The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits

Abstract

Late-onset alzheimer’s disease is the most common form of dementia, and it arises from complex genetic and environmental interactions. Preclinical models that replicate the slow progression and long prodromal phase of late-onset alzheimer’s disease are critical for identifying early therapeutic targets. The LOAD2 mouse model, developed on the C57BL/6J genetic background, integrates key late-onset alzheimer’s disease genetic risk factors: APOE4, Trem2*R47H, and an App allele encoding humanised amyloid beta. This study aimed to characterise key disease-relevant phenotypes of LOAD2 mice during ageing. Behavioural assays were conducted on 18- and 24-month-old LOAD2 and age-matched C57BL/6J wild-type control mice. At 18 months, LOAD2 mice exhibited significantly reduced locomotor activity compared to wild-type controls. However, this difference was diminished at 24 months as wild-type mice displayed an age-related decline in total distance travelled. Similarly, anxiety-like behaviour was elevated in 18-month-old LOAD2 mice relative to wild-type controls, but this difference was no longer evident at 24 months due to increased anxiety levels in aged wild-type mice. In contrast, spatial working memory and associative fear learning were intact in both LOAD2 and wild-type mice at 18 and 24 months of age, indicating no age- or genotype-dependent deficits in these forms of memory. Both groups of mice performed equally poorly in novel object and novel location recognition tasks at both ages. Thus, compared to age-matched wild-type mice, LOAD2 mice exhibit early locomotor deficits and heightened anxiety, but not overt cognitive impairment beyond that of normal ageing. These phenotypes are reminiscent of prodromal symptoms of late-onset alzheimer’s disease in humans.

Keywords

Late-onset alzheimer’s disease ageing prodromal symptoms locomotion anxiety learning and memory

Introduction

Alzheimer’s disease (AD) is a neurodegenerative disorder that occurs in two major forms: a relatively rare, genetically determined early-onset form, and a far more common late-onset form that accounts for the majority of dementia cases (Baumgart et al., 2015; Sloane et al., 2002). Compared to early-onset familial AD, late-onset alzheimer’s disease (LOAD) is characterised by a slow progression and a prolonged prodromal phase, underscoring the need for preclinical models that better reflect the complexity of the human condition. As such, a novel AD mouse model (LOAD2) has been developed based on the C57BL/6J background, integrating three well-known LOAD genetic factors: the human APOE4 allele, the R47H variant in Trem2 (triggering receptor expressed on myeloid cells 2), and App encoding humanised amyloid beta (hAβ) (Kotredes et al., 2024). The APOE4 allele is common in humans, occurring in approximately 14% of the global population, and is the strongest genetic risk factor for LOAD (Lumsden et al., 2020; Ward et al., 2012), while the Trem2 R47H variant is rare, but it confers one of the largest effect sizes among non-APOE risk alleles (Ayer et al., 2019; Cheng-Hathaway et al., 2018). The human Aβ sequence, while not a LOAD risk factor per se, differs from the mouse version at three amino acid residues (5, 10, and 13), and evidence suggests that it is more amyloidogenic than murine Aβ (Xu et al., 2015). Therefore, this model offers a unique opportunity to investigate how known susceptibility alleles in combination with hAβ affect behaviour and cognition across ageing, independent of overt plaque or tau pathology (Kotredes et al., 2024).

Although cognitive impairment remains a hallmark of AD, early disruptions in locomotor function and anxiety regulation are increasingly recognised as contributors to disease progression and patient distress (Laganà et al., 2022; Mendez, 2021; Ogawa et al., 2018; Thaliath and Pillai, 2025). However, despite their clinical relevance, such behavioural domains remain underexplored in preclinical models. Here, we aimed to examine age-dependent alterations in locomotor activity and anxiety-related behaviour in the LOAD2 mouse model, alongside cognitive tests assessing spatial working memory, novelty recognition memory, and associative fear memory. We hypothesised that LOAD2 mice would display an earlier onset of age-related behavioural abnormalities compared to wild-type mice, resembling prodromal AD symptoms in humans.

Methods

Animal model

To model LOAD, we used the LOAD2 mouse strain (B6J.APOE4.Trem2*R47H.hAβ triple homozygous, JAX #030670) developed by the MODEL-AD consortium (Kotredes et al., 2024). C57BL/6J mice, also obtained from JAX, were used as age-matched wild-type (WT) controls. The LOAD2 line and WT controls were maintained independently by within-strain breeding in the same facility. Behavioural testing was conducted at 18 and 24 months of age to model LOAD in humans, for which prevalence rises substantially between the ages of 65 and 85 years (Bekris et al., 2010; Fox, 2007). Mice were bred and maintained at The Centre for Phenogenomics (TCP; Toronto, Canada) under a 12 h light/dark cycle (lights on at 08:00), with ad libitum access to standard rodent chow diet (Teklad 2918X) and water. Animals were group-housed (2–5 per cage). All procedures adhered to Canadian Council on Animal Care (CCAC) guidelines and were approved by the institutional Animal Care Committee at TCP.

Genotyping

Genomic DNA was prepared from tail or ear biopsies collected at weaning using the HotSHOT alkaline lysis method. Briefly, samples were incubated in 50-μL alkaline lysis buffer (25 mM NaOH, 0.2 mM disodium ethylenediaminetetraacetic acid (EDTA)) at 95°C for 10–30 min, then neutralised by addition of 50-μL neutralisation buffer (40 mM Tris-HCl), without pH adjustment. The Jackson Laboratory genotyping protocols for Trem2 R47H and humanised APP (stock 030670; protocol 28934) rely on real-time polymerase chain reaction (PCR). Given equipment availability and the added technical complexity of real time PCR, we instead used conventional PCR-based genotyping for these alleles. Genotyping was performed by conventional PCR or tetra-primer amplification refractory mutation system PCR (tetra ARMS-PCR), following established protocols (Peng et al., 2018; Truett et al., 2000; Ye, 2001). Primers were designed based on published sequences for humanised APOE4 (Foley et al., 2022), Trem2 R47H (Xiang et al., 2018), and humanised APP (hAβ) (Kotredes et al., 2024), using the Primer1 web tool (http://primer1.soton.ac.uk/primer1.html), and are listed in Supplementary Table 1. Since the hAβ primers used in the previous study (Kotredes et al., 2024) cannot distinguish mutant from WT, we designed and used alternative primer sets (Supplementary Table 1). For the Trem2 and APP alleles, we implemented a restriction enzyme digest of the amplicon generated by the outer primer pair as an alternative genotyping approach. PCR was run using a touchdown cycling programme as follows: initial denaturation was performed at 95°C for 5 min, followed by 10 touchdown cycles of 95°C for 30 s, annealing at 65°C with a 0.5°C decrease per cycle, and extension at 68°C for 10 s. This was followed by 28 amplification cycles of 95°C for 15 s, 60°C for 15 s, and 72°C for 10 s, with a final extension at 72°C for 2 min and a hold at 4°C. For tetra ARMS-PCR, the outer primer pair and inner allele-specific primer pair were mixed at a 1:10 ratio. Reactions were prepared using DNA Fast Extract 2× HS-Red Taq Mix (Wisent Bioproducts; Cat. 801-200-HS), and PCR products were resolved on 2% agarose gels.

Body weight and survival analysis

Body weights were recorded at 18 and 24 months of age. Survival analysis was conducted using observational records from the animal facility. Mice reaching humane endpoints were recorded as deceased at the corresponding age. Animals euthanized for experimentation were not counted as deaths.

Behavioural assays

Behavioural testing was performed in three independent cohorts: one cohort (A) was assessed at 18 months of age, and two separate cohorts (B and C) were assessed at 24 months of age. Cohorts comprised different mice, and no animals were tested in more than one cohort (Figure 1(a)). This design was selected to minimise carryover effects from prior test experience. In particular, fear conditioning involves an aversive foot shock, and prior foot shock or behavioural stress exposure can modulate later fear learning and related behavioural readouts (Cnops et al., 2022; Iwasaki et al., 2015; McIlwain et al., 2001). Therefore, independent cohorts were used for each age to ensure that comparisons between 18 and 24 months were not confounded by repeated stress exposure or training history. All behavioural experiments were conducted by blinded investigators following a 5-day habituation period (3 min/day handling). Testing was carried out at the TCP Neurobehaviour Core in the following order: open field test (OFT), elevated zero maze (EZM), Y-maze, novel object recognition (NOR) test, novel location recognition (NLR) test, and contextual fear conditioning (CFC) (Figure 1(a)).

Figure 1.

Study design, body weight, and survival in aged WT and LOAD2 mice. (a) Schematic of cohort structure and experimental timeline. Both male and female C57BL/6J wild-type (WT) and LOAD2 mice were bred and aged to either 18 or 24 months. Following acclimation and 5-day handling period, mice underwent a battery of behavioural tests. There were three independent cohorts (one cohort at 18 months and two separate cohorts at 24 months), and each animal was tested at only one age. Figure created in BioRender Kang, S. (https://BioRender.com/8ktbhvp) is licensed under CC BY 4.0. (b) For the 18-month-old group, both WT and LOAD2 mice showed sex-related differences in body weight, with males weighing significantly more than females (WT male: n = 13; WT female: n = 17; LOAD2 male: n = 20; LOAD2: female n = 16). However, there were no genotypic differences in body weight within each sex. In the set of 24-month-old subjects, no sex- or genotype-specific differences in body weight were evident (WT male: n = 9; WT female: n = 18; LOAD2 male: n = 9; LOAD2 female: n = 9). (c) Survival curves for male and female mice of each genotype. Female mice showed significantly decreased survival relative to males within each genotype, but no genotype-dependent differences were detected (WT male: n = 165; WT female: n = 139; LOAD2 male: n = 187; LOAD2 female: n = 169). All data are presented as mean ± SEM. Statistical analyses are detailed in the main text. Where individual data points are shown, filled symbols are males and open symbols are females. *p < 0.05; **p < 0.01.

OFT

Locomotor activity was assessed using SuperFlex Open Field arenas (42 cm × 42 cm × 30 cm; Omnitech Electronics, Columbus, OH, USA) under uniform lighting (200 lux). Each session lasted 20 min. Infrared beam breaks were recorded to calculate total distance travelled. Data were analysed using Fusion software (Omnitech Electronics).

EZM test

Anxiety was assessed in the EZM (Noldus Information Technology, Netherlands), which consisted of a circular platform (50 cm diameter; 5 cm track width) elevated 50 cm from the floor. The maze contained two open and two closed quadrants. Mice were allowed to explore freely for 5 min and an automated tracking system (EthoVision XT; Noldus) recorded time spent in open and closed areas.

Y-maze spontaneous alternation test

Y-maze spontaneous alternation was assessed in a ‘Y’-shaped white acrylic maze as a measure of spatial working memory. Mice were placed in the centre of the maze and allowed to explore freely for 5 min. Arm entries and total distance travelled were recorded using an automated tracking system (EthoVision XT, Noldus). Spontaneous alternation was defined as consecutive entries into three different arms, scored on overlapping triplet sequences. For example, entries A to C to B constitute an alternation, whereas A to B to A do not. Percent alternation was calculated using the formula

A l t e r n a t i o n I n d e x (%) = (\frac{number of alternations}{total possible alternations}) \times 100

where total possible alternations equalled the number of arm entries minus two.

The theoretical chance level for this metric in a three-arm maze was calculated as 22.2%, derived under the assumption of random arm selection with equal probability among the three arms, where the probability that three consecutive choices are all different is $1 \times \frac{2}{3} \times \frac{1}{3} = \frac{2}{9} = 22.2 %$ . Accordingly, alternation performance was interpreted relative to this chance level.

NOR and NLR tests

Novelty recognition memory was assessed using the NOR and NLR tasks, adapting commonly used rodent protocols (Antunes and Biala, 2012; Ennaceur and Delacour, 1988). Testing was conducted in an open field arena under uniform lighting conditions and was tracked using EthoVision XT (Noldus).

On the first day of NOR, mice were habituated to the empty arena for 5 min to reduce novelty-induced exploration and to provide a baseline measure of locomotor activity (total distance travelled). On the following day, mice completed a 10-min training trial with two identical, nontoxic objects placed in the arena. After a 1-h inter-trial interval, mice were returned to the arena for a 5-min test trial in which one familiar object was replaced with a novel object. On the next day (NLR), mice were trained for 10 min with two identical objects that were not used in NOR, with distinct distal visual cues present around the arena. After a 1-h inter-trial interval, mice were reintroduced for a 5-min test trial in which one object was moved to a novel location, while the other remained in the familiar location.

Exploration was defined as directed investigation of an object, scored as the time the nose was within 2 cm of the object. Time spent climbing or sitting on the object was not scored as exploration. Mice with less than 5 s total object exploration during the test trial were excluded. Novel object type (for NOR), or the novel location (for NLR), was counterbalanced across mice. The arena and objects were cleaned with 70% isopropanol between trials to minimise olfactory cues.

The discrimination index (DI), which quantifies the preference for the novel object/location, was calculated as follows

D i s c r i m i n a t i o n I n d e x = \frac{Time_novel}{Time_novel + Time_familiar}

CFC

CFC was conducted using a standard fear conditioning protocol. Mice received a single sequential tone-shock pairing (tone: 85 dB, 2.8 kHz, 30 s; shock: 0.75 mA, 2 s) on Day 1 of CFC. Contextual recall was tested 24 h later (Day 2) in the original chamber. Cue-specific fear was tested 2 h after the contextual recall in a novel context with tone presentations (tone: 85 dB, 2.8 kHz, 2.5 min). Freezing behaviour was quantified using automated video tracking (Med Associates, St. Albans, VT, USA).

Statistical analysis

Data are presented as mean ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism v10. Normality of each dataset was assessed with the Shapiro–Wilk test to guide the use of parametric versus non-parametric analyses. One-sample t-tests were used to determine whether performance differed from chance where appropriate (22.2% for Y-maze and DI of 0.5 for NOR and NLR). Between-group comparisons were performed using log rank (Mantel–Cox) tests for survival data, and three-way or two-way analysis of variance (ANOVA) for body weight and behavioural outcomes, respectively, with post hoc multiple comparisons as indicated in the results. For behavioural assays, both male and female mice were included, and data are displayed by sex for qualitative visualisation where individual datapoints are plotted (with filled symbols representing males, and open symbols representing females). Sex was not included as a factor in the prespecified statistical models, and analyses therefore pooled males and females, except for body weight and survival, where sex differences were expected and sample sizes permitted sex-stratified analyses. All mean, SEM, and N numbers for the behavioural data from this study (OFT, EZM, Y-maze, NOR, NLR, and CFC) are listed in Supplementary Data 1.

Results

Sex- and age-matched LOAD2 and WT mice exhibit similar body weight and survival

Body weight was analysed using three-way ANOVA with sex, age (18 vs. 24 months), and genotype (WT vs. LOAD2) as factors (Figure 1(b)). The analysis detected main effects of sex (F(1,103) = 17.25, p < 0.001) and genotype (F(1,103) = 10.90, p = 0.001), with no main effect of age (F(1,103) = 0.245, p = 0.622). There was a sex × age interaction (F(1,103) = 4.453, p = 0.037), whereas sex × genotype, age × genotype, and the three-way interaction were not significant (all p ⩾ 0.189). Tukey’s multiple comparisons showed that males weighed more than females at 18 months in both WT (n = 13/17 (male/female), p = 0.006) and LOAD2 (n = 20/16 (male/female), p = 0.030) cohorts, while no sex differences were detected in those aged to 24 months (WT: n = 9/18 (male/female), p = 0.996; LOAD2: n = 9/9 (male/female), p = 0.955). In contrast, genotypic differences (WT vs. LOAD2) were not detected within age- and sex-matched groups (18-month-old males: p = 0.997; 24-month-old males: p = 0.362; 18-month-old females: p = 0.771; 24-month-old females: p = 0.498). Thus, although the three-way ANOVA indicated an overall genotype effect when averaged across sex and age, we conclude that LOAD2 mice do not differ significantly in body weight from their age- and sex-matched WT counterparts. The underlying raw data and complete statistical outputs for this analysis are provided in Supplementary Data 2.

Survival analysis showed sex-dependent differences within each genotype, with males exhibiting longer lifespans than females (Log-rank Mantel–Cox test; WT male (n = 165) vs. female (n = 139): χ²(1) = 7.850, p = 0.005; LOAD2 male (n = 187) vs. female (n = 168): χ²(1) = 11.60, p < 0.001; (Figure 1(c)). However, there was no effect of genotype on survival (WT vs. LOAD2 males: χ²(1) = 0.311, p = 0.577; WT vs. LOAD2 females: χ²(1) = 0.198, p = 0.657; Figure 1(c)).

Accelerated onset of locomotor and anxiety deficits in LOAD2 mice

Locomotor activity was assessed using the OFT in WT and LOAD2 mice at 18 or 24 months. Total distance travelled over 20 min was analysed by two-way ANOVA with age and genotype as factors (Figure 2 (a) and (b)). There was a significant age × genotype interaction (F(1,44) = 11.56, p = 0.001) and a main effect of genotype (F(1,44) = 6.242, p = 0.016), with no main effect of age (F(1,44) = 0.257, p = 0.615). Post hoc multiple comparisons (Šidák’s test) showed that at 18 months LOAD2 mice travelled less than WT controls (WT: 40.70 ± 2.47 m, n = 10; LOAD2: 27.74 ± 2.79 m, n = 9; p < 0.001), whereas no genotypic difference was detected at 24 months (WT: 32.12 ± 1.95 m, n = 12; LOAD2: 34.10 ± 1.70 m, n = 17; p = 0.730). Within-genotype comparisons revealed an age-related reduction in distance travelled in WT mice (p = 0.016), whereas locomotor activity did not differ significantly between 18- and 24-month-old cohorts in LOAD2 mice (p = 0.084). Together, these findings suggest that LOAD2 mice undergo an earlier onset of locomotor hypoactivity compared to WT mice.

Figure 2.

Earlier emergence of reduced locomotor activity and increased anxiety in LOAD2 mice. (a) Schematic of the open field test (OFT) apparatus. (b) Total distance travelled by WT and LOAD2 mice, aged either 18 or 24 months, during a 20 min OFT session. For the 18-month-old cohort, LOAD2 mice exhibited significantly reduced locomotor activity compared to WT controls (WT: n = 10 (male 5, female 5); LOAD2: n = 9 (male 4, female 5)). A genotypic difference was not observed in the 24-month-old subjects (WT: n = 12 (male 6, female 6); LOAD2: n = 17 (male 9, female 8)). WT mice showed an age-related decline in distance travelled between 18 and 24 months, while LOAD2 mice travelled a similar distance for both age groups, suggesting an early reduction in locomotor activity in LOAD2 mice. (c) Schematic of the elevated zero maze apparatus. (d) Percentage time spent in the open quadrants of the EZM by WT and LOAD2 mice at 18 or 24 months. LOAD2 mice aged 18 months spent significantly less time than WT mice in the open quadrants, indicating elevated anxiety levels (WT: n = 10 (male 5, female 5); LOAD2: n = 9 (male 4, female 5)). However, no genotypic difference was evident in the 24-month group (WT: n = 12 (male 6, female 6); LOAD2: n = 17 (male 9, female 8)). WT mice exhibited a significant increase in anxiety-like behaviour between 18- and 24-month cohorts, while anxiety levels in LOAD2 mice remained unchanged, suggesting a premature onset of affective symptoms in the LOAD2 model. All data are presented as mean ± SEM. Statistical analyses are detailed in the main text. Where individual data points are shown, filled symbols are males and open symbols are females. *p < 0.05; **p < 0.01; ***p < 0.001. The schematics in panels (a) and (c) were created in BioRender. Kang, S. (https://BioRender.com/8ktbhvp) is licensed under CC BY 4.0.

Anxiety-like behaviour was assessed using the EZM, which exploits rodents’ natural aversion to open and elevated spaces (Kulkarni et al., 2007). The percentage of time spent in the open quadrants was analysed by two-way ANOVA with age and genotype as factors (Figure 2(c) and (d)). There was a significant age × genotype interaction (F(1,44) = 6.742, p = 0.013), whereas the main effects of age and genotype were not significant (age: F(1,44) = 3.242, p = 0.079; genotype: F(1,44) = 1.225, p = 0.274). Post hoc pairwise comparisons (Šidák’s test) indicated that LOAD2 mice spent less time in the open quadrants than WT controls at 18 months (WT: 13.55 ± 2.34%, n = 10; LOAD2: 7.13 ± 1.81%, n = 9; p = 0.041), but not at 24 months (WT: 5.93 ± 1.19%, n = 12; LOAD2: 8.51 ± 1.46%, n = 17; p = 0.432). When comparing across age by genotype, WT mice showed a significantly reduced open quadrant time in the 24- versus 18-month-old cohorts (p = 0.008), whereas LOAD2 mice did not show a significant age-related difference in this measure (p = 0.814). This result suggests that increased anxiety-like behaviour emerges earlier in LOAD2 mice, possibly due to an altered trajectory of age-dependent anxiety (Lou et al., 2025; Nolte et al., 2019).

Both WT and LOAD2 mice have intact spatial working memory but impaired novelty recognition at 18 and 24 months

To assess spatial working memory, the Y-maze spontaneous alternation test was conducted in WT and LOAD2 mice aged for either 18 or 24 months (Figure 3(a)). One-sample t-tests against chance level (theoretical mean = 22.2%; see methods) showed that alternation was above chance in all groups, indicating intact working memory in both genotypes at both ages (WT 18 months: 51.52 ± 4.02%, n = 10, t(9) = 7.286, p < 0.001; LOAD2 18 months: 59.54 ± 2.88%, n = 9, t(8) = 12.99, p < 0.001; WT 24 months: 61.27 ± 3.18%, n = 12, t(11) = 12.27, p < 0.001; LOAD2 24 months: 61.16 ± 2.55%, n = 17, t(16) = 15.29, p < 0.001; Figure 3(b)). Furthermore, two-way ANOVA analysis showed no main effects of age (F(1, 44) = 3.127, p = 0.226) or genotype (F(1, 44) = 3.127, p = 0.084) on spontaneous alternation, as well as no age × genotype interaction (F(1, 44) = 1.599, p = 0.213), suggesting performance in this task was similar for all groups. To evaluate overall task engagement during the test, the total number of alternations, defined as total arm entries minus 2, was also analysed (Supplementary Figure 1(a)). Two-way ANOVA analysis showed a main effect of age on this measure (F(1, 44) = 4.546, p = 0.039), but no effect of genotype (F(1, 44) = 0.281, p = 0.599) and no age × genotype interaction (F(1, 44) = 2.823, p = 0.100). Post hoc pairwise comparisons (Šidák’s test) revealed that the total number of alternations was significantly reduced in WT mice aged 24 months compared to those aged 18 months (18 months: 24.00 ± 3.10; 24 months: 16.33 ± 1.23; p = 0.022), indicative of an age-dependent decline in overall task engagement.

Figure 3.

Preserved spatial working memory but impaired novelty recognition memory in both WT and LOAD2 aged mice. (a) Schematic of the Y-maze spontaneous alternation task. (b) Percent spontaneous alternation in the Y-maze for 18- and 24-month-old groups of WT and LOAD2 mice. The dashed line indicates chance performance (22.2%). One-sample t-tests showed alternation above chance in all groups, consistent with intact spatial working memory (WT 18 months: n = 10 (male 5, female 5); LOAD2 18 months: n = 9 (male 4, female 5); WT 24 months: n = 12 (male 6, female 6); LOAD2 24 months: n = 17 (male 9, female 8)). (c) Schematic of the novel object recognition (NOR) task with a 1-h inter-trial interval. (d) NOR discrimination index (DI) during the test trial in WT and LOAD2 mice for 18- and 24-month-old groups. The dashed line indicates chance performance (DI = 0.5). One-sample t-tests did not detect novel object discrimination in any group (18 months WT: n = 10 (male 5, female 5); 18 months LOAD2: n = 8 (male 4, female 4); 24 months WT: n = 10 (male 5, female 5); 24 months LOAD2: n = 11 (male 3, female 8)). (e) Schematic of the novel location recognition (NLR) task with a 1-h inter-trial interval. (f) NLR discrimination index during the test trial for 18- and 24-month-old groups of WT and LOAD2 mice. The dashed line indicates chance performance (DI = 0.5). One-sample t-tests did not detect novel location discrimination in any group (18 months WT: n = 10 (male 5, female 5); 18 months LOAD2: n = 8 (male 4, female 4); 24 months WT: n = 11 (male 6, female 5); 24 months LOAD2: n = 15 (male 7, female 8)). All data are presented as mean ± SEM. Statistical analyses are detailed in the main text. Where individual data points are shown, filled symbols are males and open symbols are females. ***p < 0.001. The schematics in panels (a), (c), and (e) were created in BioRender. Kang, S. (https://BioRender.com/8ktbhvp) is licensed under CC BY 4.0.

To assess novelty recognition memory, 18- or 24-month-old WT and LOAD2 mice were tested in the NOR and NLR tasks with a 1-h retention interval. Performance was quantified using the DI, with one-sample t-tests used to compare against chance level in each group (theoretical mean DI = 0.5) (Figure 3(c)–(f)). In the NOR test (Figure 3(c)), none of the groups differed from chance, indicating no reliable preference for the novel object (WT 18 months: 0.47 ± 0.05, n = 10, t(9) = 0.761, p = 0.466; LOAD2 18 months: 0.59 ± 0.12, n = 8, t(7) = 0.716, p = 0.497; WT 24 months: 0.57 ± 0.09, n = 10, t(9) = 0.841, p = 0.422; LOAD2 24 months: 0.65 ± 0.07, n = 11, t(10) = 2.220, p = 0.051; Figure 3(d)). Similarly, in the NLR task (Figure 3(e)), DIs did not differ from chance in any group, indicating no preference for the novel location (WT 18 months: 0.40 ± 0.07, n = 10, t(9) = 1.487, p = 0.171; LOAD2 18 months: 0.53 ± 0.09, n = 8, t(7) = 0.330, p = 0.751; WT 24 months: 0.41 ± 0.07, n = 11, t(10) = 1.285, p = 0.228; LOAD2 24 months: 0.49 ± 0.06, n = 15, t(14) = 0.258, p = 0.800; Figure 3(f)). For both the NOR and NLR tasks, two-way ANOVA analysis showed no significant main effects of age (NOR: F(1, 35) = 1.161, p = 0.289; NLR: F(1, 40) = 0.032, p = 0.859) or genotype (NOR: F(1, 35) = 1.555, p = 0.221; NLR: F(1, 40) = 2.277, p = 0.139), and no age × genotype interaction (NOR: F(1, 35) = 0.066, p = 0.799; NLR: F(1, 40) = 0.131, p = 0.719). Thus, WT and LOAD2 mice performed equally poorly in these tasks at both 18 and 24 months, likely reflecting a natural decline in recognition memory in ageing mice (Traschütz et al., 2018).

To evaluate whether object exploratory drive differed as a function of age or genotype, total exploration time during the NOR and NLR test trials was analysed using two-way ANOVA (Supplementary Figure 1(b) and (c)). In NOR, there was a significant main effect of age on exploration time (F(1, 35) = 16.02, p < 0.001), with no main effect of genotype (F(1, 35) = 1.113, p = 0.299) and no age × genotype interaction (F(1, 35) = 2.411, p = 0.129). Post hoc pairwise comparisons (Šidák’s test) showed that WT mice aged 24 months spent less time exploring the objects than WT mice of 18 months (18 months: 62.47 ± 5.77 s; 24 months: 22.98 ± 3.57 s; p < 0.001), suggesting an age-dependent decline in novelty-seeking behaviour. Similarly, in NLR, there was a main effect of age on exploration time (F(1, 40) = 4.988, p = 0.031), but no effect of genotype (F(1, 40) = 0.032, p = 0.859), and no age × genotype interaction (F(1, 40) = 0.259, p = 0.614). These results suggest that, while there was an age-related reduction in object exploration overall, there were no genotype-dependent differences in exploratory behaviour in both the NOR and NLR tasks.

Associative fear conditioning is not impaired in aged LOAD2 mice

The CFC test is a well-established method for evaluating learning and memory in rodents (Trent et al., 2025). To assess contextual memory and associative learning in the LOAD2 mouse model, we conducted contextual and cued fear conditioning at 18 or 24 months of age, with age-matched WT mice used as controls (see Figure 4(a) for the procedure). For freezing behaviour in 18-month-old mice, two-way RM ANOVA showed main effects for testing session (F(2.147, 36.49) = 34.81, p < 0.001) and genotype (F(1, 17) = 12.06, p = 0.003), with no session × genotype interaction (F(2.147, 36.49) = 2.989, p = 0.059) (WT: n = 10, LOAD2: n = 9; Figure 4(b)). Tukey’s multiple comparisons test showed that LOAD2 mice displayed elevated freezing compared to WT mice during both the baseline session (WT: 7.10 ± 2.12%; LOAD2: 20.84 ± 5.78%; p = 0.049) and post-shock session (WT: 25.22 ± 2.90%; LOAD2: 51.69 ± 6.91%; p = 0.005), as well as during testing in the familiar context (WT: 32.59 ± 3.86%; LOAD2: 56.55 ± 7.02%; p = 0.011), suggesting elevated baseline fear or anxiety in LOAD2 mice rather than enhanced learning. However, at 24 months, there was no main effect of genotype on freezing behaviour (F(1, 26) = 0.397, p = 0.534), while there was a main effect of session (F(2.815, 73.19) = 51.07, p < 0.001) with no session × genotype interaction (F(2.815, 73.19) = 0.435, p = 0.716) (WT: n = 11, LOAD2: n = 17; Figure 4(c)). Within-group pairwise comparisons across sessions showed that WT and LOAD2 mice of both 18 and 24 months displayed elevated freezing during testing in the familiar context relative to their respective pre-shock baselines (WT 18 months: p < 0.001; LOAD2 18 months: p = 0.009; WT 24 months: p < 0.001; LOAD2 24 months: p < 0.001), as well as reduced freezing in the novel context relative to the familiar context (WT 18 months: p < 0.001; LOAD2 18 months: p < 0.001; WT 24 months: p < 0.001; LOAD2 24 months: p < 0.001), indicative of intact contextual fear memory (Figure 4(b) and (c)). Similarly, all groups exhibited increased freezing upon delivery of the sound cue in the novel context relative to the novel context alone, indicative of intact cued fear memory (WT 18 months: p = 0.006; LOAD2 18 months: p = 0.009; WT 24 months: p = 0.007; LOAD2 24 months: p = 0.002). Full time-course plots of freezing behaviour across the conditioning and testing sessions are shown in Supplementary Figure 2.

Figure 4.

LOAD2 mice have intact fear memory at 18 and 24 months. (a) Schematic of the fear conditioning paradigm. On Day 1, mice underwent a tone-shock pairing in Context A, which was preceded by a 2-min baseline period and followed by a 2.5-min post-shock period of observation. Twenty-four hours later (Day 2), contextual fear memory was assessed by re-exposing mice to the familiar conditioning environment (Context A) without tone or shock. After a 2-h interval, cued fear recall was examined in a novel environment (Context B) using tone presentations only, without foot shocks. (b) and (c) Percent freezing across sessions in WT and LOAD2 mice aged for either 18 or 24 months. At 18 months (b), LOAD2 mice exhibited higher freezing than WT during baseline, post-shock, and familiar context sessions, consistent with elevated baseline fear or anxiety rather than enhanced learning (WT: n = 10 (male 5, female 5); LOAD2: n = 9 (male 4, female 5)). However, both genotypes exhibited increased freezing relative to their respective baselines when re-exposed to the familiar context or sound cue, consistent with intact fear learning. For 24-month-old mice (c), freezing levels varied across sessions consistent with intact fear memory, but did not differ by genotype (WT: n = 11 (male 6, female 5); LOAD2: n = 17 (male 9, female 8)). (d) Contextual fear memory quantified as change in freezing (Δ% freezing), calculated as freezing in the familiar context on Day 2 minus baseline freezing on Day 1. There were no effects of age or genotype on levels of contextual fear memory. (e) Cued fear memory quantified as Δ% freezing, calculated as freezing in the novel context + sound cue minus freezing in the novel context alone. There were no effects of age or genotype on levels of cued fear memory. BL, baseline; PS, post shock; FC, familiar context; NC, novel context; NC + SC, novel context plus sound cue. Full time-course plots of freezing behaviour across the conditioning and testing sessions are shown in Supplementary Figure 2. All data are presented as mean ± SEM. Statistical analyses are detailed in the main text. Where individual data points are shown, filled symbols are males and open symbols are females. *p < 0.05; **p < 0.01; ***p < 0.001.

Since baseline freezing was not comparable across all groups, between-group comparisons were made by first calculating the change in freezing from baseline (Δ% freezing) for each mouse, which provides a more reliable index of fear memory expression by controlling for baseline differences (Jacobs et al., 2010). For contextual fear memory, Δ% freezing was calculated as freezing in the familiar context on Day 2 minus baseline freezing on Day 1, while Δ% freezing for cued fear memory was calculated as freezing in the novel context + sound cue minus freezing in the novel context alone. Two-way ANOVA analyses of these baseline-adjusted scores showed no main effects of age (contextual: F(1, 43) = 1.182, p = 0.283; cued: F(1, 43) = 3.223, p = 0.080) or genotype (contextual: F(1, 43) = 0.110, p = 0.741; cued: F(1, 43) = 0.118, p = 0.733), and no age × genotype interactions (contextual: F(1, 43) = 3.117, p = 0.085; cued: F(1, 43) = 0.155, p = 0.695), suggesting similarly preserved levels of contextual and cued fear memory in both WT and LOAD2 mice aged for either 18 or 24 months (Figure 4(d) and (e)). Session-wide raw data and the full set of within- and between-group comparisons are provided in Supplementary Data 3.

Discussion

Overall, our study provides an assessment of key, disease-relevant, behavioural phenotypes of the LOAD2 mouse model during ageing. Specifically, we examined locomotor activity, anxiety, spatial working memory, novelty recognition memory, and contextual/cued fear memory. Our findings reveal that LOAD2 mice exhibit an early onset of locomotor and affective impairments, both of which occur without overt genotype-dependent cognitive deficits and may therefore reflect symptoms observed in the prodromal stages of AD. Relative to WT mice, these locomotor and anxiety phenotypes were evident at 18 months but were no longer detectable in the cohorts aged to 24 months, likely due to converging age-related changes in the WT controls, suggesting that LOAD2 mice may undergo a form of behavioural ‘accelerated ageing’.

Motor dysfunction and anxiety are increasingly recognised as early indicators of AD, often preceding cognitive deficits (Ogawa et al., 2018). As such, reduced locomotor activity and enhanced anxiety in LOAD2 mice at 18 months may reflect faster neurodegenerative changes or disruptions in basal ganglia and prefrontal circuits, which regulate both voluntary movement and anxiety (Andrade-Guerrero et al., 2024; Mendez, 2021). Indeed, it is notable that both contextual and cued fear memory, when tested 24 h after conditioning, remained preserved across ageing in both LOAD2 and WT mice in our experiments. This finding, which is consistent with a previous study showing preserved fear memory in aged C57BL/6 mice (Gould and Feiro, 2005), further supports the idea that the locomotor and anxiety deficits in 18-month-old LOAD2 mice may reflect an early onset of age-related changes, rather than a more generalised pathology. In addition, the results of our fear conditioning experiments provide further support for the affective phenotype of LOAD2 mice. Specifically, LOAD2 mice exhibited elevated freezing behaviour at 18 months, which was likely reflective of heightened anxiety rather than enhanced memory formation since the elevated freezing was observed in the baseline session even before shock exposure. Interestingly, however, this exaggerated freezing response was not present in LOAD2 mice at 24 months. The reason for this difference in baseline freezing between 18- and 24-month-old LOAD2 mice is unknown, but may be due to the influence of other age-related changes on baseline freezing that become more prominent with advanced ageing, such as impaired vision or olfaction (Patel and Larson, 2009; Wong and Brown, 2007). We also noted that, during the conditioning sessions, both WT and LOAD2 mice aged 24 months exhibited more pronounced post-shock freezing responses than those aged to 18 months, relative to their respective baseline levels of freezing (Supplementary Figure 2). This observation is consistent with a previous report of elevated freezing behaviour after shock exposure in aged C57BL/6 mice (Gould and Feiro, 2005).

In addition to fear conditioning, LOAD2 mice did not differ from age-matched WT controls in other cognitive assays. Spontaneous alternation in the Y-maze, which assesses spatial working memory, was similarly above chance in LOAD2 and WT mice at both 18 and 24 months, indicating preserved performance across ageing in a short-term working memory retention, continuous updating task. Furthermore, both LOAD2 and WT mice aged to either 18 or 24 months performed equally poorly in the NOR and NLR tasks, indicative of a similar age-dependent deficit in novelty recognition memory for both genotypes. Indeed, other studies have shown that spontaneous alternation can remain robustly preserved in aged mice, including in an AD model (Biundo et al., 2016), while NOR performance is known to decline with ageing, due in part to reduced exploration and altered novelty seeking (Traschütz et al., 2018). Thus, while our results suggest that LOAD2 mice do not undergo overt cognitive decline beyond that of normal ageing, we cannot rule out the possibility of more subtle cognitive deficits in this model given the sensitivity of the assays employed here together with the confounding effects of age on these behavioural measures.

While genotype-dependent cognitive deficits were not evident in this study, further work is needed to examine other domains of memory and executive function in LOAD2 mice, as well as the age at which any cognitive deficits may emerge. For example, studies that include earlier ages would help define whether and when LOAD2 mice begin to diverge from WT mice in recognition memory impairment, and would better establish any shift in the timing of cognitive decline in this model. Also, it will be important to determine whether the genetic risk factors present in the LOAD2 model are, by themselves, sufficient to induce any detrimental effects on cognitive function, or whether additional environmental factors such as diet, stress, and sleep disturbance may be required to trigger cognitive symptoms. Indeed, it was shown that under environmental stressors such as high-fat diet, LOAD2 mice do exhibit learning and memory impairments, alongside broader disease-relevant biological changes, including sex-dependent neuronal loss, increased insoluble Aβ42, elevated plasma neurofilament light chain, and alterations in synaptic and lipid metabolism pathways, despite the absence of overt amyloid plaque pathology (Kotredes et al., 2024). These findings support the notion that genetic risk factors alone may be insufficient to produce cognitive dysfunction in this model, and they motivate future studies that explicitly integrate environmental factors when evaluating cognition in models of LOAD.

The present study has several limitations. First, non-littermate C57BL/6J mice were used as WT controls, which may increase the variability between groups due to potential differences in background genetics or maternal care. Although littermate controls can reduce variability related to background genetics and early life environment, generating sufficient numbers of homozygous and WT littermate animals is challenging for a model carrying multiple engineered alleles. We therefore chose to maintain separate homozygous LOAD2 and WT breeding colonies to avoid the large number of unused animals that would otherwise result from a triple heterozygous breeding strategy. Second, we used independent cohorts at 18 and 24 months rather than longitudinally tracking the same animals across ages. A longitudinal design can be advantageous for defining within animal trajectories, particularly because ageing-related changes can be non-linear and can progress at different rates across physical, affective, and cognitive domains (Yanai and Endo, 2021). However, repeated behavioural testing and handling can, by themselves, alter exploratory behaviour and emotionality, and exposure to aversive tasks such as fear conditioning can introduce stress-related carryover effects that confound later time points (Cnops et al., 2022; McIlwain et al., 2001). We therefore prioritised independent cohorts to minimise training history and repeated stress exposure as confounds in our age comparisons. Third, we did not include direct assessments of AD-relevant brain pathology, limiting our ability to link behavioural phenotypes to underlying neuropathological changes. Prior work provides important context for potential pathological substrates in this model, particularly under dietary stress (Kotredes et al., 2024). However, direct linking of behaviour to pathology ideally requires matched postmortem endpoints within the same cohorts, and this should be considered for future studies. Finally, behavioural outcomes may not capture early, pre-pathological alterations, as synaptic dysfunction can precede overt pathology and cognitive impairment. Accordingly, future work should incorporate physiological assessments of synaptic function to determine whether LOAD2 mice exhibit early circuit-level deficits that are not detectable in the behavioural assays used here.

In summary, we show that, at 18 months of age, the LOAD2 mouse model exhibits reduced locomotor activity and elevated anxiety compared to WT mice, which resemble non-cognitive symptoms observed in the early stages of AD. These findings support the LOAD2 model as a valuable tool for investigating the early changes in behaviour and neural circuits associated with LOAD, and highlight the need for future studies targeting non-cognitive symptoms in the prodromal disease phase.

Supplemental Material

sj-docx-1-bna-10.1177_23982128261430392 – Supplemental material for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits

Supplemental material, sj-docx-1-bna-10.1177_23982128261430392 for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits by Shinwon Kang, Ashish Kadia, Junhui Wang, Rachael Ingram, Patrick Tidball, Fuzi Jin, Aram Abbasian, John Georgiou and Graham L. Collingridge in Brain and Neuroscience Advances

Supplemental Material

sj-docx-2-bna-10.1177_23982128261430392 – Supplemental material for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits

Supplemental material, sj-docx-2-bna-10.1177_23982128261430392 for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits by Shinwon Kang, Ashish Kadia, Junhui Wang, Rachael Ingram, Patrick Tidball, Fuzi Jin, Aram Abbasian, John Georgiou and Graham L. Collingridge in Brain and Neuroscience Advances

Supplemental Material

sj-pdf-7-bna-10.1177_23982128261430392 – Supplemental material for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits

Supplemental material, sj-pdf-7-bna-10.1177_23982128261430392 for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits by Shinwon Kang, Ashish Kadia, Junhui Wang, Rachael Ingram, Patrick Tidball, Fuzi Jin, Aram Abbasian, John Georgiou and Graham L. Collingridge in Brain and Neuroscience Advances

Supplemental Material

sj-xlsx-3-bna-10.1177_23982128261430392 – Supplemental material for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits

Supplemental material, sj-xlsx-3-bna-10.1177_23982128261430392 for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits by Shinwon Kang, Ashish Kadia, Junhui Wang, Rachael Ingram, Patrick Tidball, Fuzi Jin, Aram Abbasian, John Georgiou and Graham L. Collingridge in Brain and Neuroscience Advances

Supplemental Material

sj-xlsx-4-bna-10.1177_23982128261430392 – Supplemental material for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits

Supplemental material, sj-xlsx-4-bna-10.1177_23982128261430392 for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits by Shinwon Kang, Ashish Kadia, Junhui Wang, Rachael Ingram, Patrick Tidball, Fuzi Jin, Aram Abbasian, John Georgiou and Graham L. Collingridge in Brain and Neuroscience Advances

Supplemental Material

sj-xlsx-5-bna-10.1177_23982128261430392 – Supplemental material for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits

Supplemental material, sj-xlsx-5-bna-10.1177_23982128261430392 for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits by Shinwon Kang, Ashish Kadia, Junhui Wang, Rachael Ingram, Patrick Tidball, Fuzi Jin, Aram Abbasian, John Georgiou and Graham L. Collingridge in Brain and Neuroscience Advances

Supplemental Material

sj-xlsx-6-bna-10.1177_23982128261430392 – Supplemental material for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits

Supplemental material, sj-xlsx-6-bna-10.1177_23982128261430392 for The LOAD2 mouse model of late-onset alzheimer’s disease exhibits an accelerated onset of locomotor and anxiety deficits by Shinwon Kang, Ashish Kadia, Junhui Wang, Rachael Ingram, Patrick Tidball, Fuzi Jin, Aram Abbasian, John Georgiou and Graham L. Collingridge in Brain and Neuroscience Advances

Footnotes

ORCID iDs

Shinwon Kang

Patrick Tidball

John Georgiou

Author contributions

S.K., A.A., J.G., and G.L.C. conceptualised and designed the study. S.K., A.K., J.W., and F.J. conducted all experiments. All data were analysed by S.K. The manuscript was written and revised by S.K., R.I., P.T., J.G., and G.L.C. All authors reviewed and agreed on the final version of the manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors acknowledge financial support from the Canadian Institutes of Health Research (CIHR) Foundation Grant #154276 (GLC). The research was supported by a Krembil Foundation Grant to GLC and JG. GLC also holds the Krembil Family Chair in Alzheimer’s Research. Additional support was provided by the Dani Reiss Family Foundation through the Neurodegeneration and Ageing Research Program.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supplemental material

Supplemental material for this article is available online.

References

Andrade-Guerrero

Martínez-Orozco

Villegas-Rojas

, et al. (2024) Alzheimer’s disease: Understanding motor impairments. Brain Sciences 14(11): 1054.

Antunes

Biala

(2012) The novel object recognition memory: Neurobiology, test procedure, and its modifications. Cognitive Processing 13(2): 93–110.

Ayer

Wojta

Ramos

, et al. (2019) Frequency of the TREM2 R47H variant in various neurodegenerative disorders. Alzheimer Disease and Associated Disorders 33(4): 327–330.

Baumgart

Snyder

Carrillo

, et al. (2015) Summary of the evidence on modifiable risk factors for cognitive decline and dementia: A population-based perspective. Alzheimer’s & Dementia 11(6): 718–726.

Bekris

C-E

Bird

, et al. (2010) Genetics of Alzheimer disease. Journal of Geriatric Psychiatry and Neurology 23(4): 213–227.

Biundo

Ishiwari

Del Prete

, et al. (2016) Deletion of the γ-secretase subunits Aph1B/C impairs memory and worsens the deficits of knock-in mice modeling the Alzheimer-like familial Danish dementia. Oncotarget 7(11): 11923–11944.

Cheng-Hathaway

Reed-Geaghan

Jay

, et al. (2018) The Trem2 R47H variant confers loss-of-function-like phenotypes in Alzheimer’s disease. Molecular Neurodegeneration 13(1): 29.

Cnops

Iyer

Parathy

, et al. (2022) Test, rinse, repeat: A review of carryover effects in rodent behavioral assays. Neuroscience & Biobehavioral Reviews 135: 104560.

Ennaceur

Delacour

(1988) A new one-trial test for neurobiological studies of memory in rats. 1: Behavioral data. Behavioural Brain Research 31(1): 47–59.

10.

Foley

Hewes

Garceau

, et al. (2022) The APOEε3/ε4 genotype drives distinct gene signatures in the cortex of young mice. Frontiers in Aging Neuroscience 14: 838436.

11.

Fox

(ed.) (2007) The Mouse in Biomedical Research (2nd edn). Amsterdam: Academic Press.

12.

Gould

Feiro

(2005) Age-related deficits in the retention of memories for cued fear conditioning are reversed by galantamine treatment. Behavioural Brain Research 165(2): 160–171.

13.

Iwasaki

Sakaguchi

Ikegaya

(2015) Brief fear preexposure facilitates subsequent fear conditioning. Neuroscience Research 95: 66–73.

14.

Jacobs

Cushman

Fanselow

(2010) The accurate measurement of fear memory in Pavlovian conditioning: Resolving the baseline issue. Journal of Neuroscience Methods 190(2): 235–239.

15.

Kotredes

Pandey

Persohn

, et al. (2024) Characterizing molecular and synaptic signatures in mouse models of late-onset Alzheimer’s disease independent of amyloid and tau pathology. Alzheimer’s & Dementia 20(6): 4126–4146.

16.

Kulkarni

Singh

Bishnoi

(2007) Elevated zero maze: A paradigm to evaluate antianxiety effects of drugs. Methods and Findings in Experimental and Clinical Pharmacology 29(5): 343–348.

17.

Laganà

Bruno

Altomari

, et al. (2022) Neuropsychiatric or Behavioral and Psychological Symptoms of Dementia (BPSD): Focus on prevalence and natural history in Alzheimer’s disease and frontotemporal dementia. Frontiers in Neurology 13: 832199.

18.

Lou

Zhu

, et al. (2025) Sex-specific behavioral changes in middle-aged mice. Scientific Reports 15(1): 21992.

19.

Lumsden

Mulugeta

Zhou

, et al. (2020) Apolipoprotein E (APOE) genotype-associated disease risks: A phenome-wide, registry-based, case-control study utilising the UK Biobank. eBioMedicine 59: 102954.

20.

McIlwain

Merriweather

Yuva-Paylor

, et al. (2001) The use of behavioral test batteries: Effects of training history. Physiology & Behavior 73(5): 705–717.

21.

Mendez

(2021) The relationship between anxiety and Alzheimer’s disease. Journal of Alzheimer’s Disease Reports 5(1): 171–177.

22.

Nolte

Yan

(2019) Anxiety and task performance changes in an aging mouse model. Biochemical and Biophysical Research Communications 514(1): 246–251.

23.

Ogawa

Kaneko

Sato

, et al. (2018) Sarcopenia and muscle functions at various stages of Alzheimer disease. Frontiers in Neurology 9: 710.

24.

Patel

Larson

(2009) Impaired olfactory discrimination learning and decreased olfactory sensitivity in aged C57Bl/6 mice. Neurobiology of Aging 30(5): 829–837.

25.

Peng

Wang

Luo

, et al. (2018) A novel and quick PCR-based method to genotype mice with a leptin receptor mutation (db/db mice). Acta Pharmacologica Sinica 39(1): 117–123.

26.

Sloane

Zimmerman

Suchindran

, et al. (2002) The public health impact of Alzheimer’s disease, 2000-2050: Potential implication of treatment advances. Annual Review of Public Health 23: 213–231.

27.

Thaliath

Pillai

(2025) Non-cognitive symptoms in Alzheimer’s disease and their likely impact on patient outcomes. A scoping review. Current Treatment Options in Neurology 27(1): 41.

28.

Traschütz

Kummer

Schwartz

, et al. (2018) Variability and temporal dynamics of novel object recognition in aging male C57BL/6 mice. Behavioural Processes 157: 711–716.

29.

Trent

Abdullah

Parwana

, et al. (2025) Fear conditioning: Insights into learning, memory and extinction and its relevance to clinical disorders. Progress in Neuro-Psychopharmacology and Biological Psychiatry 138: 111310.

30.

Truett

Heeger

Mynatt

, et al. (2000) Preparation of PCR-quality mouse genomic DNA with Hot Sodium Hydroxide and Tris (HotSHOT). Biotechniques 29 (1): 52–54.

31.

Ward

Crean

Mercaldi

, et al. (2012) Prevalence of Apolipoprotein E4 genotype and homozygotes (APOE e4/4) among patients diagnosed with Alzheimer’s disease: A systematic review and meta-analysis. Neuroepidemiology 38(1): 1–17.

32.

Wong

Brown

(2007) Age-related changes in visual acuity, learning and memory in C57BL/6J and DBA/2J mice. Neurobiology of Aging 28(10): 1577–1593.

33.

Xiang

Piers

Wefers

, et al. (2018) The Trem2 R47H Alzheimer’s risk variant impairs splicing and reduces Trem2 mRNA and protein in mice but not in humans. Molecular Neurodegeneration 13(1): 49.

34.

Ran

Fromholt

, et al. (2015) Murine Aβ over-production produces diffuse and compact Alzheimer-type amyloid deposits. Acta Neuropathologica Communications 3(1): 72.

35.

Yanai

Endo

(2021) Functional aging in male C57BL/6J mice across the life-span: A systematic behavioral analysis of motor, emotional, and memory function to define an aging phenotype. Frontiers in Aging Neuroscience 13: 697621.

36.

(2001) An efficient procedure for genotyping single nucleotide polymorphisms. Nucleic Acids Research 29(17): e88.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB

0.20 MB

0.26 MB

0.04 MB

0.02 MB