et al. 2016). However, the release of HMGB1 promotes oligomeric aggregation of pathogenic tau which can, in turn, induce cellular senescence, tauopathy progression, and cognitive deficits (Gaikwad et al. 2021). Mitigating neuroinflammation through anti-inflammatory therapies like minocycline reduces neuropathology in mouse models of AD, further supporting a mechanistic role of the innate immune system in disease progression (Stirling et al. 2005; Seabrook et al. 2006; Choi et al. 2007; Noble et al. 2009). A few epidemiological studies (Delanty and Vaughan 1998; in’t Veld et al. 2001; Zandi et al. 2002; Kotilinek et al. 2008; Vlad et al. 2008; Imbimbo et al. 2010), but not all (ADAPT Research Group et al. 2007; Szekely et al. 2008), have demonstrated a reduced prevalence of AD among users of immunosuppressants like non-steroidal anti-inflammatory drugs (NSAIDs), with the strongest effect seeming to have been in patients on the longest duration (Delanty and Vaughan 1998; Etminan et al. 2003; Vlad et al. 2008). To that end, a 2-year double-masked pharmaco-prevention trial, which enrolled 195 AD family history-positive elderly with a mean age of 63 years, administered naproxen 2x daily or placebo to their clinical trial participants. Unfortunately, naproxen-treated individuals had more adverse events, and no reduction in the rate of AD progression compared to placebo controls and no notable treatment effects on any underlying neuroimaging or CSF biomarkers present (Meyer et al. 2019). Despite these null findings, interest remains high in using NSAIDs as potential AD treatments (Hershey and Lipton 2019; Rivers-Auty et al. 2020).

Microglia, the brain’s resident macrophages, play a critical role in physiology and disease (Wolf et al. 2016). They are particularly important orchestrators of neuroinflammation in several neurodegenerative disorders, like AD (Hickman et al. 2018; Gao et al. 2023). Single-cell RNAseq revealed the remarkable complexity of microglial responses that can occur in a wide variety of neurodegenerative disease states (Provenzano et al. 2021). Ethanol acts as a DAMP, leading microglia to acquire complex phenotypes that can either facilitate recovery (Marshall et al. 2013) or promote injury following ethanol exposure (Chastain and Sarkar 2014). The presence of ethanol and DAMPs induces pro-inflammatory polarization of microglia, characterized by the upregulation of Iba-1 expression, increased cell surface receptors like CD68, and the production of pro-inflammatory cytokines (Chastain and Sarkar 2014). Altered microglia reactivity to ethanol is believed to promote neurogenesis during recovery from ethanol exposure (Nixon et al. 2008). Additionally, ethanol has been shown to induce microglial phagocytosis, which may aid in debris clearance and recovery following ethanol-induced neurotoxicity (Fernandez-Lizarbe et al. 2009). However, pro-inflammatory microglia are also associated with neuronal death in models of chronic or repeated binge ethanol exposure in mice and rats (Chastain and Sarkar 2014; Yang et al. 2014). Microglia contribute to neuronal death in these models through potent production of pro-inflammatory cytokines. For instance, depletion of microglia in vivo in mice and ex vivo in rat brain slice cultures mitigates both pro-inflammatory cytokine expression and ethanol-related injury (Walter et al. 2017; Coleman et al. 2020). Furthermore, in vivo and ex vivo ethanol exposure decreases microglia phagocytosis of Aβ in mice and rats (Marsland et al. 2022). These data suggest dysregulated microglia inflammatory responses to ethanol promote neurotoxicity.

Microglia also play critical roles in both promoting and mitigating AD pathology. Similar to ethanol exposure, microglia acquire diverse phenotypes throughout both aging (Provenzano et al. 2021) and AD progression (Hamelin et al. 2016; Mathys et al. 2017; Miao et al. 2023). In AD, damage-associated microglia (DAMs) upregulate genes associated with phagocytic activity aimed at reducing Aβ (Keren-Shaul et al. 2017). Soluble and oligomeric forms of Aβ as well as pathologic tau can induce pro-inflammatory polarization of microglia to promote disease progression (Brelstaff et al. 2021; Bo et al. 1995; Walter et al. 2007; Jin et al. 2008; Song et al. 2011; Solito and Sastre 2012; Dalgediene et al. 2013; Majerova et al. 2014; Koss et al. 2016; Yang et al. 2017; Luciunaite et al. 2020; Zhang et al. 2024). However, in the proximity of Aβ plaques, microglia can enhance phagocytic activity (Hamelin et al. 2016; Wang et al. 2016; Keren-Shaul et al. 2017; Ennerfelt et al. 2022). Microglia may also play a central role in spreading pathogenic tau, since their depletion suppresses tau propagation (Asai et al. 2015). Notably, both Aβ and tau may act together to induce unique microglia subtypes that could define both early- and late-stage AD (Kim et al. 2022). As AD advances, chronic stimulation of microglia by Aβ can ultimately impair phagocytosis and cause release of pro-inflammatory cytokines that promote tau pathology and neurodegeneration (Miao et al. 2023). Pathological tau can also prime neuroinflammation through NF-κb and IL-1β signaling (Jiang et al. 2021). Although these data suggest microglia dysfunction is driven by AD neuropathology, additional evidence shows microglia dysregulation precedes neuropathology. For example, genes associated with impaired microglia phagocytosis are identified as strong risk factors for LOAD (Efthymiou and Goate 2017; Novikova et al. 2019). Additionally, microglia lose their dynamic functions with advanced aging and respond to stimuli with exaggerated pro-inflammatory responses but reduced phagocytic capacity (i.e., “priming”) (Norden and Godbout 2013). Therefore, additional activation by environmental exposures, such as heavy alcohol use, may exaggerate microglia priming and leave these cells in a chronically stimulated state that could accelerate or promote AD pathology, increasing risk for the disease.

Emerging preclinical evidence suggests that alcohol exposure in adolescence or advanced age has profound impacts on neuroinflammation and associated AD-like neuropathology. Intermittent ethanol exposure in a 3xTg genetic mouse model of AD elevates cytokine expression, markers of microglia reactivity, and promotes Aβ accumulation and neuronal death in adulthood compared to vehicle-exposed mice (Barnett et al. 2022). Treatment with the anti-inflammatory compound minocycline prevented this enhanced pathology. These data suggest adolescent exposure has lasting impacts on AD progression into adulthood that are neuroimmune-dependent. A persistent increase in innate immune genes also occurs in wild-type mice exposed to binge ethanol (Vetreno and Crews 2012), which can lead to deficits in cholinergic neuronal populations and cognitive inflexibility later in adulthood (Coleman et al. 2011, 2014). This is consistent with studies in aged wild-type mice and rats, which consistently find that ethanol increases markers of neuroinflammation and neuronal death more so in aged brains compared to young brains (Kane et al. 2013; Marsland et al. 2022; Anton et al. 2024). These studies suggest a heightened risk of neurodegeneration due to ethanol-induced increases in neuroinflammation in advanced age. Together these data find differences in the effects of ethanol exposure on neuroimmune function positioned to exacerbate AD outcomes.

10.3.3. Metabolic Dysfunction in Brain with AUD and AD: A Possible Link with Neuroinflammation

Both AUD and AD feature metabolic dysfunction. Glucose is the essential energy source of the brain (Mergenthaler et al. 2013), and in the adult brain, neurons have the biggest appetite (Howarth et al. 2012). Fortunately, fluorodeoxyglucose positron emission tomography (FDG-PET) has evolved to be a sensitive neuroimaging biomarker. Studies using FDG-PET have found hypometabolism in key brain regions over the course of both AUD (Thanos et al. 2008; Tomasi et al. 2019) and AD (Mosconi et al. 2008; Fouquet et al. 2009; Kobylecki et al. 2015). In human AUD brain, the loss in FDG-PET correlates strongly with alcohol-related brain damage across brain regions (Tomasi et al. 2019). In AD, reductions in FDG uptake predict the conversion from mild cognitive impairment to AD (Fouquet et al. 2009), although recent work suggests this may not be the case (Smailagic et al. 2018). There is also evidence that Aβ and tau drive cognitive decline early in AD, while glucose hypometabolism drives decline at later stages of the disease (Hammond et al. 2020). These reductions in glucose uptake and/or utilization suggest a shift in the primary metabolic pathways used by neurons in both diseases.

In the setting of AUD, there is evidence to suggest a shift toward acetate metabolism (Volkow et al. 2015), likely due to the oxidative metabolism of alcohol. The oxidative metabolism of alcohol results in the production of acetate (Wilson and Matschinsky 2020), which is either converted to acetyl-CoA for use in the tricarboxylic acid cycle (TCA) or converted to malonyl-CoA by acetyl-CoA carboxylase to promote lipogenesis. Interestingly, although it is well known that alcohol promotes lipidosis in the peripheral tissues such as the liver, we recently found that ethanol-induced neuronal lipidosis promotes AD pathology, with pro-inflammatory microglia driving neuronal lipidosis (Barnett et al. 2024). Thus, alcohol-induced alterations to metabolic processes may directly exacerbate AD pathology.

The underlying causes of the metabolic shifts observed in AUD and AD warrant further investigation. Altered glucose utilization or a shift to an alternative energy source such as acetate or lipids can lead to neuronal dysfunction with impairments in neurotransmitter production, synaptic function, and neuronal hyperexcitability (Rorbach-Dolata and Piwowar 2019) that can result in neuron death and cognitive abnormalities (Suh et al. 2003; McDonald et al. 2023). We believe studying the impact of pro-inflammatory microglia on neuronal function is of particular importance following recent evidence that neurons can perform glycolysis independently (Li et al. 2023) and that microglia may take up more glucose than astrocytes from the periphery, which can also be detected using FDG-PET (Xiang et al. 2021; Gnorich et al. 2023). Further, microglia take up glucose at high levels in vivo (Xiang et al. 2021; Gnorich et al. 2023), and pro-inflammatory microglia engage in high levels of glycolysis (Orihuela et al. 2016). Changes in glial cell activation state may therefore directly alter neuronal metabolism and activity during disease progression, and as such may represent a viable approach to mitigating neurodegeneration arising from ethanol exposure and/or AD. This is complicated by known differences in neuronal metabolism at different stages of life (Dienel 2019), which have yet to be fully described. Investigation into the interactions between age, ethanol exposure, and metabolic outcomes may reveal how ethanol exposure throughout the lifespan impacts AD development and subsequently AD-related pathologies.

10.3.4. Pathological Tau Formation and Spread

In AD, although Aβ plaques are likely the first pathology to form, pTau levels have been shown to be more predictive of AD onset (Malpetti et al. 2020; Binette et al. 2022) and to correlate better with the progression of AD-associated neurodegeneration (brain atrophy) and cognitive decline (Thijssen et al. 2020, 2021; Teunissen et al. 2021). In fact, many consider AD to be an amyloid-induced tauopathy. Microtubules and their accessory proteins, referred to as microtubule-associated proteins (MAPs), exist throughout the entire organism and play key roles in mitotic and meiotic spindle formation, neuronal development, and polarization, among many other diverse regulatory cellular processes (Goodson and Jonasson 2018). Microtubule-associated protein tau (encoded by the gene Mapt) is primarily associated with stabilizing these microtubules and regulating axonal transport in the brain (Wang and Mandelkow 2016). This interaction between microtubules and tau is tightly regulated by post-translational modifications (PTMs), such as phosphorylation throughout development (Yu et al. 2009). However, in AD, as well as other neurodegenerative tauopathies, tau undergoes a variety of PTMs, like hyperphosphorylation (Šimić et al. 2016), throughout the preclinical, prodromal, and symptomatic phases of AD. Recent work has suggested that these PTMs can be heterogeneous, but that they undergo a characteristic pattern of pathogenesis (Wesseling et al. 2020). Importantly, tau can undergo a variety of other PTMs that include acetylation, ubiquitination, methylation, oxidation, and nitration (Alquezar et al. 2021; Carroll et al. 2021), among others. Tau is a highly modifiable protein with more than 80 potential phosphorylation sites (Noble et al. 2013). However, in AD, tau becomes abnormally hyperphosphorylated at a ratio 3 times greater than physiological tau (Wang et al. 2013), which then aggregate into neurofibrillary tangles (NFTs) (Dujardin et al. 2020; Wesseling et al. 2020). The accumulation of PTMs by tau is dependent on its length, which varies due to alternative splicing. For instance, the generated tau protein can differ in the number N-terminal inserts (0N, 1N, or 2N), and the microtubule binding repeats (MTBRs, 3R or 4R), with 0N3R tau only present during fetal development (Andreadis 2005, 2006; Liu and Gong 2008; McMillan et al. 2008). All six of these isoforms can be present throughout the lifespan, but 3R and 4R isoforms are typically present in a 1:1 ratio (Goedert et al. 1989). However, the ratio of 3R to 4R tau plays a crucial role in maintaining tau homeostasis (Ginsberg et al. 2006; Conrad et al. 2007). In AD this balance is disrupted resulting in a 2:1 ratio of 4R to 3R, and this imbalance is thought to facilitate PTM of tau (Alquezar et al. 2021). Pathological tau (pTau) is thought to sequentially develop and spread throughout the human brain, originating from within the entorhinal cortex (EC), hippocampus (HP), and limbic areas before spreading to neocortical areas (Braak and Braak 1991a, b; Mufson et al. 2016). This spread can occur in various ways, such as through extracellular synaptic vesicle release and subsequent endocytosis (Vogels et al. 2019, 2020; Robert et al. 2021). Regardless, relative neural activity is a key component of this process (Pooler et al. 2013; Holmes et al. 2014; Yamada et al. 2014; Wu et al. 2016). Thus, pTau production and spread, which are exacerbated by neural activity, are two pathological hallmarks of AD pathogenesis.

Even though levels of CSF tau, indicative of neurodegeneration, are transiently elevated in Wernicke’s encephalopathy (WE), this neuronal damage is different than AD (Matsushita et al. 2008). Emerging clinical reports using CSF-based biomarkers have found that older, cognitively intact participants in a frequent drinking group (≥ 1 time/week) had higher CSF p-Tau/Aβ42, and higher abnormalities in pTau and total-Tau levels, suggesting that AD biomarkers like pTau could be a strong indicator of future AD risk in this alcohol-consuming population (Wang et al. 2021). While other studies suggest that these cognitive deficits and abnormal CSF profiles from older demented alcohol-dependent patients could be masking an AD diagnosis (Azuar et al. 2021). Recent preclinical evidence agrees that excessive alcohol use can exacerbate intraneuronal Aβ and pTau in areas like the EC potentially due to lysosomal dysfunction (Tucker et al. 2022) and/or neuroinflammation (Barnett et al. 2022; Anton et al. 2024). Regardless, elucidating alcohol’s role in pathogenic tau development and spread as an AD risk factor will be crucial moving forward.

10.3.5. Excitatory Inhibitory Balance and Tauopathy, Linked Pathologic Processes

An observation common to both AUD (Correas et al. 2021) and AD (Lauterborn et al. 2021; Javed et al. 2022; Scaduto et al. 2023; Soula et al. 2023) is neural hyperexcitability, driven by disruptions in excitatory inhibitory balance (EIB). Glutamatergic and GABAergic neurons are the main excitatory and inhibitory neurotransmitters in the central nervous system (CNS), and they play a crucial role in regulating EIB (Sears and Hewett 2021). Proper inhibitory signaling is necessary for regulating the global spatiotemporal network activity and proper neural processing (Buzsáki et al. 2007). GABAergic neurons, which have significant diversity (Markram et al. 2004; Huang and Paul 2019), primarily modulate inhibitory activity in the brain by interacting with excitatory neurons (Fritschy and Brünig 2003; Sears and Hewett 2021). Dysfunction in GABAergic interneurons, particularly parvalbumin-expressing (PV) interneurons (Bartos et al. 2007), contribute to hyperexcitability in neuronal networks associated with AD (Verret et al. 2012; Cattaud et al. 2018; Petrache et al. 2019; Chung et al. 2020; Mattson 2020; Xu et al. 2020; Aouci et al. 2022).

Excessive alcohol use and the development of alcohol dependence can be characterized by excessive neural activity and somatic withdrawal (Heilig et al. 2010; Koob and Volkow 2016). This dysfunctional EIB is thought to be a result of alcohol’s direct facilitation of GABA_A receptors (Valenzuela and Jotty 2015; Olsen and Liang 2017) and inhibition of glutamatergic receptors (Möykkynen and Korpi 2012) acutely leading to hypoactivity (Dharavath et al. 2023). In a study of young adult binge drinkers, not only were cortical GABA levels reduced (Marinkovic et al. 2022), but there were also alterations in theta power and synchrony (Huang et al. 2022), suggesting that the GABA system is particularly vulnerable with continued high alcohol use. Alcohol-induced hyperexcitability, which (1) occurs as a compensatory adaptation to its acute CNS depressing actions (Valenzuela 1997; Correas et al. 2021), (2) is exacerbated by multiple withdrawal periods (Becker 1998), and (3) is observed in individuals with a history of excessive alcohol use (Gimenez-Gomez et al. 2023), could be a potential mechanism of alcohol-induced increases in pTau production and spread. Hence, GABAergic neurons, pivotal for maintaining inhibitory tone, may undergo selective targeting and modification during repetitive alcohol exposure and ensuing withdrawal phases, potentially rendering them more susceptible to pTau aggregation—this may represent a key pathogenic process linking alcohol misuse throughout the lifespan to an increased risk of AD.

Excessive neuronal activity is a well-known driver of pTau spread (Brunello et al. 2019). In mouse models of tauopathy, inhibitory cells are thought to be the first to undergo tauopathy aggregation, while excitatory synapses stayed relatively intact leading to increased hyperexcitability (Shimojo et al. 2020; Kudo et al. 2023). Moreover, when extracellular vesicles containing pTau were injected into the brains of C57BL6/J mice, pTau preferentially accumulated in GABAergic interneurons (Ruan et al. 2020), suggesting a potential vulnerability within inhibitory neurons that drives hyperexcitability in AD. There are also reports of reductions in glutamate clearance from the synaptic cleft which would contribute to further increases in hyperexcitability (Hunsberger et al. 2015). Emerging work building on these findings is focusing on using EIB as a biomarker for MCI (Cope et al. 2022; Javed et al. 2022). Elevated neural hyperexcitability in the hippocampus (HP) and frontal cortex (FC) of adults with a first-degree AD-diagnosed relative is consistent with these findings (Bassett et al. 2006). Interestingly, in patients with mild cognitive impairment (MCI), hyperactivity shifts into hypoactivity during the transition from MCI to AD (Dickerson et al. 2005; Celone et al. 2006) as a direct result of neurodegeneration. While the cell-type specificity of pTau vulnerability remains an active area of investigation, alterations in EIB may be among the first observable markers associated with AD pathology.

Pathogenic tau (pTau) in AD is observed early in AD in layer II of the entorhinal cortex (EC) (Braak and Braak 1991a, b; Kaufman et al. 2017, 2018) and then propagates throughout interconnected networks like the hippocampus (van Groen et al. 2003; Nilssen et al. 2019; Ohara et al. 2023) at rates thought to be dictated by relative neuronal activity (Pooler et al. 2013; Yamada et al. 2014; Wu et al. 2016). This relationship between neural activity and tau is likely what mediates the ability of CSF total-tau levels to predict seizure activity in AD cases (Tábuas-Pereira et al. 2019), as well as the recent observations of seizures and seizure-like activity in AD and rodent models of AD (Scharfman 2012). For example, recent studies in the 5xFAD mouse model of AD found that decreasing hyperexcitability was effective at reducing the development of molecular pathology (Barbour et al. 2024) and that this may be due directly to decreases in pTau propagation known to be exacerbated by seizures in the same mouse model (Barbour et al. 2023). In fact, there is now emerging evidence from multiple groups that seizure activity is a common feature of mouse models of AD (Palop et al. 2007; Roberson et al. 2011; Bezzina et al. 2015; Lisgaras and Scharfman 2022, 2023; Anna et al. 2023; Hole et al. 2024). Taken together these data support neuronal network hyperexcitability as a key underlying mechanism behind tau seeding and spreading in the context of AD and associated tauopathies.

Increases in neural EIB (i.e., hyperexcitability) have been observed in many animal models of alcohol use and are ascribed to compensatory alterations in neurotransmission resulting from repeated alcohol exposures (Mihic and Harris 1995; Tabakoff and Hoffman 1996; Valenzuela and Harris 1997; Valenzuela 1997). These alterations to EIB appear to occur throughout the brain (Pati et al. 2020, 2022; Downs et al. 2022), and as discussed above in the context of seizures, this relative increase in neural activity may be directly responsible for driving increased tau spread. We are actively exploring this possibility and to date have found some initial evidence that at least alcohol consumption in the P301S model of AD may indeed increase EIB (Maphis et al. 2024). There is a clear need for further preclinical research on the role of alcohol-induced alterations to EIB dysfunction and AD pathology, particularly focused on brain regions known to be involved in the development of both AUD and AD. In addition to exploring alcohol/AD relationships with preclinical models, advanced neuroimaging technologies and new sensitive blood-based biomarkers should be combined with accurate measures of alcohol use history in the clinic. Unfortunately, time is likely running out for millions of individuals—there is an urgent need to combine preclinical and clinical efforts toward uncovering the mechanisms responsible for alcohol-induced increases in AD risk.

10.4. Alcohol and Alzheimer’s Disease: Future Directions

There is now mounting evidence from the past two decades which strongly supports alcohol use as a major risk factor for AD. Studies examining the effects of alcohol misuse in adolescence (Pascale et al. 2022), young adulthood (de Goede et al. 2021; Kekkonen et al. 2021), midlife (Anttila et al. 2004; Sabia et al. 2018; Kivimäki et al. 2020; Chosy et al. 2022; Zhao et al. 2024), and advanced age (Heymann et al. 2016; Kivimäki et al. 2020) all indicate that alcohol exacerbates risk for AD. However, to date there are many gaps in our knowledge that need to be addressed. First, epidemiological studies often fail to track patterns of alcohol use throughout the lifespan. Second, these studies typically only assess AD-related outcomes at one or two points in time. Third, many large studies on alcohol use critical to our understanding of AD progression have not assessed alcohol use past the age of 65, when LOAD is most commonly diagnosed. Fourth, while there are active human longitudinal studies on the impact of alcohol use during adolescence and young adulthood when drinking rates are highest and the brain is still developing (Lisdahl et al. 2018), we are unaware of any that measure AD-related outcomes. All the above issues would be addressed with comprehensive lifespan studies that use neuroimaging, blood-based biomarkers, and alcohol use assessments throughout the preclinical and prodromal phases of AD. We argue that additional characterization of oxidative stress, neuroinflammation, metabolic dysregulation, pTau pathogenesis, and EIB is particularly beneficial to include.

There is a lack of preclinical studies examining the impact of alcohol on AD so more research is needed to gain even fundamental knowledge on how alcohol affects the development and progression of AD. Many alcohol/AD studies utilize the 3xTg animal model (Castano-Prat et al. 2019; Hoffman et al. 2019; Frausto et al. 2022; Tucker et al. 2022; Walter et al. 2022; Sanna et al. 2023), characterized by mutations in both APP and tau, while others have used tauopathy models featuring tau mutations exclusively (P301S, PS19) (Catavero et al. 2022; Downs et al. 2022; Maphis et al. 2024). While these mutations are clearly useful in driving pathogenesis, they do so at very young ages, which is inconsistent with the majority of AD cases (i.e., LOAD) (Drummond and Wisniewski 2017; Tai et al. 2020). These studies in transgenic mice are currently and will continue to be informative to this emerging field, but we should be prepared to quickly contextualize them using alternative approaches. For example, conducting a study to assess alcohol’s influence on survival rates in a range of AD models could provide invaluable knowledge on which specific pathogenic pathways in AD are most perturbed in the context of alcohol. These data would also aid in prioritizing mouse models and molecular systems that better recapitulate LOAD. Moreover, there is a need for improved mouse models of LOAD. Fortunately, the national consortium, MODEL-AD (Wilcock and Lamb 2024), is actively working on creating and validating novel models for LOAD using humanized apolipoprotein epsilon 4 varient (APOEε4, strongest genetic predictor of LOAD, Sienski et al. 2021), as well as incorporation of non-mutant humanized Aβ and tau. These models will be particularly instrumental in evaluating the effects of alcohol exposure on AD at a variety of developmental stages and can do so in a timeframe that is greatly accelerated compared to human studies.

In summation, both epidemiological and pre-clinical studies implicate alcohol use as a risk factor for AD and AD-related dementia. These studies identify key molecular pathways that include oxidative stress, neuroimmune dysfunction, metabolic dysregulation, pTau formation/spread, and excitatory/inhibitory balance. We have outlined some potential future work that we believe will be essential to fully define the nature of alcohol-related risk in AD, which we are hopeful will identify key pathological mechanisms setting the stage for the development of therapeutic targets.