Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Mar 14.
Published in final edited form as: J Alzheimers Dis. 2024;97(4):1751–1763. doi: 10.3233/JAD-230465

Senescence targeting methods impact Alzheimer’s disease features in 3xTg mice

Pei Y Ng a, Cheng Zhang b,c, Hu Li b,c, Darren J Baker a,c,d,e,*
PMCID: PMC10939718  NIHMSID: NIHMS1968907  PMID: 38306030

Abstract

Background:

Cellular senescence has been associated with neurodegenerative disease and clearance of senescent cells using genetic or pharmaceutical strategies (senolytics) has demonstrated beneficial effects in mouse models investigating individual disease etiologies of Alzheimer’s disease (AD). However, it has remained unclear if senescent cell clearance in a mouse model exhibiting both plaque and tau pathologies modifies the disease state (3xTg).

Objective:

To investigate the effects of senescent cell clearance in the 3xTg mouse model.

Methods:

3xTg mice were treated with senolytics (ABT263 (navitoclax; NAVI), a combination of dasatinib and quercetin (D+Q)), or subjected to transgene-mediated removal of p16-expressing cells (via INK-ATTAC).

Results:

Senolytic treatments consistently reduced microgliosis and ameliorated both amyloid and tau pathology in 3xTg mice. Using RNA sequencing, we found evidence that synaptic dysfunction and neuroinflammation were attenuated with treatment. These beneficial effects were not observed with short-term senolytic treatment in mice with more advanced disease.

Conclusion:

Overall, our results further corroborate the beneficial effects senescent cell clearance could have on AD and highlight the importance of early intervention for the treatment of this debilitating disease.

Keywords: Alzheimer’s disease, cellular senescence, tauopathy, amyloid-β

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disorder pathologically defined by the presence of β-amyloid (Aβ)-containing plaques and tau-containing neurofibrillary tangles [1]. It is the most common cause of dementia and the greatest risk factor for developing AD is increasing age [1,2]. With the general global aging trend [3], understanding the link between age and AD risk is as important as it has ever been, especially if this link is modifiable. Senescence, a state of irreversible cell-cycle arrest accompanied by an inflammatory phenotype, has been linked to a variety of age-related diseases [4–7] and provides one possible explanation for why increased age predisposes to AD development. Cells with senescent features have been identified in the brains of AD patients, although the exact role senescence has in the pathogenesis of AD in humans remains difficult to study due to difficulties in sample collection and a lack of specific markers for senescence in vivo [8].

The evidence linking senescence with AD has largely been established in mouse models [4,9–12], where removal of senescent cells leads to improvements in cognitive function or amelioration of AD pathology. For example, in the P301S mouse model of tauopathy, senescent cell removal (via INK-ATTAC transgenic or ABT263) prevented gliosis, reduced tau hyperphosphorylation, delayed neurodegeneration, and preserved memory formation [4]. In the rTg4510 (P301L) mouse model of tauopathy, senescent cell removal via senolytics reduced neurofibrillary tangle burden, ventricular enlargement, and neurodegeneration [9]. In the APP/PS1 mouse model, administration of senolytics led to less neuroinflammation, fewer Aβ plaques, and memory retention [10]. However, it is unclear what the interplay between senescence, Aβ plaques, and tau pathology is, as most studies have utilized mouse models with just one aspect of AD pathology.

In this study, we sought to investigate the effect of senescent cell clearance in the 3xTg mouse model of AD. The 3xTg mouse model is a triple transgenic mouse, being triple homozygous for PSEN1M146V, MAPTP301L, and the APP Swedish mutation [13], which models both plaque and tau pathology. The benefits of senescent cell clearance were assessed for impact on plaque burden, tau pathology, gliosis, neurodegeneration, synaptic dysfunction, neuroinflammation, and cognitive deficits using either transgenic (INK-ATTAC) or pharmacological (navitoclax (ABT263) or dasatinib/quercetin) means. These features of AD have been reported in the 3xTg mouse model [13–15]. We find that long-term removal of senescent cells (from 12–18 months), but not short-term removal (beginning 6 weeks before sacrifice), had benefits in plaque pathology, tau pathology, microgliosis, synaptic dysfunction, and neuroinflammation. However, our colony of 3xTg mice displayed genetic drift in the form of delayed onset of pathology. Notably, we found no evidence of neurodegeneration or robust senescence in 3xTg mice compared to age-matched wildtype mice, and only a mild cognitive deficit.

MATERIALS AND METHODS

Mouse strains

All experiments were reviewed and approved by the Mayo Clinic Institutional Animal Care and Use Committee. 3xTg mice were originally purchased from The Jackson Laboratory (stock no.004807). Control BL6129SF2/J mice were originally purchased from The Jackson Laboratory (stock no.101045). Control mice were age-matched to 3xTg mice. Mice from this cohort were randomly assigned to receive ABT263 (Selleckchem #S1001) or vehicle (Phosal 50 PG, Lipoid NC0130871, 60%; PEG400, Sigma 91893, 30%; ethanol, 10%). ABT263 was administered by oral gavage at a dose of 50 mg per kg body weight on a repeating regimen of five consecutive days of treatment followed by 16 days of rest. Mice from this cohort were also randomly assigned to receive a combination of dasatinib (LC labs #D-3307) and quercetin (sigma #Q4951) or vehicle (PEG400, Sigma 91893, 30%; 0.9% NaCl;70%). Dasatinib and quercetin were administered by oral gavage at a dose of 12 mg per kg for dasatinib and 50 mg per kg for quercetin twice a week, with 2–3 days between treatments.

To generate 3xTg;ATTAC mice, C57BL/6;ATTAC mice generated as previously described [6,7] were crossed to 129S1/SvImJ mice. F2;ATTAC hybrid mice were then crossed to the 3xTg mice until 3xTg;ATTAC mice homozygous for the co-injected APPSwe and tauP301L transgenes, and the PS1M146V Psen1 knock-in mutation were obtained. Control BL6/129;ATTAC mice were also generated from the F2;ATTAC x 3xTg cross. Control mice were age-matched and had a similar genetic background resulting from the F2;ATTAC x 3xTg cross. Mice from this cohort were randomly assigned to receive AP20187 (AP; B/B homodimerizer; Clontech) or vehicle (100% EtOH, 4%; PEG400, Sigma 91893, 10%; 2% tween 20, Fisher BP337500; 86%) intraperitoneally twice a week, with 2–3 days between treatments. Dosing of AP was 2 mg per kg body weight.

Mice were treated from 12 months to 18 months of age for the prophylactic treatment and treated from 16.5 months to 18 months of age (for 6 weeks) for the “short-term” treatment. Mice were housed in a 12h:12h light:dark cycle environment and had ad libitum access to standard irradiated pelleted chow (LabDiet product 5053).

Single cell preparation, FACS, and Quantitative RT-PCR

Dissociation and subsequent sorting of cerebral tissue in 18-month-old BL6/129;ATTAC and 3xTg;ATTAC mice was performed as described [4]. Cells were sorted directly into lysis buffer and RNA was extracted using RNeasy micro kits according to the manufacturer’s instructions (Qiagen #74004). Subsequent cDNA synthesis and qRT-PCR for Casp8 and Gfp were performed as described [4].

Immunofluorescence and immunohistochemistry

Mice were transcardially perfused with ice-cold DPBS (gibco #14190144) or 4% paraformaldehyde (PFA, Sigma-Aldrich #P6148) followed by ice-cold DPBS (gibco #14190144). 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 (300g sucrose, 300 ml ethylene glycol, 500 ml PBS) at −20°C.

For immunofluorescence, Iba1 staining (wako #019–19741; 1:200), Gfap staining (abcam #ab4674; 1:1000), phospho-tau S202/S205 or AT8 (ThermoFisher #MN1020; 1:500), and synaptophysin staining (abcam #ab16659; 1:250) were carried out on free-floating brain sections from bregma −2.78mm to −3.28mm as previously described [4], with the addition of TrueBlack (Biotium #23007) quenching of autofluorescence. Images were acquired on a Zeiss LSM 780 confocal system using a multi-track configuration. Mean fluorescence intensity measurements were calculated based on maximum intensity projection of 30 x-y optical sections (z step = 0.79 μm).

For immunohistochemistry, Aβ1–42 (Sigma-Aldrich #AB5078P; 1:200) was carried out as previously described [4] on brain sections from bregma −2.78mm to −3.28mm. In short, fixed brain sections underwent sodium citrate antigen retrieval (Vector #H-3300), followed by blocking with 3% bovine serum albumin (BSA; Sigma-Aldrich #A7906), and an avidin/biotin blocking kit (Vector #SP-2001). The brain sections then underwent overnight incubation at 4°C with the primary antibody. Subsequently, the slides were incubated with a biotin secondary (Jackson Immunoresearch #111–065-144) for 45 mins at room temperature, followed by a peroxide block (Sigma-Aldrich #H3410), and incubation with HRP-Avidin (Vector #A-2004). Finally, slides were developed using a DAB substrate kit (Vector #SK-4100). Nissl staining (MPBio #10510–54-0) was carried out on brain sections from bregma −2.78mm to −3.28mm as previously described [4].

Bulk RNA-sequencing analysis

Crushed frozen hippocampi were homogenized through QIAshredder (Qiagen #79656), and RNA was extracted using the RNeasy Mini Kit (Qiagen #74104), using their standard protocol. The quality of extracted RNA was assessed using an RNA Integrity Number (RIN) and a DV200 value generated by the Agilent Fragment Analyzer. Qualitative concentration was determined using a Qubit Fluorometer.

RNA libraries were prepared using 200 ng of total RNA according to the manufacturer’s instructions for the Illumina Stranded mRNA Sample Prep Kit (Illumina, San Diego, CA). The concentration and size distribution of the completed libraries were determined using an Agilent Bioanalyzer DNA 1000 chip (Santa Clara, CA) and Qubit fluorometry (Invitrogen, Carlsbad, CA). Libraries were sequenced at 50 million fragment reads per sample following Illumina’s standard protocol Illumina’s standard protocol using the Illumina NovaSeq™ 6000 and S1 flow cell. The flow cells were sequenced as 100 X 2 paired-end reads using NovaSeq S1 sequencing kit and NovaSeq Control Software v1.8.0. Base-calling was performed using Illumina’s RTA version 3.4.4.

Fastq files containing paired-end RNA-seq reads were aligned with STAR [16] against the UCSC mouse reference genome mm10 using 2-pass mode. Gene level counts were obtained using the subRead featureCounts program (v1.4.6) [17] and gene models from the UCSC mm10 annotation. Differential expression analysis was performed using the DESeq2 R package with default parameters [18]. Gene set enrichment analysis (GSEA) software was used for pathway analysis [19]. To run GSEA, we first obtained gene lists ranked by log2 fold change from differential analysis results. We used these gene lists as input for the GSEA-Preranked function in the GSEA java package against gene sets from MSigDB. To run GSEA against the SenMayo gene panel, we used the SenMayo gene panel as a gene set in place of the MSigDB gene sets and followed the same procedure.

Barnes maze

Barnes maze was carried out as previously described [15]. In short, 18-month-old 3xTg and age-matched BL6/129 mice were acclimatized to the room with the maze the day before. A training phase was conducted where the mice were acclimatized to the Barnes maze apparatus. Barnes maze was then carried out three times per mouse over four days, where the time taken for the mice to escape was averaged over the three trials per day. Mice were given 180s to escape, and if the mice did not escape the escape time was recorded as this maximum of 180s. Animal tracking was carried out using Noldus Ethovision XT.

Statistical analysis

Prism software was used for all statistical analysis. A one-way ANOVA with Tukey’s multiple comparisons test was used in Figures 1, 2, and 3, and Supplementary figures 1, 2, and 3. A two-way ANOVA with Tukey’s multiple comparisons test was used in Supplementary figure 2. A Student’s two-tailed unpaired t-test was used in Supplementary figure 1. For consistency in these comparisons, the following denotes significance in all figures: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 1. Treatment with ABT263 resulted in attenuation of amyloid plaques and phospho-tau.

Figure 1.

Open in a new tab

(A) Study design for treatment of 3xTg mice. (B) Breeding scheme to generate 3xTg;ATTAC mice. (C) AT8 staining in the CA1 region of the hippocampus, with representative images on the left and quantification of MFI on the right. (D) Aβ1–42 staining in the CA1 region of the hippocampus, with representative images on the left and quantification of number of plaques and size of plaques on the right. Scale bars, 100 μm. Data are mean±SEM. *P < 0.05; **P < 0.01, ****P < 0.0001 (one-way ANOVA with Tukey’s multiple comparisons test).

Figure 2. Treatment with D+Q and AP resulted in attenuation of amyloid plaques and phospho-tau.

Figure 2.

Open in a new tab

(A) AT8 staining in the CA1 region of the hippocampus, with representative images on the left and quantification of MFI on the right, for D+Q treatment. (B) AT8 staining in the CA1 region of the hippocampus, with representative images on the left and quantification of MFI on the right, for AP treatment. (C) Aβ1–42 staining in the CA1 region of the hippocampus, with representative images on the left and quantification of number of plaques and size of plaques on the right, for D+Q treatment. (D) Aβ1–42 staining in the CA1 region of the hippocampus, with representative images on the left and quantification of number of plaques and size of plaques on the right, for AP treatment. Scale bars, 100 μm. Data are mean±SEM. *P < 0.05; **P < 0.01, ***P < 0.001 (one-way ANOVA with Tukey’s multiple comparisons test).

Figure 3. Treatment with ABT263, D+Q, and AP resulted in attenuation of microgliosis.

Figure 3.

Open in a new tab

(A) Iba1 and Gfap staining in the CA1 medial region of the hippocampus, with representative images on the left, and quantification of MFI and number of cells on the right, for ABT263. (B) Iba1 and Gfap staining in the CA1 medial region of the hippocampus, with representative images on the left, and quantification of MFI and number of cells on the right, for D+Q. (C) Iba1 and Gfap staining in the CA1 medial region of the hippocampus, with representative images on the left, and quantification of MFI and number of cells on the right, for AP. Data are mean±SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (one-way ANOVA with Tukey’s multiple comparisons test).

RESULTS

Removal of senescent cells attenuates Aβ and tau pathology

To study if the removal of senescent cells could alleviate Aβ and tau pathology, we used three different methods of senescent cell removal (Figure 1A). We used two different types of senolytics, the Bcl-2 family inhibitor ABT263 (Navitoclax or NAVI) [20], and the combination of dasatinib and quercetin (hereafter D+Q) which targets a variety of anti-apoptotic pathways [21]. We also utilized the INK-ATTAC system, by crossing the INK-ATTAC transgene (hereafter ATTAC) to the 3xTg mouse model (Figure 1B), such that administration of AP20187 (hereafter AP) eliminates p16INK4a-expressing senescent cells [4]. Treatment was initiated at 12-months of age in a prophylactic approach, as both plaque and tau pathology were rarely seen in our cohort at this time point. Only female 3xTg mice were used, as male 3xTg mice had less and more variable pathological burden, both in our hands (data not shown) and as previously reported [22,23]. An 18-month end-point was chosen, as that was the time point where we could reliably detect both Aβ and phosphorylated tau pathology. There was a higher-than-expected loss of mice after 18 months of age regardless of whether they were control or 3xTg. By performing biopsies on these animals to determine the cause of death, we found that the majority had splenomegaly or tumors (most commonly in the liver or spleen). Due to this high dropout of mice, we did not assess animals after 18-months of age. Since we generated all our animals through breeding in-house from original founder lines, it is possible that a cancer-causing mutation was inadvertently introduced in the original breeding animals that were subsequently inbred.

Tau pathology was assessed through immunofluorescence staining of AT8 (phosphor-tau Ser202, Thr205) to detect tau hyperphosphorylation. As expected, we found an increased burden in 3xTg mice compared to age-matched BL6/129 mice (Figure 1C). Importantly, ABT263 administration reduced AT8 positivity in both the CA1 medial and lateral regions (Figure 1C). The CA1 medial and lateral regions of bregma −2.78mm to −3.28mm were chosen as those were the regions where we could consistently find pathology (Supplementary figure 1A). We next sought to assess the effect of ABT263 administration on plaque burden by immunohistochemistry staining of Aβ1–42. Although there was a slight (but not significant) reduction in the number and size of plaques in the CA1 medial region, the CA1 lateral region demonstrated a statistically significant reduction in plaque deposition (Figure 1D).

Next, we tested if this reduction in tau and plaque pathology was recapitulated using alternative senescent cell removal strategies. We tested the senolytic combination of D+Q or the transgenic INK-ATTAC genetic approach on 3xTg mice. INK-ATTAC mice express a drug-inducible Caspase 8 molecule that promotes cell death in p16-expressing cells when treated with AP [4,6,7]. In response to either D+Q or AP treatment, there was a reduction in AT8 burden in both the CA1 medial (non-significant) and lateral (significant) regions (Figures 2A and B). In assessing the plaque burden, only the size of plaques in the CA1 lateral region was significantly reduced with D+Q treatment (Figures 2C and D). One thing to note is the variability in phospho-tau and plaque burden between the different cohorts. For example, there were very few plaques present in the 3xTg;ATTAC cohorts compared to the 3xTg cohorts (Figures 1D, 2C, and 2D) and less phospho-tau burden (Figures 1C, 2A, and 2B). This perhaps resulted from the extensive breeding needed to combine the ATTAC transgene onto the 3xTg background, which could have resulted in a subtle genetic drift that could have influenced pathology. Furthermore, others have reported significant variability in this mouse model when housed at different locations. This has been reported not just as sexual dimorphism [22], but also in terms of delayed onset of pathology [24]. In our cohort at 12 months of age, plaques were rarely detected, and AT8 staining was also variably detected, unlike original reports reporting 100% of mice having detectable plaque at this timepoint or even earlier [13,14].

Having established that removing senescent cells from mice prior to disease onset impacted pathology (including plaques and phospho-tau), we next tested if short-term treatment with senolytics impacted established disease. To do this, we treated 3xTg or age-matched control BL6/129 mice with ABT263, D+Q, or their respective vehicles for 6 weeks before sacrifice at 18 months (Supplementary figure 1B). This short-term treatment with senolytics resulted in minimal reduction in plaque (Supplementary figure 1C) or phospho-tau (Supplementary figure 1D) burden. Hence, early intervention to remove senescent cells is crucial to have an effect on reducing plaque and phospho-tau burden.

Removal of senescent cells reduces microgliosis

Gliosis has been observed with AD [25] and correlates with disease pathology in a variety of mouse models of AD including 3xTg [4,14,26,27]. To determine if senescent cell removal was able to affect gliosis, immunofluorescence staining was carried out for microgliosis (Iba1 [28]) and astrogliosis (Gfap [29]). Treatment with ABT263 (Figure 3A), D+Q (Figure 3B), and elimination of p16-positive senescent cells via ATTAC (Figure 3C) were able to attenuate microgliosis, as measured by fewer Iba1+ cells and lower Iba1 mean fluorescence intensity (MFI). However, only the elimination of p16-positive senescent cells via ATTAC was able to reduce astrogliosis to a significant degree (Figure 3C), possibly due to the mild astrogliosis phenotype in the other cohorts and variability in response to senolytic treatment.

Senescent cell clearance attenuates synaptic deficits and neuroinflammation

To elucidate the possible mechanisms by which senescent cell removal could have resulted in the attenuation of pathology, we performed RNA sequencing on the hippocampi of BL6/129 and 3xTg mice treated with ABT263, D+Q, and their respective controls. However, we did not observe a robust induction in senescence-related genes including Cdkn1a or Bcl2 in the RNAseq from our cohort of mice (Supplementary figure 2A) and Cdkn2a was not detected. Although there was a strong induction of possible SASP components like Il1a, Il1b, and Serpine1, the upregulation of these SASP components could also occur in neuroinflammation and is not specific to senescence. Additionally, these SASP components did not seem to be attenuated with senolytic administration (Supplementary figure 2A). Overall, although all three senescent cell removal strategies resulted in a similar reduction of Aβ, phospho-tau, and microgliosis, senescent cell accumulation in 3xTg mice and subsequent removal with senolytic treatment could not be detected on a transcriptional level. Perhaps the senescent cell population contributing to pathology was too small to be detected by bulk RNA sequencing. Therefore, we turned to measuring levels of Casp8 and Gfp, both markers of the ATTAC transgene, in different cell populations isolated by flow sorting of brains from 18-month-old BL6/129;ATTAC and 3xTg;ATTAC mice. This would give us a better idea of which cellular population was expressing the ATTAC transgene and hence which cellular population could be senescent and sensitive to elimination with administration of AP, which is similar to what we have done previously [4]. We found that microglia from 3xTg;ATTAC mice had a significant increase of Casp8 and Gfp compared to microglia isolated from age-matched BL6/129;ATTAC mice (Supplementary figures 2B and C). Astrocytes from 3xTg;ATTAC mice showed a similar trend, although the increase in Casp8 and Gfp was not statistically significant (Supplementary figures 2B and C). There was no increase in Casp8 and Gfp in the oligodendrocyte or neuron-enriched fraction (data not shown). These results suggest that microglia and perhaps astrocytes are predisposed to senescence in aged 3xTg mice.

The relatively mild senescence phenotype could be related to the general delay of pathology seen in the cohort, where plaque deposition as detected by Aβ1–42 and phospho-tau was only reliably detected at 18 months instead of the 6–12 months in other cohorts [13,14]. Additionally, Nissl staining revealed that there was no overt neurodegeneration in this cohort (Supplementary figure 2D), also further demonstrating the delay in pathology onset in the cohort. This lack of neurodegeneration in our cohort adds to the conflicting results from different studies regarding the presence or absence of neurodegeneration in 3xTg mice. For example, some studies have demonstrated neurodegeneration in 3xTg mice [30–33] while others show cognitive impairment and pathology without significant neuronal loss [34,35].

Memory deficits were also mild in the cohort (Supplementary figure 2E). This was assessed through the Barnes maze, which measures deficits in spatial reference memory and learning [15]. 3xTg mice assigned to receive the vehicle for ABT263 did not show any cognitive deficit compared to control BL6/129 mice, whereas 3xTg mice assigned to receive the vehicle for D+Q and AP did. The 3xTg mice that received D+Q and AP then subsequently showed an improvement in memory compared to the vehicle-treated mice. Hence, although senescent cell removal resulted in a partial memory improvement for some mice, this was difficult to fully access due to the delay in pathology for some mice in the cohort.

Gene set enrichment analysis (GSEA) revealed that 3xTg mice had a downregulation of synapse-related pathways (Figure 4A, for leading-edge subsets see Supplementary file 1), suggesting a synapse dysfunction in 3xTg mice which has been previously reported [13]. 3xTg mice treated with senolytics also had an upregulation of synapse-related pathways compared to vehicle-treated 3xTg mice (Figure 4B, for leading-edge subsets see Supplementary file 2), suggesting that this synapse dysfunction could be attenuated with removal of senescent cells. Functional annotation of genes that were downregulated in vehicle-treated 3xTg mice compared to vehicle-treated BL6/129 mice also suggest that there was synapse loss in 3xTg mice and a deficit in neurogenesis (Figure 4C and Supplementary file 3). Synaptophysin staining also suggests a synaptic deficit in 3xTg mice compared to BL6/129 mice and a rescue of this deficit with ABT263 treatment (Supplementary figures 3A and B). However, this synaptic deficit and rescue were variable, which was also reflected in the z-scores of the top-enriched synapse-related pathway from the functional annotation in figure 4C, GO:0050804 modulation of chemical synaptic transmission (Supplementary figure 3C, see Supplementary file 4 for gene names and z scores). The lack of a consistent synapse dysfunction may also be related to the delay in pathology in this cohort. This variability in response was also reflected in another of the top pathways from the functional annotation in Figure 4C, GO 0007399 nervous system development (Supplementary figure 3D, see Supplementary file 4 for gene names and z scores), where the deficit in neurogenesis seen in 3xTg mice (Figure 4C) was reversed only in some mice receiving senolytic therapy.

Figure 4. 3xTg treated mice have synaptic deficits, are more neuroinflammatory, and have some features of senescence when compared to age-matched BL6/129 mice.

Figure 4.

Open in a new tab

(A) Gene ontology pathway enrichment of BL6/129 vehicle-treated mice compared to 3xTg vehicle-treated mice using GSEA. (B) Gene ontology pathway enrichment of 3xTg mice treated with ABT263 and D+Q compared to 3xTg vehicle-treated mice using GSEA. (C) Functional annotation analyses of significantly upregulated genes of vehicle-treated 3xTg mice compared to vehicle-treated BL6/129 mice. (D) Functional annotation analyses of significantly downregulated genes of vehicle-treated 3xTg mice compared to vehicle-treated BL6/129 mice. An FDR cut-off q-value of 0.25 was used.

Additionally, GSEA suggests that several inflammatory pathways are upregulated in vehicle-treated 3xTg mice compared to vehicle-treated BL6/129 mice (Figure 4A), which were then downregulated when compared to senolytic-treated 3xTg mice (Figure 4B). Functional annotation of upregulated genes of vehicle-treated 3xTg mice compared to BL6/129 mice also suggests that the brains of 3xTg mice were more inflammatory (Figure 4D and Supplementary file 3). The top pathway enriched in this functional annotation, GO 0002376 immune system process (Supplementary figure 4A, see Supplementary file 4 for gene names and z scores), shows that the genes upregulated in this process in 3xTg VEH mice were downregulated only in some mice treated with senolytics.

There were also some senescence-related pathways that were increased in 3xTg mice compared to BL6/129 mice, which were then decreased with senolytic treatment. For example, the SenMayo panel, a novel gene set with predictive ability for senescent cell detection across tissues and species [36], was upregulated in 3xTg VEH mice compared to BL6/129 VEH mice (Figure 4A), and subsequently downregulated in 3xTg mice treated with senolytics compared to 3xTg mice treated with vehicle (Figure 4B). Functional annotation of genes upregulated in 3xTg mice treated with vehicle compared to BL6/129 mice also suggested an upregulation of other pathways associated with senescence in 3xTg mice, such as regulation of proliferation and regulation of apoptosis (Figure 4D). When looking at the z-scores of the top pathways associated with these annotations, GO 0042127 regulation of cell population proliferation (Supplementary figure 4B, see Supplementary file 4 for gene names and z scores) and GO 0043069 regulation of programmed cell death (Supplementary figure 4C, see Supplementary file 4 for gene names and z scores) respectively, senolytic treatment seems to have variable impacts on these pathways, with only some mice showing a reversal of these pathways with senolytic treatment.

DISCUSSION

In this study, we demonstrate that long-term administration of ABT263, D+Q, and AP attenuated various AD-related pathologies including Aβ1–42, AT8, and microgliosis. These interventions also partially attenuated spatial memory deficits. RNA sequencing analyses suggest that the three treatments also attenuated deficits of synaptic transmission, and neuroinflammation in 3xTg mice. However, it is still unclear if these beneficial effects were due to the removal of senescent cells. The most straightforward common target of the three treatments is the elimination of senescent cells. Certainly, the beneficial effects seen with the elimination of p16-positive senescent cells via the ATTAC system suggest that the mechanism of action that led to the attenuation of pathologies is related to senescence, rather than an off-target effect of the senolytic treatments.

However, directly interrogating specific senescence-related genes like Cdkn1a, and SASP factors like Il1a, Il1b, and Serpine1 via bulk RNA sequencing did not demonstrate a consistent induction in 3xTg mice compared to BL6/129 mice and a subsequent reduction with senolytic treatment. Additionally, perhaps the mild disease phenotype in this cohort of 3xTg mice made it more difficult to detect senescence in the disease model and any subsequent effect senescent cell clearance might have had. Some studies have shown no increase in the transcript levels of various senescence markers in 3xTg mice [9,37]. However, we show increased expression of the ATTAC transgene in microglia and astrocytes isolated from 18-month-old 3xTg mice, a time point that is in a more advanced stage of disease than what had been previously investigated. In addition, we show an upregulation of various senescence-associated pathways such as the SenMayo pathway in 3xTg hippocampal tissues, whereas previous studies that failed to detect elevated levels of senescence markers focused on other brain regions such as the cortex or cerebellum. More focused strategies, such as focusing on specific cell types or brain regions in older 3xTg mice, may be needed to detect senescence in the 3xTg mouse model.

The mild disease phenotype demonstrated in this cohort of 3xTg mice, encompassing the delay of disease phenotype, lack of neurodegeneration, inconsistent memory deficits, inconsistent astrogliosis, and inconsistent synaptic dysfunction also highlights the considerable disease variability and genetic drift of the 3xTg mouse model. It is still unclear what exactly is causing this genetic drift. It has been suggested that the drift could have occurred due to the segregation of alleles in the mixed strain background, or changes in the transgene copy number within the single site of integration [24,38]. Alternatively, perhaps the different housing conditions and hence microbiota of the mice could have affected disease progression [39]. Regardless, this genetic drift is important to keep in mind in comparing different studies with the 3xTg mouse model.

Overall, this study adds to the growing body of evidence that strategies targeting senescent cells can lead to an amelioration of AD-related pathologies [4,9,10]. None of the senolytic strategies seem particularly advantageous over any other and each strategy has its limitations. For example, ABT263 and D+Q could have possible off-target impacts and ATTAC-mediated elimination is reliant on genetic modification. This study also highlights the importance of the intervention timing, as short-term treatment did not lead to a reduction in AD-related pathologies, unlike the prophylactic treatment which was initiated before the overt presence of pathologies. With senolytics already being used in clinical trials for AD [40–42], the question of when senolytic treatment should be administered is more pertinent than ever. This is especially true given that patients are often diagnosed with AD based on clinical symptom presentation, which typically manifests after considerable tau and amyloid pathology in the brain [43,44].

Supplementary Material

SuppFile3

Supplementary file 3. GO terms of pathways and associated genes of functional annotation analyses in Figures 4C and 4D.

SuppFile2

Supplementary file 2. Leading edge subsets of GSEA pathways in Figure 4B.

SuppFile1

Supplementary file 1. Leading edge subsets of GSEA pathways in Figure 4A.

SuppFile4

Supplementary file 4. Gene names and z scores of heatmaps in Supplementary Figures 3 and 4.

5

ACKNOWLEDGEMENTS

We would like to thank T. Thao, and R. Fierro Velasco for genotyping and animal support and the Mayo Clinic Genome Analysis Core for RT-qPCR instrumentation and assistance with the bulk RNA-seq experiment. This work was supported by the National Institutes of Health (R01AG053229, R01AG068076 to D.J.B.), the Glenn Foundation for Medical Research (D.J.B.) and the Agency for Science, Technology and Research, Singapore (stipend supporting P.Y.N.).

Footnotes

CONFLICT OF INTEREST

D.J.B. has a potential financial interest related to this research. He is a co-inventor on patents held by Mayo Clinic, patent applications licensed to or filed by Unity Biotechnology, and a Unity Biotechnology shareholder. Research in the Baker laboratory has been reviewed by the Mayo Clinic Conflict of Interest Review Board and is being conducted in compliance with Mayo Clinic Conflict of Interest policies. The other authors declare no competing interests.

DATA AVAILABILITY STATEMENT

The gene expression dataset analyzed in this study is available in the NCBI Gene Expression Omnibus repository (https://www.ncbi.nlm.nih.gov/geo/) under the accession number GEO: GSE231332. Other data supporting the findings of this study are available on request from the corresponding author.

REFERENCES

  • [1].Knopman DS, Amieva H, Petersen RC, Chételat G, Holtzman DM, Hyman BT, Nixon RA, Jones DT (2021) Alzheimer disease. Nat Rev Dis Primers 7, 1–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [2].Guerreiro R, Bras J (2015) The age factor in Alzheimer’s disease. Genome Med 7, 106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [3].Nations United (2022) World Population Prospects. https://population.un.org/wpp, Last updated 2022, Accessed on March 8, 2023. [Google Scholar]
  • [4].Bussian TJ, Aziz A, Meyer CF, Swenson BL, van Deursen JM, Baker DJ (2018) Clearance of senescent glial cells prevents tau-dependent pathology and cognitive decline. Nature 562, 578–582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [5].Xu M, Bradley EW, Weivoda MM, Hwang SM, Pirtskhalava T, Decklever T, Curran GL, Ogrodnik M, Jurk D, Johnson KO, Lowe V, Tchkonia T, Westendorf JJ, Kirkland JL (2017) Transplanted Senescent Cells Induce an Osteoarthritis-Like Condition in Mice. J Gerontol A Biol Sci Med Sci 72, 780–785. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Baker DJ, Wijshake T, Tchkonia T, LeBrasseur NK, Childs BG, van de Sluis B, Kirkland JL, van Deursen JM (2011) Clearance of p16Ink4a-positive senescent cells delays ageing-associated disorders. Nature 479, 232–236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [7].Baker DJ, Childs BG, Durik M, Wijers ME, Sieben CJ, Zhong J, Saltness RA, Jeganathan KB, Verzosa GC, Pezeshki A, Khazaie K, Miller JD, van Deursen JM (2016) Naturally occurring p16(Ink4a)-positive cells shorten healthy lifespan. Nature 530, 184–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [8].Liu R-M (2022) Aging, Cellular Senescence, and Alzheimer’s Disease. Int J Mol Sci 23, 1989. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [9].Musi N, Valentine JM, Sickora KR, Baeuerle E, Thompson CS, Shen Q, Orr ME (2018) Tau protein aggregation is associated with cellular senescence in the brain. Aging Cell 17, 12840–12840. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Zhang P, Kishimoto Y, Grammatikakis loannis, Gottimukkala K, Cutler RG, Zhang S, Abdelmohsen K, Bohr VA, Sen JM, Gorospe M, Mattson MP (2019) Senolytic therapy alleviates Aβ-associated oligodendrocyte progenitor cell senescence and cognitive deficits in an Alzheimer’s disease model. Nat Neurosci 22, 719–728. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [11].Hou Y, Wei Y, Lautrup S, Yang B, Wang Y, Cordonnier S, Mattson MP, Croteau DL, Bohr VA (2021) NAD+ supplementation reduces neuroinflammation and cell senescence in a transgenic mouse model of Alzheimer’s disease via cGAS-STING. Proc Natl Acad Sci U S A 118, e2011226118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [12].Hu Y, Fryatt GL, Ghorbani M, Obst J, Menassa DA, Martin-Estebane M, Muntslag TAO, Olmos-Alonso A, Guerrero-Carrasco M, Thomas D, Cragg MS, Gomez-Nicola D (2021) Replicative senescence dictates the emergence of disease-associated microglia and contributes to Aβ pathology. Cell Rep 35, 109228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [13].Oddo S, Caccamo A, Shepherd JD, Murphy MP, Golde TE, Kayed R, Metherate R, Mattson MP, Akbari Y, LaFerla FM (2003) Triple-Transgenic Model of Alzheimer’s Disease with Plaques and Tangles: Intracellular Aβ and Synaptic Dysfunction. Neuron 39, 409–421. [DOI] [PubMed] [Google Scholar]
  • [14].Belfiore R, Rodin A, Ferreira E, Velazquez R, Branca C, Caccamo A, Oddo S (2019) Temporal and regional progression of Alzheimer’s disease-like pathology in 3xTg-AD mice. Aging Cell 18, e12873. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [15].Attar A, Liu T, Chan W-TC, Hayes J, Nejad M, Lei K, Bitan G (2013) A Shortened Barnes Maze Protocol Reveals Memory Deficits at 4-Months of Age in the Triple-Transgenic Mouse Model of Alzheimer’s Disease. PLOS ONE 8, e80355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR (2013) STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Liao Y, Smyth GK, Shi W (2013) The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote. Nucleic Acids Res 41, e108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].Love MI, Huber W, Anders S (2014) Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15, 550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [19].Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102, 15545–15550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Zhu Y, Tchkonia T, Fuhrmann-Stroissnigg H, Dai HM, Ling YY, Stout MB, Pirtskhalava T, Giorgadze N, Johnson KO, Giles CB, Wren JD, Niedernhofer LJ, Robbins PD, Kirkland JL (2016) Identification of a novel senolytic agent, navitoclax, targeting the Bcl-2 family of anti-apoptotic factors. Aging Cell 15, 428–435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Zhu Y, Tchkonia T, Pirtskhalava T, Gower AC, Ding H, Giorgadze N, Palmer AK, Ikeno Y, Hubbard GB, Lenburg M, O’Hara SP, LaRusso NF, Miller JD, Roos CM, Verzosa GC, LeBrasseur NK, Wren JD, Farr JN, Khosla S, Stout MB, McGowan SJ, Fuhrmann-Stroissnigg H, Gurkar AU, Zhao J, Colangelo D, Dorronsoro A, Ling YY, Barghouthy AS, Navarro DC, Sano T, Robbins PD, Niedernhofer LJ, Kirkland JL (2015) The Achilles’ heel of senescent cells: from transcriptome to senolytic drugs. Aging Cell 14, 644–658. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Dennison JL, Ricciardi NR, Lohse I, Volmar C-H, Wahlestedt C Sexual Dimorphism in the 3xTg-AD Mouse Model and Its Impact on Pre-Clinical Research. J Alzheimers Dis 80, 41–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Carroll JC, Rosario ER, Kreimer S, Villamagna A, Gentzschein E, Stanczyk FZ, Pike CJ (2010) Sex differences in β-amyloid accumulation in 3xTg-AD mice: Role of neonatal sex steroid hormone exposure. Brain Res 1366, 233–245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Javonillo DI, Tran KM, Phan J, Hingco E, Kramár EA, da Cunha C, Forner S, Kawauchi S, Milinkeviciute G, Gomez-Arboledas A, Neumann J, Banh CE, Huynh M, Matheos DP, Rezaie N, Alcantara JA, Mortazavi A, Wood MA, Tenner AJ, MacGregor GR, Green KN, LaFerla FM (2022) Systematic Phenotyping and Characterization of the 3xTg-AD Mouse Model of Alzheimer’s Disease. Frontiers in Neuroscience 15,. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Akiyama H, Arai T, Kondo H, Tanno E, Haga C, Ikeda K (2000) Cell Mediators of Inflammation in the Alzheimer Disease Brain. Alzheimer Disease & Associated Disorders 14, S47. [DOI] [PubMed] [Google Scholar]
  • [26].Yoshiyama Y, Higuchi M, Zhang B, Huang S-M, Iwata N, Saido TC, Maeda J, Suhara T, Trojanowski JQ, Lee VM-Y (2007) Synapse loss and microglial activation precede tangles in a P301S tauopathy mouse model. Neuron 53, 337–351. [DOI] [PubMed] [Google Scholar]
  • [27].Gómez-Nicola D, Fransen NL, Suzzi S, Perry VH (2013) Regulation of Microglial Proliferation during Chronic Neurodegeneration. J Neurosci 33, 2481–2493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Jurga AM, Paleczna M, Kuter KZ (2020) Overview of General and Discriminating Markers of Differential Microglia Phenotypes. Frontiers in Cellular Neuroscience 14,. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Eng LF, Ghirnikar RS (1994) GFAP and astrogliosis. Brain Pathol 4, 229–237. [DOI] [PubMed] [Google Scholar]
  • [30].Villamil-Ortiz JG, Cardona-Gomez GP (2015) Comparative analysis of autophagy and tauopathy related markers in cerebral ischemia and Alzheimer’s disease animal models. Front Aging Neurosci 7, 84. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [31].Esquerda-Canals G, Roda AR, Martí-Clúa J, Montoliu-Gaya L, Rivera-Hernández G, Villegas S (2019) Treatment with scFv-h3D6 Prevented Neuronal Loss and Improved Spatial Memory in Young 3xTg-AD Mice by Reducing the Intracellular Amyloid-β Burden. J Alzheimers Dis 70, 1069–1091. [DOI] [PubMed] [Google Scholar]
  • [32].Sasaguri H, Nilsson P, Hashimoto S, Nagata K, Saito T, De Strooper B, Hardy J, Vassar R, Winblad B, Saido TC (2017) APP mouse models for Alzheimer’s disease preclinical studies. The EMBO Journal 36, 2473–2487. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [33].Shekari A, Fahnestock M (2021) Chapter 13 - Cholinergic neurodegeneration in Alzheimer disease mouse models. In Handbook of Clinical Neurology, Swaab DF, Buijs RM, Kreier F, Lucassen PJ, Salehi A, eds. Elsevier, pp. 191–209. [DOI] [PubMed] [Google Scholar]
  • [34].Rohn TT, Vyas V, Hernandez-Estrada T, Nichol KE, Christie L-A, Head E (2008) Lack of Pathology in a Triple Transgenic Mouse Model of Alzheimer’s Disease after Overexpression of the Anti-Apoptotic Protein Bcl-2. J Neurosci 28, 3051–3059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [35].Wirths O, Zampar S (2020) Neuron Loss in Alzheimer’s Disease: Translation in Transgenic Mouse Models. Int J Mol Sci 21, 8144. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].Saul D, Kosinsky RL, Atkinson EJ, Doolittle ML, Zhang X, LeBrasseur NK, Pignolo RJ, Robbins PD, Niedernhofer LJ, Ikeno Y, Jurk D, Passos JF, Hickson LJ, Xue A, Monroe DG, Tchkonia T, Kirkland JL, Farr JN, Khosla S (2022) A new gene set identifies senescent cells and predicts senescence-associated pathways across tissues. Nat Commun 13, 4827. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [37].Dorigatti AO, Riordan R, Yu Z, Ross G, Wang R, Reynolds-Lallement N, Magnusson K, Galvan V, Perez VI (2022) Brain cellular senescence in mouse models of Alzheimer’s disease. Geroscience 44, 1157–1168. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [38].Goodwin LO, Splinter E, Davis TL, Urban R, He H, Braun RE, Chesler EJ, Kumar V, van Min M, Ndukum J, Philip VM, Reinholdt LG, Svenson K, White JK, Sasner M, Lutz C, Murray SA (2019) Large-scale discovery of mouse transgenic integration sites reveals frequent structural variation and insertional mutagenesis. Genome Res 29, 494–505. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [39].Zhang H, Chen Y, Wang Z, Xie G, Liu M, Yuan B, Chai H, Wang W, Cheng P (2022) Implications of Gut Microbiota in Neurodegenerative Diseases. Front Immunol 13, 785644. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [40].Kirkland JL (2023) ALSENLITE: An Open-Label Pilot Study of Senolytics for Alzheimer’s Disease, Identifier NCT04785300. https://clinicaltrials.gov/study/NCT04785300. [Google Scholar]
  • [41].Lipsitz L (2023) Senolytics To Alleviate Mobility Issues and Neurological Impairment in Aging, Identifier NCT05422885. https://clinicaltrials.gov/study/NCT05422885. [Google Scholar]
  • [42].Wake Forest University Health Sciences (2022) Phase II Clinical Trial to Evaluate the Safety and Feasibility of Senolytic Therapy in Alzheimer’s Disease, Identifier NCT04685590. https://clinicaltrials.gov/study/NCT04685590. [Google Scholar]
  • [43].Weller J, Budson A (2018) Current understanding of Alzheimer’s disease diagnosis and treatment. F1000Res 7, F1000 Faculty Rev–1161. [Google Scholar]
  • [44].Jack CR, Knopman DS, Jagust WJ, Shaw LM, Aisen PS, Weiner MW, Petersen RC, Trojanowski JQ (2010) Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol 9, 119–128. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

SuppFile3

Supplementary file 3. GO terms of pathways and associated genes of functional annotation analyses in Figures 4C and 4D.

NIHMS1968907-supplement-SuppFile3.xlsx (46.2KB, xlsx)
SuppFile2

Supplementary file 2. Leading edge subsets of GSEA pathways in Figure 4B.

NIHMS1968907-supplement-SuppFile2.xlsx (28KB, xlsx)
SuppFile1

Supplementary file 1. Leading edge subsets of GSEA pathways in Figure 4A.

NIHMS1968907-supplement-SuppFile1.xlsx (51.6KB, xlsx)
SuppFile4

Supplementary file 4. Gene names and z scores of heatmaps in Supplementary Figures 3 and 4.

NIHMS1968907-supplement-SuppFile4.xlsx (216.3KB, xlsx)
5
NIHMS1968907-supplement-5.pdf (4MB, pdf)

Data Availability Statement

The gene expression dataset analyzed in this study is available in the NCBI Gene Expression Omnibus repository (https://www.ncbi.nlm.nih.gov/geo/) under the accession number GEO: GSE231332. Other data supporting the findings of this study are available on request from the corresponding author.

RESOURCES