Caloric restriction improves spatial learning deficits in tau mice

Abstract

Background:

Caloric restriction (CR) has been recognized for its benefits in delaying age-related diseases and extending lifespan. While its effects on amyloid pathology in Alzheimer's disease (AD) mouse models are well-documented, its effects on tauopathy, another hallmark of AD, are less explored.

Objective:

To assess the impact of a short-term 30% CR regimen on age-dependent spatial learning deficits and pathological features in a tauopathy mouse model.

Methods:

We subjected male PS19 tau P301S (hereafter PS19) and age-matched wildtype mice from two age cohorts (4.5 and 7.5 months old) to a 6-week 30% CR regimen. Spatial learning performance was assessed using the Barnes Maze test. Tau pathology, neuroinflammation, hippocampal cell proliferation, and neurogenesis were evaluated in the older cohort by immunohistochemical staining and RT-qPCR.

Results:

CR mitigated age-dependent spatial learning deficits in PS19 mice but exhibited limited effects on tau pathology and the associated neuroinflammation. Additionally, we found a decrease in hippocampal cell proliferation, predominantly of Iba1+ cells.

Conclusion:

Our findings reinforce the cognitive benefits conferred by CR despite its limited modulation of disease pathology. Given the pivotal role of microglia in tau-driven pathology, the observed reduction in Iba1+ cells under CR suggests potential therapeutic implications, particularly if CR would be introduced early in disease progression.

Introduction

Caloric restriction (CR), a dietary regimen that reduces calorie intake without malnutrition, has been widely studied for its potential benefits in enhancing health and longevity. Research spanning a spectrum of organisms, from yeast and worms to flies, rodents, and non-human primates, shows that CR can delay the onset of age-related diseases and extend lifespan, impacting an array of biological processes such as metabolism, oxidative stress, and inflammation [1-3]. The beneficial effects of CR extend across a range of tissues, including the heart, liver, muscle, fat, and, importantly, the brain [4]. Several studies have indicated that CR can modulate neuroplasticity, stimulate neuroprotective pathways, enhance cognitive performance, and potentially alleviate age-related cognitive decline [5]. CR might protect the brain’s health and function by mitigating the age-related accumulation of oxidative damage to DNA, protein, and lipids: a common pathological feature of many neurodegenerative diseases [6].

Alzheimer's disease (AD), one of the most prevalent neurodegenerative disorders and the fifth leading cause of death, is characterized by a gradual deterioration of cognitive function, memory, and behavioral changes. Alongside severe neuronal loss, one of the main features of AD is the abnormal accumulation of misfolded amyloid-beta (Aβ) and hyperphosphorylated tau protein, forming senile plaques and neurofibrillary tangles, respectively, in the brain [7]. Considering the significant toll of AD and the current lack of effective treatment, research has been directed toward exploring the potential therapeutic benefits of CR using various transgenic mouse models of AD. Several studies employing AD mouse models of amyloid-only pathology have reported promising results following both short-term and long-term CR regimens of 30-40%, including reduced Aβ plaque load and amelioration of associated pathological features such as gliosis, synaptic dysfunction, and cognitive deficits. Long-term CR initiated at 3 months of age completely prevented amyloid plaque formation in 12-month-old female transgenic Tg2576 mice [8]. Another study using the same mouse model also reported a significant reduction in hippocampal Aβ burden in 15-month-old females [9]. In a study with male APP/PS1 mice, a prolonged 68-week CR regimen initiated as early as 4 weeks of age resulted not only in a significant reduction of Aβ plaque load in the hippocampus but also attenuated microgliosis and improved cognitive performance, as demonstrated in the Morris Water Maze test [10]. Furthermore, another study found that shorter CR regimens (6 and 15 weeks) also reduced Aβ plaque count and astrogliosis in male APP and APP/PS1 mice [11]. CR continued to exhibit beneficial effects even when it was initiated at a later stage in AD pathology in APP/PS1 mice (13-14 months). Specifically, compared to age-matched ad libitum-fed mice, 18 weeks of CR reduced Aβ plaque load in both the neocortex and the hippocampus by one third [12].

While the impact of CR on amyloid pathology in AD mouse models has been well investigated, its effect on tauopathy, a key pathological feature of many neurodegenerative disorders including AD, progressive supranuclear palsy, corticobasal degeneration, and frontotemporal dementia, remains less understood. One study in the triple transgenic mouse model 3xTgAD, that exhibits both amyloid plaques and tau tangles, showed that a 40% CR regimen initiated at 3 months of age ameliorated age-related behavioral deficits by 17 months [13]. In addition to reductions in hippocampal levels of Aβ1-40 and Aβ1-42, phospho-tau levels were also significantly decreased. Another study in a conditional double knockout mouse model of presenilin-1 and presenilin-2 (PS1/PS2), which displays tau hyperphosphorylation, brain atrophy, synaptic dysfunction, and cognitive deficits, showed that 4 months of CR ameliorated cognitive deficits and attenuated ventricular enlargement, hippocampal atrophy, astrogliosis, and tau hyperphosphorylation [14]. Furthermore, in Tg4510 tau P301L mice, CR initiated at 3 months partially rescued memory deficits by the age of 6 months, substantially improving short-term memory in novel object recognition and associative memory in contextual fear conditioning tests [15]. Despite these promising behavioral improvements, the pathophysiological markers of tauopathy including astrocyte/microglia activation and tau/phospho-tau expression, remained unchanged under CR. The cognitive improvements in these tauopathy mouse models highlight the importance of studying CR's effects in the context of tauopathy, particularly since tau burden correlates more strongly with neurodegeneration and cognitive impairment severity than Aβ deposition [16-19].

Given the limited research on the effects of CR on tauopathy mouse models, our study aimed to explore the impact of a short-term 30% CR regimen on spatial learning deficits and the pathological features associated with the PS19 mouse model. Our selection of the 6-week duration was based on prior studies showing that short-term CR can effectively reduce amyloid burden and associated pathology in AD mouse models. This short-term approach may offer practical insights into the therapeutic potential of CR for tauopathies, including its ability to modulate cognitive deficits and pathological markers within a relatively short time frame.

Materials and methods

Mouse strains

All experiments were reviewed and approved by the Mayo Clinic Institutional Animal Care and Use Committee. MAPT^P301SPS19 (PS19) mice, which carry the P301S human MAPT gene encoding T34-tau isoform (1N4R) driven by the mouse prion protein, Prnp, promoter [20] were purchased from the Jackson Laboratory (stock no. 008169) and bred to C57BL/6 for at least three generations. Littermate wild-type (WT) mice from the same genetic background were used as controls. Prior to enrollment in the experiments described below, mice were co-housed in a 12 h:12 h light:dark cycle environment and had ad libitum access to water and food (LabDiet #5053).

Calorie restriction (CR) and ad libitum (AL) feeding regimens

Experimental groups consisted of two independent cohorts: 4.5-month-old (younger cohort, hereafter young) and 7.5-month-old (older cohort, hereafter old) PS19 and age-matched WT male mice. Before initiating calorie restriction, all mice were singly housed for a week to acclimatize and received approximately 5g of powdered standard laboratory chow (EURodent Diet 14%, 5LF2, LabDiet, St. Louis, MO, USA) daily, based on the dietary recommendations for adult mice. This initial food intake was adjusted to the body weight of individuals and quantified to determine the food amount required to maintain a stable weight for one week. Thereafter, mice were randomly assigned to either AL or CR groups. AL mice received food amounts tailored to their body weights, while CR groups received a 30% initial reduction in calorie intake. Calorie restriction was adjusted weekly to maintain body weights between 19-21 g. All mice were fed daily at zeitgeber time (ZT) +8, and their body weights were recorded twice weekly, and were maintained on their feeding regimens for approximately 6 weeks.

Barnes maze

Barnes maze testing was performed as previously described with some modifications [21]. Briefly, the apparatus featured a 120 cm diameter circular platform, elevated 75 cm from the floor, with 20 evenly spaced 5 cm circular holes around its perimeter. One hole led to an escape box (a darkened chamber) while the others were false holes. The maze was illuminated by four 75 W overhead lights and four distinct shapes were placed around it as visual cues. After each trial the setup was cleaned, aired, and the platform was randomly rotated, though the location of the escape box remained consistent. All mice were acclimated to the testing room for 1 h before sessions. On the first day, mice were habituated by placing them in the center of the maze under a clear glass beaker for 30 s. They were then gently guided by slowly moving the beaker, first to a false hole to explore, and then to the adjacent hole, which was the escape hole. If they did not enter the escape hole independently within 1 min, they were gently nudged with the beaker to enter and remained there for 1 min before being returned to their cage. After a 30-min rest, acquisition training began. At the start of each trial, the mouse was placed in the center of the maze, covered by a bucket. After 5 seconds, the bucket was lifted, allowing the mouse to explore the maze to find the escape box. If not located within 1 min, the mouse was guided towards it. If the mouse was reluctant to enter, it was gently place inside. The hole was then covered for another 1 min before the mouse was returned to its home cage. Mice underwent 3 trials daily for 4 days, with a 30-min inter-trial interval. Due to a restricted ~3 h testing window available due to the feeding schedule and approaching dark phase, individual trial times were limited to 1 minute. Consequently, the latency to escape the maze (total latency) could not be adequately assessed, as the timeframe might have been insufficient for the majority of mice to enter the escape hole, therefore we used the latency to locate the escape hole (primary sniff latency) to measure learning performance. Distance and velocity were recorded for locomotor activity. Trials were recorded from above (Panasonic WV-CP294) and analyzed using TopScan Version 3.00 (CleverSys Inc, Reston, VA, USA) video tracking software.

Quantitative RT–PCR

RNA extraction, cDNA synthesis and RT–qPCR analysis were performed on hippocampi samples from mouse brains as previously described [22]. Primers used to amplify Gfap, S100b, Pai1, Il1b, p21, p16, p19 and Tbp were as previously described [22-24]. The following additional primer was used: human MAPT forward 5’-TGCTTTTACTGACCATGCGA-3’, reverse −5’-AAGACCAAGAGGGTGACACG-3’. Expression was normalized first to Tbp.

Immunofluorescence staining

Mice were transcardially perfused with ice-cold DPBS (Gibco #14190144) until fluid run-off was clear. Brains were stored in 4% PFA overnight at 4 °C and then cryoprotected by incubating in a 30% sucrose solution for 48 h at 4 °C. Samples were sectioned into 30-μM-thick coronal sections and stored in antifreeze solution (300 g sucrose, 300 ml ethylene glycol, 500 ml PBS) at −20 °C. Phospho-tau S202/S205 (1:500, Thermo Fisher #MN1020), Gfap (1:1000, Novus #3NBP1-05198), Iba1 (1:500, Wako #019-19741), PSA-NCAM (1:250, Thermo Fisher #14-9118-82) immunohistochemistry staining was performed on free-floating sections from bregma −1.6 mm to −2.5 mm as previously described [24]. For cellular proliferation assay mice were injected with EdU (75 mg/kg i.p) once daily for 5 consecutive days before euthanasia. Imaging of EdU-positive cells was performed following the manufacturer’s instructions (Invitrogen Click-iT EdU Alexa Fluor 488 Imaging Kit, C10337). Images were acquired on a Zeiss LSM 780 confocal system using multi-track configuration (Fig. 2, Fig. 3 and Supplementary fig. 3) and using an Olympus BX53 Fluorescence microscope and DP80 digital camera (Fig. 4). Stereological quantification of cellular proliferation and neurogenesis in the subgranular zone (SGZ) and granule cell layer was conducted as previously described [25]. Briefly, EdU+ cells, PSA-NCAM+ cells and double-positive cells were counted to determine the total cell number. For each analyzed dentate gyrus (DG), its area was measured from the middle z-plane by manually outlining its perimeter and capturing the area using the area measurement feature provided by the ImageJ software. The volume of the DG section was calculated by multiplying its area by the thickness. The cell count was then normalized to the calculated volume to obtain the total cell density in the DG per mm³. For the co-localization quantification of EdU;Iba1 and EdU;Gfap, we followed the methodology described above. Additionally, to obtain the hippocampal area excluding DG, we measured the hippocampal area and then subtracted the area of the DG.

Figure 2.

Calorie restriction does not affect tau pathology or neuroinflammation in 9-month-old PS19 mice. (A) Immunostaining and quantification of phosphorylated tau using AT-8 antibody (S202/T205) in the dentate gyrus (DG) of PS19 mice; n=8/group. (B) RT-qPCR analysis of the expression of astrocyte, pro-inflammatory, and senescence markers in the hippocampi of PS19 and WT mice under AL or CR diets; n=6-11/group. (C) Representative photomicrographs of Gfap (upper panel) and Iba1 (lower panel) staining in the DG of PS19 and WT mice under AL or CR. Cell nuclei were labeled with DAPI (blue). Scale bars = 100 μm. Data are mean ± s.e.m and analyzed with unpaired two-sided t-test with Welch’s correction (A) and one-way ANOVA with Tukey’s multiple comparisons test (B).

Figure 3.

Calorie restriction attenuates the increased cell proliferation in the dentate gyrus of 9-month-old PS19 mice. (A) Representative photomicrographs and quantification of EdU-positivity in the dentate gyrus (DG) of PS19 and age-matched WT mice under AL or CR diets. Mice were pulsed with EdU for five consecutive days prior to termination; n=4-8/group. (B) Representative photomicrographs and quantification of PSA-NCAM-positive cells in the DG of PS19 and age-matched WT mice under AL or CR diets; n=4-7/group. (C) Representative photomicrographs and quantification of double immunostaining for EdU and PSA-NCAM in PS19 mice under AL or CR diets. Note the minimal co-localization of EdU with PSA-NCAM+ cells in the PS19 AL group (white arrows pointing to Edu+;PSA-NCAM+ cells located in the subgranular zone); n=5/group. Cell nuclei were labeled with DAPI (blue). Scale bars = 100 μm. All images, except for the far right image in panel C (captured at 40x magnification), were taken at 20x magnification with consistent z-stack configurations across all samples. Data are mean ± s.e.m and analyzed by one-way ANOVA with Tukey’s multiple comparisons test (A-B) and with unpaired two-sided t-test with Welch’s correction (C). *p<0.05, **p<0.01, ****p<0.0001.

Figure 4.

Calorie restriction decreases Iba1+ cell proliferation in the hippocampus of 9-month-old PS19 mice. (A-B) Representative photomicrographs and quantification of co-localization of EdU (green) with Iba1 (red) and Gfap (blue) staining in the dentate gyrus (DG, A) and in hippocampal area excluding DG, with the CA3 region depicted in (B) at bregma −1.8mm for PS19 AL and −2.2 mm for PS19 CR; n=3/group. Note the predominant co-localization of EdU+ cells with Iba1 (white arrows, insets); co-localization of EdU+ cells with Gfap was rare (no co-localization depicted). Cell nuclei were labeled with DAPI (gray). Scale bars =100 μm, insets =20 μm. Data are mean ± s.e.m and analyzed by unpaired two-sided t-test with Welch’s correction. *p<0.05, **p<0.01.

Statistical analysis

All statistical analysis was performed in GraphPad Prism (version 9, GraphPad Software Inc., San Diego, CA, USA). A student’s two-tailed unpaired t-test (Fig. 2A, Fig. 3C, Fig. 4 and Supplementary Fig. 1A-B) was used to determine statistical significance between two groups. A one-way ANOVA with Tukey’s post-hoc for multiple comparisons test was performed to assess the statistical significance for more than two groups (Fig. 1I-J, Fig. 2B, Fig. 3A-B). Two-way ANOVA with Tukey’s (or Sidak’s) multiple comparisons test was used in all other figures. Statistical significance was defined as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Investigators were blinded during experiments and when evaluating outcomes, with occasional exceptions where blinding was unfeasible.

Figure 1.

Calorie restriction mitigates age-dependent decline in spatial learning and improves locomotor performance in 9-month-old PS19 mice. (A-B) Schematic of the experimental timeline for calorie-restricted (CR) or unrestricted (AL) diets initiated at 4.5 months (young cohort) and 7.5 months (old cohort) in PS19 and age-matched WT mice. Barnes maze test was conducted during week 5. (C-H) Behavioral parameters measured include primary sniff latency (C, F), total distance travelled (D, G) and average speed (E, H); n=8-11/group. (I) Comparison of total distances travelled on day 4 by individual 9-month-old mice that sniffed (black circles) or did not sniff (grey circles) at the escape hole. (J) Distance traveled to the first sniff on day 4 in 9-month-old mice that successfully located the escape hole; n=3-11/group. Data are mean ± s.e.m and analyzed by two-way ANOVA with Tukey’s multiple comparisons test (C-H) and one-way ANOVA with Tukey’s multiple comparisons test (I, J). Symbols: # indicates PS19 AL vs WT AL and * PS19 CR vs PS19 AL comparison. *p<0.05, **p<0.01, ***p<0.001.

Results

Calorie restriction mitigates age-dependent decline in spatial learning and improves locomotor performance in 9-month-old PS19 mice

To determine whether caloric restriction impacted disease-dependent pathology in the PS19 mouse model of neurodegeneration, we subjected 4.5-month-old (young) and 7.5-month-old (old) PS19 and age-matched WT male mice to a 30% CR regimen for a duration of 6 weeks (Fig. 1A). The control groups consisted of age-matched PS19 and WT mice that were not calorie-restricted (ad lib; AL). An array of studies using various CR paradigms, including those used in our study [26], have demonstrated that prolonged restriction does not cause overt health issues in mice [27-29]. In line with these findings, CR mice were systematically evaluated twice weekly and showed no signs of inactivity, loss of grooming, hunched posture, ruffled coat, dehydration, or diarrhea, while displaying complete food intake and normal water consumption. As previously reported [30], PS19 mice from both age groups had initial body weights lower than age-matched WT mice (Supplementary Fig. 1A,B). However, following the initial weight loss, PS19 CR mice, similar to WT CR mice, maintained stable body weights (Supplementary Fig. 1C,D). Collectively, these data indicate that our 6-weeks CR was without major health issues.

During the fifth week of CR, mice were subjected to a 4-day Barnes Maze assessment (Fig. 1B). We assessed spatial learning performance using the latency to locate the escape hole (primary sniff latency) and evaluated locomotor performance by distance and velocity measurements.

We first assessed the young cohort to test if 6-month-old PS19 mice exhibited spatial learning impairments and whether CR initiated at 4.5 months of age would mitigate these deficits. At this age, PS19 AL mice demonstrated a learning curve, as shown by a significant decrease in primary latency on day 4 relative to day 1 (p=0.02), that was comparable to the WT AL group (Fig. 1C). We did not observe a difference with CR, as the learning curves of both PS19 and WT CR mice were similar to those of AL mice (Fig. 1C).

We also evaluated the locomotor performance of PS19 mice at 6 months of age. Both PS19 and WT mice under CR displayed somewhat higher locomotor activity compared to their respective AL groups on day 1, evidenced by longer travel distances and increased moving speed, although these differences were not statistically significant (Fig. 1D, E). However, by day 4, total distance traveled and speed across all groups were comparable.

Collectively, these findings indicate that 6-month-old PS19 mice have not yet developed measurable deficits in spatial learning performance. A 6-week CR regimen initiated at 4.5 months of age does not impact spatial learning, as CR mice performed comparably to AL mice. CR appears to increase locomotor activity particularly during the early stages of Barnes maze testing, independent of genotype.

To contrast these results, we next evaluated spatial learning and locomotor performance on an independent cohort of 9-month-old PS19 mice with the same CR regimen initiated at 7.5 months of age. Unlike in the 6-month-old mice, older PS19 AL mice failed to learn the task. Their primary sniff latencies remained consistent throughout the testing period, and by day 4, the latencies were significantly longer than those of the WT AL mice (p=0.01) (Fig. 1F). In contrast to AL feeding, PS19 CR mice displayed a decrease in primary latencies across all days, and by day 4, the latencies were comparable to those of the WT AL group.

Compared to the younger cohort, 9-month-old PS19 AL mice exhibited significantly lower locomotor activity on days 3 and 4 relative to WT AL mice (Fig. 1G, H). In contrast, the distance traveled and speed of PS19 CR mice across days 2-4 were largely comparable to WT AL mice. On day 1, a diet-related increase in locomotor activity independent of genotype was observed, consistent with observations in the 6-month-old group (Fig. 1G, H).

To account for the possibility that the reduced primary sniff latencies on day 4 in PS19 CR mice might reflect longer travel distances to find the escape hole (Fig. 1I) rather than actual learning improvements, we measured the distance each mouse that successfully located the escape hole covered before its first sniff (Fig. 1J). The PS19 CR mice that sniffed at the escape hole traveled distances similar to those of the WT mice. Although PS19 AL mice traveled shorter distances compared to all other groups, more than half failed to locate the escape hole within the 60-second trial period.

Taken together, these results suggest that a 6-week CR intervention started at 7.5 months alleviates spatial learning deficits and the decline in locomotor activity present in PS19 mice at 9 months of age.

Calorie restriction does not affect hippocampal tau pathology or the associated neuroinflammation in 9-month-old PS19 mice

A hallmark of PS19 mice is the development of hyperphosphorylated tau protein aggregates in the hippocampus from as early as six months, resulting in neurofibrillary tangles, neuronal loss, and cognitive impairments [20]. With the DG playing a key role in hippocampal cognitive processes [31], particularly in spatial learning and memory [32-35], we investigated whether the spatial learning improvements with CR corresponded with reduced tau pathology. To this end, we assessed phosphorylated tau levels in 9-month-old PS19 mice on AL or CR diets by immunohistochemical staining for AT-8 antibody (Ser202, Thr205). Quantification revealed no reduction in AT-8 staining intensity in the PS19 CR group (Fig. 2A). Additionally, hippocampal tau mRNA expression was comparable between the groups (Supplementary Fig. 2).

In addition to tau pathology, PS19 mice exhibit progressive gliosis and accumulation of senescent cells with disease progression [20, 24]. To investigate whether CR mitigates this process, RT-qPCR was performed on the hippocampi of 9-month-old mice for markers of astrocytes Gfap and S100b, the pro-inflammatory genes Pai1 and Il1b, and senescence markers p21, p16 and p19 (Fig. 2B). The results did not show a reduction in the mRNA expression levels of these markers in the PS19 CR group. Consistent with the mRNA data, visual evaluation of immunohistochemical staining for Gfap and Iba1 (a microglial marker) revealed no major differences in staining intensity between the PS19 CR and PS19 AL groups or between the WT CR and WT AL groups (Fig. 2C).

Collectively, these findings indicate that a 6-week CR regimen initiated at 7.5 months does not alleviate tau pathology or the associated neuroinflammation present in PS19 mice at 9 months of age.

Calorie restriction decreases cell proliferation in the DG of 9-month-old PS19 mice

Since the mitigation of spatial learning deficits in 9-month-old PS19 CR mice did not coincide with reductions in hyperphosphorylated tau or neuroinflammation, we investigated the impact of CR on hippocampal neurogenesis, which is thought to be integral to spatial navigation tasks [36]. Neurogenesis was assessed by immunohistochemistry staining for proliferating cells after using a 5-day ethynyl deoxyuridine (EdU) administration prior to sacrifice. Because EdU is incorporated into newly synthesized DNA, EdU positivity indicates proliferation [37]. We then examined the rates of EdU-positivity and the presence of PSA-NCAM staining, which is a marker of immature neurons [38], in the DG. We found that CR reduced both EdU-positivity and PSA-NCAM expression, irrespective of genotype (Fig. 3A, B). Interestingly, PS19 AL mice displayed a pronounced increase in EdU-positivity, with only minimal co-localization of these proliferating cells with PSA-NCAM (Fig. 3C). In contrast, PS19 CR mice displayed EdU levels similar to those of the WT AL group, and the majority of EdU-labelled cells co-localized with PSA-NCAM, despite a reduction in PSA-NCAM+ cells. Importantly, the density of EdU+;PSA-NCAM+ cells was comparable between the groups.

These observations suggest that the increased EdU-positivity in PS19 AL mice might represent either early-stage neural progenitors not yet expressing PSA-NCAM, neural progenitors that have transitioned beyond the PSA-NCAM-expressing phase, or non-neuronal cells. CR appears to reduce the proliferation of these specific cell types.

Calorie restriction reduces Iba1+ cell proliferation in the hippocampus of 9-month-old PS19 mice

Because the distribution of EdU+ cells in PS19 AL mice extended beyond the traditional neurogenic zone (the SGZ of the DG) to be detectable throughout other areas of the hippocampus (Supplementary Fig. 3), we hypothesized that the increased rates of EdU-positivity might reflect the proliferation of glial cells. To investigate this, we performed co-staining experiments for Gfap (a marker for astrocytes) and Iba1 (a marker for microglia). While few cells co-localized EdU with Gfap, we found a significant co-localization of EdU+ cells with Iba1 in both the DG and other hippocampal regions of PS19 AL mice, suggesting an increased proliferation of Iba1+ cells (Fig. 4A, B). Within the DG, most EdU+;Iba1+ cells were found in the granular cell layer (Fig. 4A, inset), with a smaller fraction also observed in the SGZ (not depicted). In the PS19 CR group, there was a significant decrease in the amount of EdU+;Iba1+ cells with co-localization accounting for less than 10% of EdU+ cells, in contrast to the ~ 25% in the PS19 AL group.

In the hippocampal area excluding the DG, the proportion of proliferating cells that co-localized with Iba1 in the PS19 AL group was nearly double that observed in the DG, representing approximately half of the total EdU+ cells (Fig. 4B). Although the PS19 CR group showed a significant reduction in EdU+;Iba1+ cells, the proportion of these cells relative to total EdU+ cells was comparable to the PS19 AL group. This suggests that while CR reduces Iba1+ cell proliferation, it might not exclusively target these cells, as it also affects the proliferation of other cell types.

The densities of EdU+;Gfap+ cells were comparable between the groups (Fig. 4A, B). However, when considering these cells as a proportion of the total EdU+ cell population, the PS19 CR group displayed a higher ratio, especially in the DG. This suggests that in the CR group, although the overall proliferation of these cells is not significantly altered, there may be a subtle shift in their differentiation pattern. A slightly larger fraction of the proliferating cells might be committing to the astrocytic lineage or to the neural lineage, given that GFAP is also expressed in a subset of neural progenitor cells, namely type 1 cells [38].

Collectively, these findings suggest that CR in PS19 mice appears to inhibit hippocampal cell proliferation leading to a decrease in Iba1+ cell proliferation.

Discussion

As the quest for an effective treatment for Alzheimer's disease continues, the potential therapeutic benefits of lifestyle modifications, such as dietary interventions, have garnered significant attention in recent years. While several studies have demonstrated the beneficial effects of CR in reducing brain amyloid deposits in several mouse models of AD, its impact on tau pathology - another hallmark of AD and related tauopathies - remains less explored. In the present study, we show that a 6-week 30% CR regimen improved spatial learning deficits in 9-month-old PS19 mice without impacting the levels of phosphorylated tau or the associated neuroinflammation in the hippocampus. Notably, CR attenuated the increased cell proliferation observed in PS19 mice, predominantly of Iba1+ cells.

The amelioration of age-associated spatial learning impairments in PS19 mice under CR, without changes in key pathological features, aligns with previous findings in Tg4510 tau P301L mice, where CR partially rescued cognitive deficits without altering phosphorylated tau levels or the associated tau pathology [15]. On the other hand, studies in other mouse models of tau hyperphosphorylation and neurodegeneration have shown that CR improved cognitive functions alongside alleviating the disease phenotype [13, 14]. These outcomes were typically linked to the early initiation of CR, before the onset of pronounced tau phosphorylation and evident neurodegeneration in these models [39-41]. In contrast, introducing CR at 4 months in Tg4510 tau P301L mice likely coincided with an advanced disease stage as this mouse model typically sees the onset of gliosis and consistent histological evidence of tau deposits by the third month [42]. Similarly, starting CR at 7.5 months in our PS19 tau P301S mice may have corresponded with a more advanced disease state, as significant microgliosis and tau pathology are likely present by 6 months of age [20]. This suggests that for CR to mitigate tau pathology and its associated neurodegeneration effectively, it might need to be introduced either before the disease's onset or very early in its progression.

The spatial learning improvements observed in PS19 CR mice, despite no significant alterations in key pathophysiological markers, suggest that the benefits of CR might extend beyond direct modulation of disease pathology. This observation aligns with the broader perspective in AD in humans, where although the progression of plaques and tangles is often causally linked to the cognitive symptoms, a discrepancy between the two exists in some individuals who show substantial cognitive decline yet have only a minor number of lesions [43]. While several studies have shown CR’s effect on reducing Aβ plaques, to our knowledge, only two have concurrently assessed cognitive outcomes. Such an emphasis on plaque and tangle reductions potentially obscures the possibility that CR might mediate cognitive benefits through alternative pathways.

Accumulating evidence underscores the importance of adult hippocampal neurogenesis in learning and memory [44-46], with several studies linking enhanced neurogenesis to cognitive improvements in rodents [47-49]. Contrary to previous research where dietary restriction was associated with enhanced neurogenesis [50-54], we found that CR decreased both EdU-positivity and the amount of PSA-NCAM-positive immature neurons in the DG. Despite this reduction, the proportion of newly generated PSA-NCAM-positive immature neurons in PS19 CR mice was similar to that in PS19 AL mice, suggesting that the generation of new immature neurons remained unaffected. The decline in PSA-NCAM expression under CR raises two possibilities: reduced survival of immature neurons or enhanced neuronal maturation. Given PSA-NCAM’s role in hippocampus-related synaptic plasticity [55, 56], and learning and memory functions [57-60], the preserved hippocampal-dependent spatial learning in our CR mice suggests that its decrease is unlikely to be detrimental. It could instead indicate a more efficient neuronal maturation process under CR conditions, characterized by a quicker loss of the development marker PSA-NCAM and the emergence of mature neural cell markers. Thus, despite reduced cell proliferation, CR might promote the survival and maturation of newborn neurons, aligning with previous studies showing CR regimens favor neuronal survival over neural stem cell proliferation [51, 53, 61].

Expanding our observations beyond the neurogenic zone, we found CR to confer an overall decline in hippocampal cell proliferation. Previous research has shown that CR reduces cell proliferation in various mitotic tissues, a phenomenon commonly attributed to decreased metabolic rates and cellular turnover, potentially reducing tumorigenesis risk and prolonging lifespan [62-65]. Specifically, in PS19 mice, CR predominantly reduced the number of proliferating Iba1+ cells, indicating a decrease in the proliferation of microglia or infiltrating monocytes, as Iba1 is a marker for both microglia and blood-derived macrophages [66]. However, the overall Iba1+ cell count in the hippocampus remained comparable to that in AL mice (Supplementary Fig. 4), which may suggest an elevated turnover rate of microglia in PS19 mice. CR could reduce this effect, which would mirror its observed effects in other mitotically competent cell types.

How CR’s modulation of microglial proliferation and lifespan may affect spatial learning performance is yet to be understood. In the PS19 mouse model, microgliosis is one of the earliest manifestations of disease, preceding the appearance of neurofibrillary tangles and neuronal loss [20]. Accumulating evidence suggests that activated microglia play an important role in the progression of tau-mediated neuropathology, including tau spreading and tau-induced synaptic loss [20, 67]. Accordingly, interventions targeting microglial activity, such as the use of the CSF1R inhibitor JNJ-40346527 or the inactivation of microglial NF-κB pathway, have demonstrated protective effects against tau-induced neurodegeneration and functional and cognitive decline in PS19 mice [68, 69]. Additionally, removal of senescent microglia has been shown to prevent tau-dependent pathology and the associated cognitive decline, underscoring the critical role of microglial senescence in tau-mediated neuropathology [24]. Considering these findings, CR’s modulation of microglial proliferation could be pivotal for cognitive performance. Yet, CR’s limited impact on neuroinflammatory and senescence markers suggests it may not effectively counteract chronic microglial activation. Previous research has demonstrated that CR can inhibit microglial activation in several animal models, including cortical injury, LPS-induced microglial activation, aging, and in APP/PS1 mice [10, 50, 70-72], when initiated before the onset of pathology. Similarly, in PS19 mice, interventions aimed at either inhibiting microglia proliferation or removing senescent microglia were typically implemented prior to microgliosis onset. Thus, the timing of CR initiation appears crucial for its therapeutic potential in modulating microglial activation.

Effective treatments for human tauopathies are currently lacking, with most drug candidates for non-AD tauopathies targeting tau pathology [73], mirroring the treatment strategy in AD that focuses on amyloid burden reduction [74]. However, these strategies often fall short in addressing the behavioral and cognitive deficits that largely contribute to the patient’s substantial functional disability and loss of independence [75, 76]. In this context, the potential of CR to enhance cognitive functions and possibly improve locomotion, extending beyond the direct modulation of hallmark disease pathologies, becomes increasingly significant. While CR has shown promise in rodent models of AD and related tauopathies, its translatability to human disease remains uncertain. Accumulating evidence indicates that moderate CR in humans recapitulates outcomes observed in rodent studies, including significant cardiometabolic health improvements [77-79] and cognitive benefits [80]. This suggests that some mechanisms of CR are well conserved across species. Still, the feasibility of implementing CR early in the progression of tauopathies and adherence to long-term nutrient restriction may prove challenging in a clinical setting. Therefore, exploring the underlying mechanisms behind CR’s cognitive benefits and identifying pharmacological agents that can replicate these effects could pave the way for novel therapeutic approaches for tauopathies.

In conclusion, our study shows that a short-term 30% CR regimen ameliorates spatial learning deficits in PS19 mice without affecting phosphorylated tau levels or associated neuroinflammation in the hippocampus. Notably, we observed a reduction in Iba1+ cell proliferation under CR, potentially due to metabolic changes influencing the cellular turnover of what might be brain-resident microglia. Given the central role of microglia in tau-driven neuropathology, such alterations could be essential for spatial learning preservation. However, our findings suggest that CR may be more effective as a preventive strategy than an intervention in more advanced-stage disease. Future research should probe the effects of CR in PS19 mice when introduced early in the onset of the disease. Although our study focused on a hippocampal-dependent cognitive function, exploring other brain regions implicated in cognitive functions and behavior affected by tauopathies is warranted.

Limitations:

Our study has several limitations. First, our choice to use only male mice was based on our observation that the phenotype appeared earlier and was more pronounced compared to their female counterparts. This raises the question of whether the effects observed in male PS19 mice would be observed in females. Second, the 60-second duration for BM trials may not allow all PS19 mice sufficient time to approach and sniff the escape hole. Extending the BM trial duration would provide mice with more time and allow for measuring the escape latency. Additionally, we did not evaluate spatial memory, which is another crucial aspect of cognitive function that should be explored in future studies. Third, the heightened locomotor activity in PS19 CR mice, potentially arising from food anticipatory behavior or increased alertness due to reduced satiety, may confound the interpretation of Barnes maze results. Addressing this in future studies could involve integrating tests for anxiety-related behaviors, detailed movement analyses via video tracking, and cross-validation with additional cognitive tasks that are less affected by motor activity, to ensure a more accurate differentiation between anxiety-related behaviors and true cognitive performance. Fourth, while Iba1 is commonly used as a microglial marker, we cannot entirely exclude the contribution of peripheral monocytes that might have infiltrated the tissue. Future studies using additional markers or techniques could help differentiate between brain-resident microglia and brain infiltrating monocytes.

Supplementary Material

SFig 1

Supplementary Figure 1. Effects of calorie restriction on body weights in PS19 and age-matched WT male mice in young and old cohorts. (A-B) Starting body weights for 4.5-month-old (A) and 7.5-month-old (B) PS19 and WT mice before CR implementation. (C-D) Body weights trends over 6 weeks for PS19 and WT mice on CR or AL diets in young (C) and old (D) cohorts. Body weight variations were analyzed by comparing diet effect within each genotype. Data are mean ± s.e.m and analyzed by unpaired two-sided t-test with Welch’s correction (A-B) and two-way ANOVA with Sidak’s multiple comparisons test (C-D). *p<0.05, **p<0.01, p<0.001***, p<0.0001****.

SFig 2

Supplementary Figure 2. Comparison of total tau mRNA expression levels in the hippocampus of 9-month-old PS19 mice under AL and CR diets. There were no significant differences between the groups.

SFig 3

Supplementary Figure 3. Increased EdU-positivity in the hippocampus of 9-month-old PS19 mice. Representative photomicrographs show widespread EdU+ cells (green) throughout the hippocampus in PS19 mice on AL diet, with a noticeable reduction in mice under CR. Cell nuclei were labeled with DAPI (blue). Scale bars =100 μm.

SFig 4

Supplementary Figure 4. Quantitative analysis of Iba-1 immunostaining in the hippocampus of 9-month-old PS19 and age-matched WT mice under AL and CR diets. No diet-related effect on Iba-1+ cell densities was observed across genotypes. Analysis was performed using on 20x magnification images, with three images per animal and n=3-8 per group.

Data availability statement

Data supporting the findings of this study are available on request from the corresponding author.