Sex-specific effects of chronic alcohol consumption across the lifespan in the transgenic Alzheimer’s Disease (TgF344-AD) rat model

Abstract

Alcohol consumption across the lifespan contributes to mood fluctuations and cognitive dysfunction, two neurobehavioral features also associated with Alzheimer’s Disease and Related Dementias (ADRD). Yet, few studies have used rodent models to determine how a history of ethanol consumption across the lifespan might contribute to neurobehavioral and neuropathological features of ADRD. We exposed Wild Type (WT) and transgenic Fischer 344 AD rats (TgF344-AD) that have been genetically modified to express the human Amyloid Precursor Protein (APP) and presenilin-1 genes with mutations, to ethanol using a chronic, intermittent ethanol consumption model. Beginning at P28, rats were given a single bottle 10% ethanol solution for 2 consecutive days, followed by 2 days of tap water. This pattern (2 days on, days off) was repeated for a total of 12 cycles until rats reached the age of ~3 months, and repeated at 6 (Exp 1 and Exp 2) and 9 months of age (Exp 2). In experiment 1, ethanol consumption decreased alternations in a spontaneous alternation task in females, only at the 3-month time point, whereas TgF344-AD females showed increased contextual fear conditioning in the test of retention and reinstatement tests at 6 months of age. In experiment 2, a battery of anxiety-like behaviors (Elevated Plus Maze, Marble Burying, and Novelty Induced Hypophagia) were assessed following a 2-week abstinence period at 3, 6, and 9 months of age in ethanol-consuming rats. Data from the EPM and marble burying tasks revealed evidence of heightened anxiety-like behavior in Tg-F344-AD rats that varied by sex and age, with no significant effects of ethanol. In the novelty-induced hypophagia task, males with a history of ethanol consumption had a lower latency to approach a familiar, salient reward at 3 months old, but effects of ethanol were overall minimal. Examination of dorsal hippocampal gene expression at 6 months of age under basal conditions also revealed predominantly genotype and sex-specific effects on inflammation- and AD-related genes (App, Il-6, Bace1, Rage, Lrp-1). When examined at 9 months old following LPS challenge, ethanol increased inflammatory genes in males (Il-1β, Il-6) in the hippocampus, whereas ethanol decreased several inflammatory and AD-related genes (Hmgb1, Rage, Bace1, Lrp-1) in TgF344-AD females. Overall, these data provide further evidence that females are especially vulnerable to AD, and that a history of ethanol consumption had selective, rather than global, effects on AD- and inflammation-related genes following an inflammatory stimulus.

Introduction:

Alcohol use in adolescence is strongly associated with developing an Alcohol Use Disorder (AUD) in adulthood (Pitkanen, Lyyra, & Pukkinen, 2015). AUD diagnoses are also on the rise among aging populations (>50 years of age) (Han et al., 2017). Similarly, hospital admissions for heavy alcohol use significantly increased in adult and aging populations (45+ years of age) from 1993 to 2010 (Sacco et al., 2015). These data indicate that alcohol misuse and AUDs are on the rise among adult and elderly populations, effects that may be particularly problematic when alcohol misuse begins in early adolescence (Swendensen et al., 2012; Han, Moore, Sherman, Keyes, & Palamer, 2017; Keyes et al., 2019). This is especially concerning because heavy and chronic alcohol consumption across the lifespan may pose a significant risk factor for developing Alzheimer’s Disease and Related Dementias (ADRD; Harwood et al., 2010; Deng et al., 2006; Xu et al., 2009). Consistent with this, chronic alcohol misuse was recently identified as one of the top modifiable risk factors in the development of dementia world-wide (Livingston, 2023).

The amyloid beta hypothesis posits that Aβ deposits form neurotoxic, insoluble plaques that contribute to the neuropathological outcomes associated with AD (Hardy & Higgins, 1992; Chen et al., 2017; McGirr et al., 2020). Familial AD (FAD), a genetic subset of AD, has been linked to mutations and overexpression of presenilin 1, presenilin 2, and amyloid precursor protein (App), which ultimately results in accumulation of Aβ plaques (Kang et al., 1987; Borchelt et al., 1996). Many transgenic models of AD exist in both mice and rats, most of which express one or more FAD-related genes, including the 3x mouse model (Billings et al., 2005; Stover et al., 2015), the 5x mouse model (Richard et al., 2015; Giannoni et al., 2016; Jawhar et al., 2012), and the TgF344-AD rat model (Cohen et al., 2013; Pentkowski et al., 2018). Aβ plaque accumulation, cognitive impairments, memory deficits, and behavioral disruptions associated with AD in patients have been recapitulated these animal models (Cohen et al., 2013, Pentkowski et al., 2018).

Cognitive impairments and memory disruptions are hallmark symptoms of AD in human patients and are currently the principal diagnostic criteria for determining stage and severity of the disease (Reviewed by Jiménez-Balado & Eich, 2021). In a mouse model of AD, deficits in performance in a spontaneous alternation task, a measure of cognition and spatial working memory, were observed in adult transgenic-AD, but not WT (wild type), mice (Miedel et al., 2017). Similarly, cognitive degradation was observed in a battery of cognitive tasks in a mouse model of AD across time, where no impairments were observed in early adulthood, then impairments emerged in middle adulthood, and increased in severity during later aging (Filali & Lalonde, 2009). These cognitive impairments are associated with age-dependent Aβ plaques and neurofibrillary tangles that appeared throughout the cortex, hippocampus, and cerebellum of transgenic AD models (Cohen et al., 2013; Marutle et al., 2002). TgF344-AD rats have shown deficits in reversal learning in the Morris water maze (Rorabaugh et al., 2017), the Barnes maze, and the novel object recognition test, indicating deficits in cognitive flexibility and spatial memory, consistent with clinical populations (Cohen et al., 2013). Taken together, these data indicate that animal models of FAD recapitulate hallmark symptoms of cognitive deficits and deterioration associated with AD.

Clinically significant anxiety occurs in approximately 40% of patients with AD, and anxiety diagnoses often precede cognitive impairments in those later diagnosed with AD (Mendez, 2021). Animal models of FAD also recapitulate the affective and behavioral changes associated with the development of AD across time. TgF344-AD rats were found to have increased anxiety-like behavior in the EPM (Elevated Plus Maze), as measured by reduced percent time spent in open arm, as early as 4–6 months of age, before significant deficits in spatial memory occurs (Pentkowski et al., 2018). Male and female TgF344-AD rats (9–11 MO) showed increased anxiety-like behavior in the EPM and decreased motivation in a sucrose consumption test, suggesting affective dysregulation, while no differences were observed in spatial working memory using the Y-maze (Tournier et al., 2020). At 12MO, TgF344-AD rats displayed increased anxiety-like behavior in an open field test, increased anxiety-like behavior in an EPM, and increased depressive like behaviors in the sucrose preference test and increased time spent immobile in a forced swim task (Wu et al., 2020). Similarly, TgF344-AD rats (6–8MO, males) had increased anxiety-like behavior in the EPM, as well as contextually-conditioned fear compared with WT rats (Pentkowski et al., 2022). These studies provide evidence that transgenic animal models replicate prodromal anxiety, as well as neuropathology of AD, such as cognitive impairments and hippocampal dysfunction, supporting the utility of transgenic models in determining risk factors and/or protective factors for developing AD/ADRD.

The relationship between chronic ethanol intake and AD neuropathology is somewhat mixed, with some evidence suggesting that mild alcohol use may decrease risk of developing AD, whereas heavy alcohol use may increase risk of developing AD/ADRD (Harwood et al., 2010; Deng, Zhou, Li, Wang, Gao, & Chen, 2006). In support of this, chronic ethanol resulted in increased expression of AD-associated genes such as App and Bace1 in adult male Sprague Dawley rats (Kim et al., 2011). Similarly, chronic ethanol increased both Aβ_1–40 and Aβ_1–42 in mice, suggesting that either ethanol itself increases Aβ peptide production, or otherwise interferes with clearance of Aβ (Gong et al., 2017). Interestingly, a study performed in our lab using wild type F344 rats found evidence of increased microglial phagocytosis of Aβ following a moderate chronic intermittent ethanol consumption procedure in females only (Marsland et al., 2023). Taken together, these data indicate that ethanol influences the ability to either produce or otherwise process Aβ fragments, though the timing, concentration, and number of exposures to ethanol likely influences vulnerability to the aggregation of Aβ fragments into larger plaques.

The basal state of the brain under AD neuropathology may differ quite dramatically from the state of the AD brain following inflammatory challenge, such as lipopolysaccharide (LPS). Peripherally generated Aβ peptides deposit into the brain and lead to AD pathogenesis (Eisele et al., 2014). Disruptions to Aβ peptide transporters across the BBB, such as ABCB1, LRP-1, and RAGE, result in changes to Aβ homeostasis in the CNS. LRP-1 and P-glycoproteins (PGP-1, also known as ABCB1) are efflux transporters that shuttle Aβ from the brain to the blood, while RAGE (also known as AGER) is the major transporter of Aβ from the blood to the brain (Deane et al., 2003; Donahue et al., 2006). LRP-1 disruptions, which occur following chronic ethanol consumption, are associated with impairment of Aβ efflux from the brain and dysregulation of microvasculature within the CNS (Garcia et al., 2022; Wei et al., 2021; Shibata et al., 2000; Donahue et al., 2006). LPS injection caused decreased LRP-1 receptor protein, decreased central and peripheral clearance of exogenous Aβ, and shifted Aβ deposits from parenchyma to the vasculature in an LRP-1 dependent manner (Wei et al., 2021; Erickson et al., 2013; Shibata et al., 2000). Further studies have found that LPS induced endotoxemia resulted in increased Aβ plaques in the brain, as well as increased inflammatory cytokine expression (Wang et al., 2018). Taken in conjunction with a history of chronic ethanol, it is likely that inflammatory challenge in AD patients will further exacerbate neuropathology of AD.

The goal of these studies was to examine the influence of chronic, lifelong ethanol consumption on the development of AD pathology and behavior in both genetically vulnerable (TgF344-AD) and wild type rats, using both cognitive (spontaneous alternation, contextual fear conditioning) and anxiety tasks (elevated plus maze, marble burying, novelty induced hypophagia) at various ages to assess the development of AD-related pathology. A variety of time points were chosen, both behaviorally as well as endpoints for PCR, to detect prodromal neurobehavioral changes. We examined inflammatory and AD-related gene expression to evaluate the impact of chronic ethanol on the physiological development of AD, both under basal conditions (Experiment 1) and following lipopolysaccharide challenge (Experiment 2). We hypothesized that anxiety-like behaviors would be present in TgF344-AD rats prior to cognitive impairments, and that both anxiety behaviors and cognitive impairments would worsen over time. Altogether, we hypothesized that chronic ethanol exposure would result in cognitive and affect dysregulation into aging, and that these behavioral changes would be caused by changes in AD and neuroinflammatory disruption in the CNS.

Methods

General Methods:

Animal Husbandry:

TgF344-AD rats expressing human APPsw and PSEN1 with delta exon 9 mutation were acquired from the Rat Research and Resource Core (RRRC) at the University of Missouri and bred on site for AD+ and AD- rats (Cohen et al., 2013). This transgenic rat model has been endorsed by the NIA for NIH-supported investigations into the etiology of familial AD, and wild type (AD-, non-transgenic) Fischer 344 rats were also used in all experiments. At P21, rats were weaned and pair-housed with same-sex partners to be genotyped. Following genotyping (Transnetyx, Cordova, TN, USA), rats were re-housed and paired with a same-sex, same-genotype, non-littermate partner for experimentation prior to P28.

All rats (n = 8–14/group) were pair-housed in a standard Plexiglas cage with chew blocks provided for enrichment, cage changes occurred every 4 days. Rats received ad libitum access to food, and colony conditions were maintained on a 12:12 light/dark cycle (lights on 07:00) at 22 +/− 1° C. They were handled for 2–3 minutes for 2 days prior to start of ethanol exposure and for 7 days prior to behavioral manipulations at each time point to acclimate to experimenter handling. Rats housed together were assigned the same experimental condition. All experimental subjects were maintained in accordance with PHS policy and the Institutional Animal Care and Use Committee (IACUC) at Binghamton University and experiments were approved prior to procedures.

Chronic intermittent ethanol exposure procedure:

To determine whether lifelong history of ethanol would increase prevalence of Alzheimer’s-related deficits, adolescent (P28-P32) TgF344-AD and WT rats were randomly assigned to a continuous access, intermittent ethanol exposure procedure as previously described, or tap water as a control group (Marsland et al., 2023A; Marsland et al., 2023B). Pair-housed rats were given access to a single bottle of ethanol (10% v/v in tap water) for a period of 48 hours, followed by a period of tap water access for 48 hours. Drinking data was recorded daily, measured as pre- and post- bottle weights to calculate estimated consumption. Ethanol consumption was estimated from grams consumed of 10% ethanol using the following formula: $\text{(mL}\text{ethanol}\text{consumed}\text{*}\text{0}{\text{.1}}_{\text{[10\%}\text{ethanol]}}\text{*}\text{0}{\text{.79}}_{\text{[specific}\text{gravity}\text{of}\text{ethanol]}}\text{)/(body}\text{weight}\text{of}\text{both}\text{partners}\text{in}\text{kg)}$. Importantly, in order to not cause long-term isolation stress, rats were pair-housed throughout the entire experiment; thus, individual ethanol consumption could not be directly measured. Instead, ethanol consumption was estimated within the cage and drinking data is reported by cage rather than by individual animal.

Each presentation of ethanol followed by water is termed a cycle, and rats were placed on this procedure for a total of 12 cycles (48 days) for three developmental epochs (1–3MO, 3–6MO, and 6–9MO). Neurobehavioral testing occurred during a 2-week period of abstinence between each cluster of 12 cycles (see Figure 1). Data from our lab using this model indicates that the procedure is well tolerated and produces consistent, moderate-to-binge levels of ethanol consumption based on both blood and brain ethanol concentrations (Marsland et al., 2023A; Marsland et al., 2023B). Estimated consumption in this experiment averaged out to be ~15 g/kg/day in the early life period (1–3MO), and ~10 g/kg/day in adults (3–6 and 6–9MO), similar to other experiments performed by our group using this drinking model (Marsland et al., 2023a; Marsland et al., 2023b; Sicher et al., 2024).

Figure 1. Experimental timelines.

(A) Timeline for experiment 1. (B) Timeline for experiment 2.

LPS Administration:

LPS from Sigma (serotype E0111:B4) was diluted to 1.0 mg/mL in sterile, pyrogen-free physiological saline (0.9%) and stored at −20° C. Three (3) hours prior to tissue collection in Experiment 2, LPS was prepared to a concentration of 100 μg/mL and delivered at a dose of 1 mL/kg in saline (i.p.). Home cage control animals were given an equivalent i.p. injection of physiological saline.

I. Cognitive Battery:

A. Spontaneous Alternation

In Experiment 1, following the cessation of the 2on/2off procedure, animals underwent a 2-week abstinence period from ethanol to fully abate withdrawal (Figure 1A). Rats were tested at both 3 and 6 months of age on a spontaneous alternation procedure in a 4-arm plus maze with black flooring and transparent exterior walls on all arms (measuring 105.5 cm × 14.4 cm × 15 cm) (Fernandez & Savage, 2017; Reitz et al., 2024). 2 days prior to spontaneous alternation, animals were food restricted to 85–90% of their free-feeding body weight to motivate animals to explore the spontaneous alternation maze (Fernandez et al., 2017; Fernandez, Steward, & Savage, 2016). On days of testing, animals were taken from home cage and given 30 minutes to habituate to the testing room under low light conditions. After habituation, animals were placed in center of maze in a brightly lit room, facing the same direction, and scored for 18 minutes on exploratory behavior in each arm of the maze. Scoring occurred live while the animal was on the maze by a double-blinded observer. Entries into each arm of the maze were recorded, and an entry was counted when the animal had all 4 paws in an arm of the maze. An alternation was scored when an animal made entries into each arm of the maze continuously, with no repeats into an arm that was already entered. Animals that made fewer entries than the threshold (12 arm entries) were not included in data analysis.

B. Contextual Fear Conditioning

In Experiment 1, rats underwent a contextual fear conditioning to evaluate hippocampal function. Animals were conditioned and tested in fear conditioning chambers (32 × 25 × 25 cm, Med Associates) made of clear polycarbonate (top, front walls), white acrylic (back wall), and stainless steel (sides, shock grids, drop pan) material. The grid floors consisted of 19 parallel 4.8-mm diameter rods situated 1 cm apart. At 3 months of age, following spontaneous alternation testing, each animal was placed in a fear chamber for a 5-minute acclimation period (pre-shock) and then exposed to 3 presentations of scrambled footshock (3 shocks, 1mA, 1s each, ITT 90s) for a period of 3 minutes. Following the shock period, there was a 5-minute consolidation period (post-shock), before being returned to home cages. Animals were then tested for retention (ToR) of fear memory 24 hours following the conditioning procedure, where they received no further shocks. At 6 months of age, animals were reintroduced to the context for a period of 5 minutes, then received a single shock (1.0 mA, 1 s each), and 1 minute to consolidate the reinstatement shock. Animals were then placed back in the chambers, unshocked, and tested for the retention of the reinstatement 24 hours later.

II. Anxiety Battery:

In Experiment 2, a non-shocked group of rats at equivalent timepoints (3, 6, and 9MO), underwent a battery of anxiety tests, including elevated plus maze (EPM), marble burying, and novelty-induced hypophagia tests (Figure 1B).

A. Elevated plus maze

On the first day of EPM testing, rats were habituated to an external procedure room in low light conditions, then placed on an EPM (48.3 × 12.7 cm open arms with 29.2 cm walls on the closed arms) and recorded on video camera for a total of 5 minutes. EPM anxiety scores were recorded by a blinded scorer from the video recording for: time spent in the closed and open arms, number of stretch-attend postures, number of head dips, and number of entrances made into the closed arms (Pentkowski et al., 2018).

B. Marble burying task

Twenty-four hours after the completion of EPM, rats were assessed on the marble burying task. Rats were placed into a novel opaque-sided cage with equivalent dimensions to the home cage, using standard wood shaving bedding, with an array of 15 marbles (1 cm diameter) placed in a 3×5 grid for 15 minutes, recorded on video camera. Marble burying scores were recorded by a blind observer immediately after animals were removed from the cage and reported as number of marbles buried (greater than 2/3 of surface area covered, estimated by observer) (Rouzer et al., 2017).

C. Novelty-induced hypophagia

Twenty-four hours after the completion of the marble burying task, rats were placed into a new, single-housed home cage for novelty induced hypophagia training. Rats had access to one Nilla Wafer (Nabisco, NJ, USA) placed in the training cage for an hour per day, for 6 days. Following the 1-hour exposure, rats were placed back with partners in their home cage. Following 6 days of training, the novelty induced hypophagia test took place. Rats were placed in a brightly lit, large plastic tub (58.42 cm x 43.18 cm x 31.75 cm) with no bedding and a Nilla Wafer for 15 minutes, and behaviors were recorded on video camera. A measurement of the weight of Nilla wafers was taken prior to and following each day of novelty induced hypophagia training to determine amount of Nilla wafer consumed during the training period. On the test day, anxiety scores were measured to include: latency to approach the Nilla wafer, latency to consume the Nilla wafer, and percent of Nilla wafer consumed during the test. All videos were scored remotely by a blinded observer (Rouzer et al., 2017).

Testing throughout lifespan:

1–2 weeks following behavioral testing, exposure to the continuous access, intermittent ethanol exposure procedure resumed to model a lifelong history of intermittent ethanol intake (Figure 1). Behavioral testing occurred at 3 and 6 months of age in all experiments. Experiment 2 continued behavioral testing until 9 months. In both Experiments 1 and 2, rats were euthanized via rapid decapitation and brains, spleens, and livers were collected for RT-PCR analysis. These time points were chosen to identify prodromal and developing symptoms of AD across the lifespan.

Tissue collection and RT-PCR:

One week following the final behavioral test in both Experiments 1 and 2, rats were rapidly decapitated (non-anesthetized), and tissue collected for PCR. Plasma was stored at −20°C until protein assay analysis. Brains were flash frozen in 2-methylbutane on dry ice, then stored at −80°C until punched on a Leica cryostat at −16° C. Spleens and Livers were gross dissected and frozen on dry ice, then stored at −80°C. Brain punches were taken using Paxinos and Watson’s 2nd Edition Brain Atlas from the dorsal hippocampus (dHPC, Bregma −4.36 mm, 2mm x 1.5mm punches, bilateral). Tissue punches were stored in 2 mL microcentrifuge tubes at −80° C for PCR analysis.

Reverse-Transcriptase Polymerase Chain Reaction

RT-PCR was conducted using procedures described in previous work (Doremus-Fitzwater, Gano, Paniccia, & Deak, 2015). Reagent (Invitrogen, Grand Island, NY) was spiked into tissue punches along with a 5 mm stainless steel bead, then homogenized using a Qiagen Tissue Lyser (Qiagen, Valencia, CA). RNA was extracted using RNeasy mini columns (Qiagen) and eluted in 65°C RNase-free water. Nanodrop (ThermoScientific) was used to determine RNA concentration and quality, and then normalized using RNase-free water. QuantiTect reverse transcription kit (Qiagen) was followed according to manufacturer’s instructions to synthesize cDNA. cDNA was amplified in a 10 μL reaction, consisting of 0.5 μL cDNA, 5 μL SYBR Green Supermix (Bio-Rad), 0.5 μL primer, and 4 μL Rnase-free water. RT-PCR was conducted using a CFX384 real-time PCR detection system (Bio-Rad) in a 384-well plate. Samples were pipetted in triplicate, and underwent a 3-minute start, then were denatured at 95° C for 30 seconds, annealed (30 seconds at 60° C) and extended (30 seconds at 72° C) for a total of 40 cycles. To ensure product alignment, samples were again denatured for 1 minute at 95° C and annealed for 1 minute at 55° C. Following annealing, samples increased temperature at a rate of 0.5° C every 15 seconds until 95° C to analyze the melt curve for specificity to target gene. Data were analyzed relative to expression of the ultimate control using the 2^ΔΔC(t) method (Livak & Schmittgen, 2001) using GAPDH as a housekeeper. However, in both experiments, GAPDH showed significant fluctuations across experimental groups as noted below. To address this, we conducted PCR reactions on several other common housekeepers and were unable to identify a stable housekeeper gene. For this reason, we analyzed and report GAPDH as a separate target gene for the sake of transparency. It should be noted, however, that the statistically significant effect on GAPDH was mathematically small and affected only a single experimental group. Therefore, all gene targets continue to be expressed relative to the housekeeper.

Statistical Analyses.

Body weight and consumption data included time as a repeated measures variable, with each phase of ethanol exposure being analyzed separately using Graphpad Prism and Statistica software. Behavioral data were analyzed with a separate 2×2 ANOVA for each sex with genotype and ethanol exposure condition as factors, as data from previous experiments in our lab led to the a priori prediction that more ethanol would be consumed in females compared with males, thus making sex-specific comparisons of the dependent measures difficult (Marsland et al., 2023B). Gene expression analyses included a separate group of true home cage controls that were not behaviorally tested, thus target genes were analyzed with a single factor ANOVA followed by Tukey’s HSD post hocs to clarify differences between groups. The alpha was set to > 0.05. Due to the number of comparisons made, results discuss only significant main effects and interactions.

Specific Methods:

Experiment 1: Cognitive Behaviors

Starting at P28-P32, male and female, wild type (WT) and transgenic (TgF344-AD) rats (n = 10, N = 80) were given a single bottle of 10% ethanol or water for 2 days using a chronic intermittent ethanol exposure procedure then given water access for 2 days, described above, for a total of 12 cycles. Following the end of the drinking cycle, rats underwent an abstinence period of 2-weeks before behavioral testing. At 3 months old, rats were tested for spontaneous alternation and contextual fear conditioning. 1–2 weeks post behavioral testing, animals were placed back on the chronic intermittent ethanol exposure procedure for another 12 cycles. At approximately 6 months of age, animals again underwent a 2-week period of abstinence before being tested for spontaneous alternation, then placed back in contextual fear conditioning chambers for retention of fear as well as a brief (single shock) fear reinstatement and tested for retention of reinstatement 24 hours later. 2 weeks post behavioral testing, animals were rapidly decapitated (unanesthetized), and tissue was taken for PCR. Thus, rats were approximately 7 months of age at the time of tissue collection under basal (unchallenged) conditions.

Experiment 2: Anxiety-like battery

Starting at P28-P32, male and female, wild type and transgenic (TgF344-AD) rats (n = 8–14, N = 126) were given a single bottle of 10% ethanol or water for 2 days using a chronic intermittent ethanol exposure procedure then given water access for 2 days, described above, for a total of 12 cycles. A select group of male and female wild type F344 rats were kept in home cages and remained undisturbed for the duration of the experiment and used as the home cage control (HCC) group for gene expression analysis at the end of the experiment. Following the end of the drinking cycle, rats underwent an abstinence period of 2-weeks before anxiety testing, as described above. 2 weeks post behavioral testing, animals were given a single injection of sterile physiological saline (HCC) or LPS (100 μg/kg). Three (3) hours later, rats were rapidly decapitated (unanesthetized), and tissue was taken for PCR. Thus, rats in Experiment 2 were approximately 10 months old at the time of tissue collection, and brains were collected under conditions of an evoked immune challenge (LPS).

Results

Spontaneous Alternation.

No effects were observed on spontaneous alternation in males at 3 MO or at 6 MO (p > 0.05) (Table 1). In females, a main effect of ethanol history was observed, such that ethanol reduced percent alternations made at 3 MO (F (1, 37) = 5.244, p < 0.05) (Table 1). No effects on spontaneous alternation in females at 6 MO (Table 1). However, it is worth noting that a significant proportion of animals did not meet criteria to be included in analysis of spontaneous alternation at the 6-month mark (31%), making analysis of spontaneous alternation at later timepoints difficult to interpret (Table 1).

Table 1.

Percent Alternations observed in the Spontaneous Alternation Task.

	Percent Alternations
	WT water	WT ethanol	AD water	AD ethanol	inx
3 MO	Mean ± SEM	Mean ± SEM	Mean ± SEM	Mean ± SEM	p value
Males	30.4 ± 3.6	26.0 ± 4.1	29.5 ± 3.0	27.0 ± 3.1	0.78
Females	34.6 ± 3.9	27.5 ± 3.8	30.7 ± 2.6	23.2 ± 2.4	0.99
6 MO
Males	35.0 ± 15.0	20.3 ± 3.2	24.5 ± 7.2	30.8 ± 6.3	0.15
Females	31.9 ± 5.3	40.2 ± 4.6	31.5 ± 8.4	29.9 ± 3.1	0.35

Values represent mean +/− SEM. Significant differences as reported by a 2-way ANOVA, and p-value for the interaction term reported in the right-hand column.

Contextual Fear Conditioning.

When Contextual Fear Conditioning data were analyzed, no differences observed in percent freezing following initial shock at 3 MO in either males (p > 0.05; Table 2) nor females (p > 0.05). Males also did not show any differences in percent freezing during the test of retention (p > 0.05; Table 2). However, a main effect of genotype was observed in the Test of Retention in females, wherein TgF344-AD females froze significantly more than wild type counterparts (F (1, 38) = 12.13, p < 0.01; Table 2). No differences were observed in percent freezing in the 5-minute acclimation period prior to the reinstatement shock at the 6-month time point in males (p > 0.05; Table 2); however, a main effect of genotype was observed, such that TgF344-AD females showed significantly more freezing during the reinstatement test compared with wild type females (F (1, 38) = 5.567, p < 0.05; Table 2). Similarly, during the test of retention to the reinstatement shock, males showed no differences in fear responses (p > 0.05; Table 2), while a main effect of genotype was observed in females, such that TgF344-AD females showed significantly higher percent freezing compared with WT females (F (1, 38) = 11.18, p < 0.01). During extinction main effects of genotype were observed in both transgenic males (F (1, 34) = 4.295, p < .05; Table 2) as well as in transgenic females (F (1, 38) = 8.004, p < 0.01) compared with wild type animals.

Table 2.

Percent Freezing observed in contextual fear conditioning.

	Contextual Fear Conditioning
3 MO	WT water	WT ethanol	AD water	AD ethanol	inx
Shock	Mean ± SEM	Mean ± SEM	Mean ± SEM	Mean ± SEM	p value
Males	39.2 ± 3.8	39.5 ± 3.4	37.2 ± 3.3	33.4 ± 3.2	0.52
Females	27.1 ± 2.7	37.2 ± 3.3	34.6 ± 3.2	33.4 ± 3.3	0.06
Test of retention
Males	69.9 ± 2.8	74.0 ± 2.5	71.8 ± 2.6	60.6 ± 2.8	0.23
Females	41.8 ± 2.8	43.7 ± 2.7	66.7 ± 2.4	59.9 ± 3.0	0.46
6 MO
Reinstatement
Males	62.3 ± 4.8	65.3 ± 3.6	73.8 ± 3.6	61.9 ± 4.3	0.27
Females	50.8 ± 5.4	44.0 ± 4.7	76.0 ± 3.5	60.5 ± 4.4	0.58
Retention of reinstatement
Males	83.4 ± 2.0	87.2 ± 1.7	79.5 ± 2.4	82.3 ± 2.2	0.88
Females	70.5 ± 2.5	70.3 ± 2.5	85.7 ± 1.7	83.0 ± 2.0	0.77
Extinction
Males	56.2 ± 2.2	58.9 ± 2.0	64.0 ± 1.9	66.0 ± 1.9	0.92
Females	48.1 ± 1.9	47.0 ± 1.9	67.1 ± 2.0	54.9 ± 2.1	0.25

Values represent mean +/− SEM. Significant differences as reported by a 2-way ANOVA, and p-value for the interaction term reported in the right-hand column. Main effect of genotype reported as bolded font.

Male gene expression data in dorsal hippocampus.

Gene expression data was adjusted to housekeeper GAPDH prior to analysis. As expected, a main effect of genotype was observed in App expression, such that TgF344-AD males had higher expression of App (F (1, 34) = 23.74, p < 0.01; Figure 2A). A main effect of genotype was also observed in expression of Il-6, such that TgF344-AD males had decreased expression of Il-6 compared with wild type males F (1, 34) = 8.288, p < 0.01; Figure 2B). No differences were observed in expression of Bace (p < 0.05, Figure 2C), Rage (p < 0.05, Figure 2C), and Lrp-1 (p > 0.05; Figure 2D) in males.

Figure 2. dHPC gene expression at 6MO.

(A) APP gene expression increased in TgF344-AD males and females (&). (B) IL-6 expression decreased in TgF344-AD males; IL-6 expression increased in TgF344-AD females. (C) BACE1 expression increased in TgF344-AD females. (D) RAGE expression increased in TgF344-AD females. (E) LRP-1 expression increased in TgF344-AD females. All data expressed as mean ± SEM.

Female gene expression data in dorsal hippocampus.

Gene expression data was adjusted to housekeeper GAPDH prior to analysis. As expected, a main effect of genotype was observed in App expression, such that TgF344-AD females (F (1, 38) = 27.58, p < 0.01) had higher expression of App (Figure 2A). A main effect of genotype was observed in Il-6, Bace, Rage, and Lrp-1 expression in females, wherein TgF344-AD females showed an increase in Il-6 (F (1, 35) = 5.471, p < 0.05; Figure 2B), Bace (F (1, 37) = 9.266, p > 0.01), Rage (F (1, 38) = 6.49, p > 0.05; Figure 2C), and Lrp-1 expression (F (1, 38) = 4.844, p < 0.05, Figure 2D) compared with wild type females. No differences were observed in any of the genes as function of ethanol consumption history.

Supplemental Data: Body weight gain and ethanol consumption.

In experiment 1, from 1 to 3 months of age, male body weights increased across time, as expected (F (3.208, 106.7) = 2458, p < 0.01; Supplemental Figure 1A). There was also a main effect of ethanol (F (1, 34) = 5.080, p < 0.05), such that animals with a history of ethanol gained less weight over time when compared with water-consuming animals (Supplemental Figure 1A). Similarly, from 3–6 months of age, male animal weights increased across time (F (3.434, 116.7) = 534.4, p < 0.01), ethanol-consuming animals gained less weight that water-consuming animals (F (1, 34) = 6.512, p < 0.05), and wild type animals weighed less compared with transgenic rats (F (1, 34) = 6.063, p < 0.05) (Supplemental Figure 1A). As expected, there was a main effect of time on consumption in the males, such that rats consumed less g/kg across time (F (3.340, 26.11) = 8.635, p < 0.01) (Supplemental Figure 1B). No effects were observed on consumption intake at the 3–6 month consumption period (Supplemental Figure 1B). In females, body weights during the 1–3 month consumption period increased across time, as expected (F (4.140, 152.8) = 3241, p < 0.01) (Supplemental Figure 1C). Similarly, from 3–6 months, there was an increase in body weight across time (F (4.569, 173.6) = 173.9, p < 0.01) as well as a main effect of genotype (F (1, 38) = 6.442, p < 0.05), such that wild type females weighed less than transgenic females (Supplemental Figure 1C). Similar to males, there was a decrease in consumption across time in females (F (4.601, 35.97) = 4.123, p < 0.01) (Supplemental Figure 1D). However, from 3–6 months old, there were no differences in consumption across time, but there was increased consumption in the wild type animals compared with TgF344-AD animals (F (1, 8) = 5.523, p < 0.05; Supplemental Figure 1D).

In experiment 2, there was a significant main effect of time on male body weights, wherein body weights increased across time from 1–3 months of age (F (2.864, 157.0) = 1147, p < 0.01; Supplemental Figure 2A), as well as 3–6 months of age (F (2.760, 145.0) = 230.1, p < 0.01) and 6–9 months of age (F (2.897, 144.6) = 141.2, p < 0.01). Similarly, there was a main effect of time on consumption in males, where consumption decreased across time from 1–3 months of age (F (4.075, 54.46) = 8.947, p < 0.01; Supplemental Figure 2B), though no differences were observed in consumption from 3–6 months (p > 0.05) nor from 6–9 months (p > 0.05). In females, there was a main effect of time on body weights, wherein body weights increased across time from 1–3 months of age (F (3.203, 160.2) = 3994, p < 0.01; Supplemental Figure 2C), as well as from 3–6 months (F (6.031, 313.6) = 140.1, p < 0.01) and 6–9 months (F (6.384, 325.6) = 89.64, p < 0.01). We also observed main effects of genotype on body weight, wherein wild type animals weighed significantly less than TgF344-AD animals at 6 months in both males (F (1, 54) = 5.076, p < 0.05; Supplemental Figure 2A) as well as females (F (1, 51) = 33.29, p < 0.01; Supplemental Figure 2C). This difference in body weight continued at 6–9 months in females (F (1, 51) = 33.29, p < 0.01), though not in males (p > 0.05). Interestingly, no differences in consumption were observed in females across time, nor as a function of genotype (p > 0.05, Supplemental Figure 2D).

Experiment 2:

Elevated Plus Maze.

In males, no differences were observed in percent time spent in the open arms of the elevated plus maze at 3 months old (p > 0.05; Figure 3A). Similarly, no differences were observed in the number of entries into the closed arms of the elevated plus maze at 3 months old (p > 0.05, Figure 3B). In females, no differences were observed in percent time spent in the open arms at 3MO (p > 0.05, Figure 3C), though an interaction was observed in the number of entries made into the closed arms in females (F (1, 51) = 4.126, p < 0.05; Figure 3D) where TgF344-AD water-consuming females made significantly more entries into the closed arms than wild type water-consuming females (p < 0.05).

Figure 3. Elevated Plus Maze.

(A) Male percent time spent in open arms; TgF344-AD (&) spent less time in open arms at 6 and 9MO. (B) Male entrances to closed arms; TgF344-AD decreased entrances to closed arms at 6 and 9MO. (C) Female percent time spent in open arms; TgF344-AD females spent less time in open arms at 6 and 9MO; TgF344-AD females with a history of ethanol spent significantly less time in the open arms (*) compared with Wild type females with a history of ethanol. (D) Female entrances to closed arms; TgF344-AD females with a history of water made significantly more entrances to the closed arms compared with WT females with a history of water (*). All data expressed as mean ± SEM.

At 6 months of age, there was a main effect of genotype in males, such that TgF344-AD males spent significantly less time in the open arms of the maze compared with wild type males (F (1, 49) = 5.794, p < 0.05). Similarly, a main effect of genotype was observed in the number of closed arm entries, such that wild type males made significantly more entries into the closed arms compared with TgF344-AD males (F (1, 49) = 4.420, p < 0.05) at 6MO. Similarly, a main effect of genotype was observed in females, such that wild type females spent significantly more time in the open arms compared with TgF344-AD females (F (1, 52) = 10.15, p < 0.01). Interestingly, an interaction was also observed at this time point (F (1, 52) = 4.579, p < 0.05). Post hoc analysis using Tukey’s multiple comparisons test revealed that wild type ethanol-consuming females spent significantly more time in the open arms when compared with TgF344-AD ethanol-consuming animals (p < 0.01). In contrast, no differences were observed in the number of entries made into the closed arms by females at 6 months (p > 0.05)

At 9 months old, there was a main effect of genotype in males, such that TgF344-AD males spent significantly less time in the open arms of the maze compared with wild type males (F (1, 49) = 7.690, p < 0.01). A main effect of genotype was also observed in the number of closed entries made, such that wild type males made significantly more entries into the closed arms compared with TgF344-AD males at the 9-month time point (F (1, 49) = 4.172, p < 0.05). Similarly, a main effect of genotype was observed at the 9-month time point, wherein wild type females continued to spend more time in the open arms as compared with TgF344-AD females (F (1, 51) = 8.492, p < 0.01), and an interaction was observed (F (1, 51) = 6.347, p < 0.05), such that wild type ethanol-consuming females spent significantly more time in the open arms compared with TgF344-AD females (p < 0.01). In contrast, no differences were observed in the number of entries made into the closed arms by females at 9 months (p > 0.05). No meaningful differences in the number of stretch-attend postures or head dips in any groups (p > 0.05).

Marble Burying task.

In males, no differences were observed in the percentage of marbles buried at the 3-month time point (p > 0.05; Figure 4A). At the 3-month time point in females, a main effect of genotype was observed, such that TgF344-AD females buried significantly more marbles than wild type females at the 3-month time point (F (1, 52) = 8.007, p < 0.01; Figure 4B). In both males and females, no differences were observed at the 6-month time point (p > 0.05). However, at the 9-month time point, a main effect of genotype was observed in males, such that TgF344-AD males buried significantly more marbles compared with wild type males (F (1, 50) = 4.367, p < 0.05). No differences were observed at the 9-month time point in females (P > 0.05).

Figure 4. Marble Burying.

(A) Males percent of marbles buried; TgF344-AD buried more marbles at 9MO. (B) Females percent of marbles buried; TgF344-AD buried more marbles at 3MO. All data expressed as mean ± SEM.

Novelty-Induced Hypophagia.

Due to the possible ceiling effects in the following results, sphericity was not assumed, and the Geisser-Greenhouse correction was applied to correct the degrees of freedom and reduce the possibility of Type 1 error (Abdi, 2010). In the test of Novelty-Induced Hypophagia, males showed no differences in the latency to approach the Nilla wafer at the 3-month time point (p > 0.05; Figure 5A). A main effect of ethanol history was observed in the latency to consume the Nilla wafer, wherein ethanol-consuming males had a shorter latency to consume the Nilla wafer when compared with water-consuming males (F (1, 43) = 4.174, p < 0.05) (Figure 5B). In females, no differences were observed in the latency to approach or to consume the Nilla wafer at 3 months of age (p > 0.05; Figure 5C, 5D). At the 6-month time point in males, there was no effects observed in the latency to approach the Nilla wafer (p > 0.05). In contrast, a significant main effect of genotype was observed in the latency to consume the Nilla wafer, wherein male TgF344-AD rats had a significantly longer latency to consume the Nilla wafer compared with wild type males (F (1, 40) = 4.544, p < 0.05). In females, no differences were observed in the latency to approach the Nilla wafer at 6 months of age (p > 0.05; Figure 5C), nor were observed in the latency to consume the Nilla wafer at 6 months of age (p > 0.05 Figure 5D). At the 9-month time point, a main effect of genotype was observed in the latency to approach the Nilla wafer, such that TgF344-AD males approached the Nilla wafer significantly faster than wild type males (F (1, 48) = 7.573, p < 0.05), though no differences in consumption latency were observed at 9MO in males (p > 0.05). In females, no differences were observed in the latency to approach the Nilla wafer at 9 months of age (p > 0.05). However, a main effect of genotype was observed in the latency to consume the Nilla wafer at 9MO, wild type females had a shorter latency to consume the Nilla wafer when compared with TgF344-AD females (F (1, 50) = 9.820, p < 0.01).

Figure 5. Novelty induced hypophagia.

(A) Male latency to approach wafer; TgF344-AD had a decreased latency to approach at 9MO. (B) Male latency to consume wafer; ethanol reduced latency to consume wafer at 3MO; TgF344-AD increased latency to consume wafer at 9MO. (C) Female latency to approach wafer; no differences observed. (D) Female latency to consume wafer; TgF344-AD increased latency to consume wafer at 9MO. All data expressed as mean ± SEM.

Gene expression.

Inflammatory genes in the dorsal hippocampus, as well as peripheral organs spleen and liver, at 9 Months. Gene expression data was adjusted to housekeeper GAPDH prior to analysis in dHPC and adjusted to cyclophilin in the spleen and liver. A one-way ANOVA was used to determine changes to gene expression from the ultimate control (wild type, water-consuming, saline-injected) within sex, with post hoc analysis performed using a Dunnett’s multiple comparison test. In females, there was a significant change in Gapdh (F(4, 56) = 5.863, p < 0.01), in the dHPC where the TgF344-AD ethanol-consuming group was significantly higher than the control (Figure 6A). Due to this difference, housekeeper has been reported as a separate target gene for transparency.

Figure 6. Inflammatory genes after LPS at 9MO.

All data expressed relative to housekeeper gene (GAPDH); statistics run as a 1 × 5 ANOVA compared to the control saline group within sex; interactions denoted with (*) (A) GAPDH expression increased in TgF344-AD females with a history of ethanol. (B) IL-1β expression increased in ethanol-consuming WT and TgF344-AD males; IL-1β expression increased in all groups in females. (C) IL-6 expression increased in ethanol-consuming WT and TgF344-AD males; IL-6 expression increased in ethanol-consuming WT and water-consuming TgF344-AD females. (D) HMGB1 expression decreased in ethanol-consuming TgF344-AD females. (E) RAGE expression decreased in ethanol and water-consuming TgF344-AD females. All data expressed as mean ± SEM.

Dorsal Hippocampus gene expression.

In males, no differences were observed in gene expression of GAPDH (Figure 6A). Gene expression of Il-1β changed significantly in males (F(4, 54) = 4.768, p < 0.05) where both wild type and TgF344-AD ethanol-consuming groups were significantly higher compared with the control (Figure 6B). As expected, gene expression of Il-6 in males changed significantly (F (4, 54) = 6.916, p < 0.01), where the wild type and TgF344-AD ethanol-consuming groups had significantly higher Il-6 compared with the control (Figure 6C). No differences were observed from control in male Hmgb1 gene expression (p > 0.05; Figure 9D). No differences were observed from control in male Rage expression (p > 0.05; Figure 6E). In males, a significant difference was found in App expression (F (4, 53) = 17.15, p < 0.01), where both TgF344-AD groups had significantly higher expression of App compared with control, as expected (Figure 7A). No differences were observed in male Bace1 expression (p > 0.05, Figure 7B). No differences were observed in male expression of Lrp-1 (Figure 7C). A significant difference was observed in expression of Abcb1 in males (F (4, 54) = 4.916, p < 0.01), where WT ethanol-consuming group and both TgF344-AD groups were significantly higher when compared with control (Figure 7D). No differences were observed in either male Trem2 expression (Figure 10E). Gene expression of Il-1β changed significantly in females (F (4, 56) = 6.638, p < 0.01), where all groups differed from the control (Figure 6B). Gene expression of Il-6 in females changed significantly (F (4, 56) = 3.040, p < 0.05), where the WT ethanol history and the TgF344-AD water-consuming groups were significantly higher than control (Figure 6C). There was a significant change in gene expression of Hmgb1 in females (F (4, 56) = 2.667, p < 0.05), where TgF344-AD ethanol-consuming females had significantly lower expression compared with controls (Figure 6D). There was a significant difference in gene expression of Rage in females (F (4, 56) = 3.634, p < 0.05), where both water-consuming and ethanol-consuming TgF344-AD animals had decreased expression of Rage compared with control female rats (Figure 6E). A significant difference was found in App expression in females (F (4, 56) = 26.03, p < 0.01), where both TgF344-AD groups had significantly higher expression of App compared with control (Figure 7A). A significant difference was observed in Bace1 expression in females (F (4, 56) = 2.720, p < 0.05), where TgF344-AD ethanol-consuming females had significantly lower expression of Bace1 compared with control (Figure 7B). A significant difference was observed in Lrp-1 expression in females (F (4, 56) = 6.979, p < 0.01), where TgF344-AD ethanol-consuming females had significantly lower expression of Lrp-1 compared with control (Figure 7C). A significant difference was observed in expression of Abcb1 in females (F (4, 56) = 3.881), where both WT groups, as well as the TgF344-AD water-consuming group, had significantly higher expression of Abcb1 compared with the ultimate control (Figure 7D). No differences were observed in female Trem2 expression (Figure 7E).

Figure 7. AD-related genes after LPS at 9MO.

All data expressed relative to housekeeper gene (GAPDH); statistics run as a 1 × 5 ANOVA compared to the control saline group within sex; interactions denoted with (*) (A) APP expression increased in TgF344-AD males and TgF344-AD females, regardless of ethanol history. (B) BACE1 expression decreased in ethanol-consuming TgF344-AD females. (C) LRP-1 expression decreased in ethanol-consuming TgF344-AD females. (D) ABCB1 expression increased in ethanol-consuming WT males, as well as both water- and ethanol-consuming TgF344-AD males; ABCB1 expression increased in water- and ethanol-consuming WT females, as well as water-consuming TgF344-AD females. (E) No differences observed in TREM2 expression. All data expressed as mean ± SEM.

A one-way ANOVA was also used to evaluate gene expression changes in the liver and spleen back to the ultimate control group. In the male liver, no differences were observed in expression change of Bace1, Lrp-1, or Rantes (p > 0.05). Cyclophilin expression was decreased (F (4, 54) = 2.974, p < 0.05). App expression was increased (F (4, 55) = 2.199, p < 0.01) in the TgF344-AD water and TgF344-AD ethanol group, compared with the control. Gene expression also changed in the genes: IκBα (F (4, 56) = 2.212, p < 0.01), Il-1β (F (4, 55) = 2.541, p < 0.01), Il-6 (F (4, 56) = 3.741 p < 0.01), Tnfα (F (4, 56) = 3.476, p < 0.01), Mcp1 (F (4, 56) = 2.955, p < 0.05), and Mip1α (F (4, 55) = 3.785, p < 0.01), where all groups which received LPS had higher expression compared with the control (Supplemental Table 1). In the female liver, no differences were observed in expression change of Cyclophilin, App, Bace1, Lrp-1, and Rantes (p > 0.05). Gene expression also changed in the following genes: IκBα (F (4, 55) = 1.968, p < 0.01), Il-1β (F (4, 55) = 1.450, p < 0.01), Il-6 (F (4, 55) = 2.281, p < 0.01), Tnfα (F (4, 55) = 1.412, p < 0.01), Mcp1 (F (4, 54) = 2.315, p < 0.01), and Mip1α (F (4, 54) = 2.098, p < 0.01), where all groups which received LPS had higher expression compared with the control (Supplemental Table 1).

In the male spleen, no differences were observed in expression change of Cyclophilin, Bace1, or Lrp-1 (p > 0.05). App expression was increased (F (4, 56) = 0.8505, p < 0.01) in the TgF344-AD water and TgF344-AD ethanol group, compared with the WT water and WT ethanol group, respectively. Gene expression also changed in the genes: IκBα (F (4, 56) = 1.332, p < 0.01), Il-1β (F (4, 56) = 2.419, p < 0.01), Il-6 (F (4, 56) = 3.255, p < 0.01), Tnfα (F (4, 56) = 1.463, p < 0.01), Mcp1 (F (4, 56) = 1.782, p < 0.01), and Mip1α (F (4, 56) = 2.231, p < 0.01), where all groups which received LPS had higher expression compared with the control. In Rantes, gene expression increased (F (4, 56) = 0.216, p < 0.01) in TgF344-AD with a history of water consumption compared with the control (Supplemental Table 2). In the female spleen, no differences were observed in expression change of Cyclophilin, Bace1, or Lrp-1 ( p > 0.05). App expression was increased (F (4, 55) = 2.333, p < 0.01) in the TgF344-AD water and TgF344-AD ethanol group, compared with the control, WT water, and WT ethanol group. Gene expression also changed in the following genes: IκBα (F (4, 56) = 1.332, p < 0.01), Il-1β (F (4, 55) = 2.402, p < 0.01), Il-6 (F (4, 55) = 2.890, p < 0.01), Tnfα (F (4, 55) = 1.777, p < 0.01), Mcp1 (F (4, 55) = 1.744, p < 0.01), and Mip1α (F (4, 55) = 2.889, p < 0.05), where all groups which received LPS had higher expression compared with the control (Supplemental Table 2).

Discussion

A unique feature of the present studies was the inclusion of both female and male transgenic TgF344-AD rats as well as wild type rats to determine vulnerability to AD following ethanol intake in both genetically typical as well as genetically vulnerable populations, as the burden of ethanol across the lifespan may differentially affect the neuropathology of disease. Another unique feature of this set of studies was the long-term ethanol consumption and evaluation of behavioral symptoms of AD across the lifespan in both cognitive tasks (spontaneous alternation, contextual fear conditioning) as well as anxiety tasks (elevated plus maze, marble burying, novelty induced hypophagia). The present studies examined gene expression in adulthood (6 months of age) under basal conditions to assess vulnerability to ethanol in prodromal AD populations, as well as in later adulthood (9 months of age) following an inflammatory challenge in both the CNS and the periphery. Consistent with reports in the literature that females are more vulnerable to AD, our studies largely found that TgF344-AD females were vulnerable to cognitive deficits when compared with wild type rats, whereas very few differences were observed in males. In contrast, increased anxiety-like behavior was observed in both males and females. Consistent with females with AD transgenes being more sensitive behaviorally, females with a history of ethanol consumption displayed consistently more reactivity to an inflammatory stimulus when challenged at 9 months of age.

Spontaneous alternation was used to assess cognitive impairments across time, with deficits in spontaneous alternation observed in transgenic AD animals as early as 6 months old (Miedel et al., 2017). In Experiment 1, female rats with a history of ethanol consumption made fewer alternations in the spontaneous alternation task at the 3-month timepoint, indicating that a history of ethanol consumption disrupted cognitive ability. Despite this, no differences were observed in males at 3 months old, nor any differences in either sex at the 6-month time point. However, it is worth noting that a significant number of animals did not make enough alternations to meet the threshold criteria to be included in the analysis of spontaneous alternation (minimum of 12 arm entrances). Low activity at the later time points could be explained by increased anxiety or otherwise by lower locomotion (Chaney et al., 2021), despite being food deprived to 85% of body weight to encourage exploratory activity, or otherwise by interference of the contextual fear conditioning task.

Contextual fear conditioning is a hippocampus-dependent task, where representation of the context is consolidated; the hippocampus then sends excitatory projections to the basolateral amygdala for integration of the shock and context; finally, excitatory projections to the central amygdala, where expression of fear is driven (Orsini & Maren, 2012). In the present experiment, female TgF344-AD rats had increased freezing behavior during the test of retention at the 3-month time point, indicating that AD-positive females had increased fear retention, whereas no differences were observed in males. Furthermore, TgF344-AD females had increased freezing behavior when reintroduced to the context chambers at 6 months of age, prior to the reinstatement shock, as well as increased freezing behavior during the test of retention to the reinstatement shock. These findings are consistent with reports that animal models of AD result in higher retention of fear memory in contextual fear conditioning paradigms (Stover et al., 2015). We also observed that TgF344-AD rats displayed a delay to extinguish fear following the extinction trial at 6 months of age in both males and females, indicating that males might display emerging behavioral changes associated with AD at 6-months of age.

The findings that TgF344-AD animals had better retention of contextual fear conditioning were initially surprising. Contextual fear conditioning is a hippocampus-dependent task, thus we hypothesized that the accumulation of Aβ plaques in the hippocampus would disrupt fear memory and decrease retention of fear during non-shocked trials. However, the increased fear retention observed may indicate that TgF344-AD rats have increased anxiety. To test this, a variety of anxiety tasks were assessed (EPM, Marble Burying, novelty induced hypophagia) at multiple time points to determine the influence of genetic vulnerability to AD as well as chronic ethanol consumption on anxiety-like behaviors. We observed few differences in anxiety measures at 3 months of age in the elevated plus maze. However, at the 6- and 9-month timepoints, both male and female transgenic rats spent decreased percent time in the open arms of the maze, indicating increased anxiety-like behavior. Interestingly, wild type females with a history of ethanol consumption had the highest time spent in the open arms of the maze, possibly signifying a decrease in risk-aversion. In the marble burying task, we again found an increase in anxiety-like behavior in the transgenic females at the 3-month timepoint, while increases in anxiety-like behavior in transgenic males were observed at the 9-month timepoint. At 3MO in the novelty induced hypophagia test, males with a history of ethanol consumption had a decreased latency to consume the Nilla wafer, indicating lower anxiety-like behavior. At the 6-month timepoint, transgenic males had a higher latency to consume the Nilla wafer, indicating increased anxiety compared with wild types. Similarly, transgenic females had increased latency to consume the Nilla wafer at the 9-month time point. It is important to note that any animal that did not consume or approach the Nilla wafer during the test was given a max latency score of 900s, which many animals reached as they aged. It is currently unclear whether this task may be limited by ceiling effects, particularly in aged animals. A lack of studies investigating novelty induced hypophagia in rats during later aging makes the present data difficult to interpret, but may suggest that this task is inappropriate for aging studies. Together, these results are consistent with other studies showing that TgF344-AD rats had increased anxiety-like behavior and conditioned fear compared with WT F344 rats (Pentkowski et al., 2022).

Next, we examined gene expression from the dorsal hippocampus at the 6-month timepoint under basal conditions. We chose a few genes of interest that related to inflammation, aging, and AD. As the transgenic model had overexpression of App, we expected to see a main effect of genotype in both males and females on APP, which we observed. IL-6, an inflammatory cytokine, is found in elevated levels in elderly patients (Roubenoff et al., 1998), as well as in postmortem AD patients (Miron et al., 2018). We expected to see increases in Il-6 in our TgF344-AD groups; however, we observed a decrease in Il-6 in male TgF344-AD animals, and an increase in Il-6 in female TgF344-AD animals. Notably, these differences were observed under non-stimulated conditions, suggesting that the inflammatory profile of AD vulnerable individuals may differ greatly in males and females at baseline (i.e., in the absence of an inflammatory challenge). BACE1, a protease in gamma-secretase which cleaves APP into Aβ in the amyloidogenic pathway (Nunan & Small, 2000), has previously been shown to increase following chronic ethanol exposure (Kim et al., 2011; Gong et al., 2017). We expected to see increases in Bace1 in transgenic males and females, and that this would be particularly high in ethanol-consuming animals; however, we observed only a main effect of increased Bace1 in transgenic females. RAGE, an influx receptor which actively transports Aβ from the blood into the brain (Deane et al., 2003), is correlated with severity of AD pathology (Xu et al., 2016). Unsurprisingly, we observed that TgF344-AD females had increased Rage expression under basal conditions. LRP-1 is a cell-surface receptor that contributes to the efflux of Aβ from the brain to the periphery across the BBB (Deane et al., 2012; Citron, 2010; Zlokovic, 2004). Interestingly, we found an increase in Lrp-1 expression in TgF344-AD females, suggesting that Aβ was being actively transported out of the brain at higher levels compared with wild types. Taken together with the behavioral data, evidence suggests that TgF344-AD females are uniquely vulnerable to the consequences of AD at the 6-month time point.

At 9 months old (Experiment 2), animals were given an LPS injection prior to tissue collection. We hypothesized that inflammatory challenges, such as LPS, may reveal the “primed” and sensitized state of microglia following a history of ethanol exposure and aging (Norden et al., 2016; Perkins, Piazza, & Deak, 2018). Following LPS challenge in the dorsal hippocampus, we observed that App increased gene expression in transgenic males and females, as expected. Of particular note, our reference housekeeping gene (Gapdh) had significant effects in the females; a number of housekeeper genes were tested and with similarly significant results (data not shown). We observed that all groups of females showed an increase in Il-1β expression, as expected; in contrast, only males with a history of ethanol consumption had increased Il-1β expression, indicating that females may be more vulnerable to increased inflammation following LPS. Similarly, we observed an increase in Il-6 expression in males following a history of ethanol consumption, while this increase in Il-6 was only observed in wild type females with a history of ethanol as well as TgF344-AD females with a water history, indicating that males with a history of ethanol consumption have an altered induced inflammatory response compared with females. In contrast to our first experiment, we observed lowered Hmgb1, Rage, Bace1, and Lrp-1 expression in transgenic females with a history of ethanol following an LPS challenge. These results indicate that while there may be less cleavage of APP to Aβ following an inflammatory insult, but there is also a drastic reduction in Aβ peptides transportation from the brain to the periphery. Overall, many of these findings are in contrast with reports from a 3xTg-AD mouse model, which found that chronic binge ethanol resulted in reductions of protective genes (Apoe, Trem2, Lpl, Ctsd) and increased proinflammatory genes in females but not males (Tucker, Pauneto, Barnett, & Coleman, 2022). The differences observed is likely due to a combination of different species reactivity, as well as the “unmasking” nature of an inflammatory stimulus following a chronic history of ethanol, thereby uncovering primed microglial states of reactivity. Despite these differences observed in the dorsal hippocampus, relatively few changes were observed in the peripheral organs of both males and females (Supplemental Tables 1 & 2).

Spleen and liver metabolism and function may be an area of vulnerability in AD patients, where alcohol damage to peripheral organs during aging may substantially contribute to the development of neurodegenerative disorders. Ethanol exposure has been shown to enhance inflammatory cytokine expression in splenic macrophages and generate an exaggerated inflammatory response in splenic cells after LPS challenge (Sureshchandra et al., 2019). Ethanol exposure has also been shown to increase chemokine expression in peripheral organs (Mandrekar et al., 2011), and alcohol induced MCP-1 signaling has been correlated with heightened microglial activation and brain damage (He & Crews, 2008; Zhang and Lou, 2019). The spleen has also been implicated in AD, as it clears excess amyloid beta from the peripheral organs; dysfunction in the spleen led to abnormal clearing of amyloid beta which may aggravate symptoms and onset of AD (Yu et al., 2022). Similarly, the liver is the main peripheral organ responsible for clearing and catabolizing Aβ peptide from the blood (Ghiso et al., 2004). Aβ peptide uptake into the liver is mainly mediated by LRP-1 and high-density lipoprotein (HDL) transportation, which is expressed in hepatocytes (Tamaki et al., 2006; Hottman et al., 2014). Hepatocyte dysfunction in aged rats is associated with low liver Lrp-1 expression, and with increased plasma Aβ peptides (Tamaki et al., 2006). While many cytokines and chemokines had increased expression in both the spleen and liver following LPS in our experiment, such as IκBα, Il-1β, Il-6, Tnfα, Mcp1, and Mip1α, few differences were observed in either the spleen or liver between groups that had received LPS injection. On the contrary, the dosage of LPS used appears to have created a ceiling effect on the cytokines tested, which overshadowed the subtle impacts of ethanol on immune function in the periphery. Due to the extended duration and constraints of the experiment, the use of differing LPS doses and further control groups were not feasible in the present study.

Similarly, a transgenic mouse model using a modified drinking in the dark procedure found that ethanol exposure altered APP processing in transgenic mice by increasing the levels of App and Bace1, as well as by increasing Aβ protein production (Huang et al., 2018). Conversely, our experiments found relatively few effects of ethanol on behavior or gene expression in TgF344-AD rats, whereas most of our findings were attributed to the genotype. A major limitation of the present set of studies was a lack of control over ethanol administration, which may explain the lack of effects observed in the present studies as a result of ethanol history. The intermittent drinking model was chosen to have a low-stress, low-intervention method of ethanol delivery with high face validity that could be implemented across the lifespan without concerns of stress associated with repeated experimenter handling, injections, or gavage. Our lab has previously used this method of alcohol administration and found the method of administration to be well tolerated and to reach significant levels of ethanol intake in which BECs were observed in the 40–100 mg/dl range (Marsland et al., 2023A; Marsland et al., 2023B). However, the ethanol content can only be estimated, as we housed 2 animals in a cage together and split the ethanol consumption to an estimation based on the body weight of both animals within the cage to prevent stress associated effects of prolonged social isolation. It is possible that ethanol preference or dominance dynamics may contribute to how much ethanol was consumed by each individual animal, which cannot be clarified using the estimates of ethanol intake used in this procedure. Future studies should examine the effect of ethanol at both higher and lower doses to fully understand the impact of chronic ethanol consumption on AD.

The present studies evaluated the impact of chronic, lifelong alcohol use on the development of AD pathology and symptoms throughout adulthood in both males and females, genetically typical and atypical populations. Overall, the present studies support the conclusions that transgenic models of AD have sex-specific behavioral effects, wherein females with AD-transgenes have more severe cognitive deficits as well as significantly increased anxiety-like behavior. Interestingly, under basal conditions, males had decreased inflammatory cytokine expression, while females had increased inflammatory cytokine expression, as well as increased AD-related protein gene expression. In contrast, under inflammatory challenge, a reduction in several inflammatory and AD-associated genes were observed, particularly in TgF344-AD females with a history of ethanol, indicating that these females may have altered microglial and Aβ interaction dynamics. Taken together, the present studies support the growing evidence that females are uniquely susceptible to AD, and this susceptibility may be due to altered APP / Aβ processing and removal from the CNS, which are altered by ethanol history.

Supplementary Material

Supplemental Figure 2

Supplemental Figure 2. Characterization of ethanol model to 9MO. (A) Body weight in male rats across experimental timeline; time increased body weights at 3 6, and 9 months; TgF344-AD weighed more at 6 months. (B) Ethanol consumption in male rats across experimental timeline; no differences observed. (C) Body weight in female rats across experimental timeline; time increased body weights at all time points; TgF344-AD increased body weights at 6 and 9 months. (D) Ethanol consumption in female rats across experimental timeline; no differences observed. All data expressed as mean ± SEM.

Supplemental Figure 1

Supplemental Figure 1. Characterization of ethanol model to 6MO. (A) Body weight in male rats across experimental timeline; time increased body weights at 3 and 6 months; ethanol consuming animals weighed less than water consuming animals at both 3 and 6 months. (B) Ethanol consumption in male rats across experimental timeline; rats consumed less ethanol across time during 1–3MO drinking period. (C) Body weight in female rats across experimental timeline; time increased body weights at all time points; TgF344-AD increased body weights at 6 MO. (D) Ethanol consumption in female rats across experimental timeline; ethanol consumption decreased across time during the 1–3MO drinking period; wild type females consumed more ethanol compared with TgF344-AD females. All data expressed as mean ± SEM.