LIVER EXERKINE REVERSES AGING- AND ALZHEIMER’S-RELATED MEMORY LOSS VIA VASCULATURE

Gregor Bieri; Karishma JB Pratt; Yasuhiro Fuseya; Turan Aghayev; Juliana Sucharov; Alana M Horowitz; Amber R Philp; Karla Fonseca-Valencia; Rebecca Chu; Mason Phan; Laura Remesal; Shih-Hsiu J Wang; Andrew C Yang; Kaitlin B Casaletto; Saul A Villeda

doi:10.1016/j.cell.2026.01.024

. Author manuscript; available in PMC: 2026 Apr 13.

Published in final edited form as: Cell. 2026 Feb 18;189(5):1499–1516.e25. doi: 10.1016/j.cell.2026.01.024

LIVER EXERKINE REVERSES AGING- AND ALZHEIMER’S-RELATED MEMORY LOSS VIA VASCULATURE

Gregor Bieri ^1,^#, Karishma JB Pratt ¹, Yasuhiro Fuseya ¹, Turan Aghayev ¹, Juliana Sucharov ^1,², Alana M Horowitz ^1,², Amber R Philp ¹, Karla Fonseca-Valencia ^1,³, Rebecca Chu ^1,², Mason Phan ¹, Laura Remesal ¹, Shih-Hsiu J Wang ^4,⁵, Andrew C Yang ^6,⁷, Kaitlin B Casaletto ^7,⁸, Saul A Villeda ^1,^2,^3,^9,^10,^11,^#

PMCID: PMC13070421 NIHMSID: NIHMS2148338 PMID: 41713415

SUMMARY

Blood factors transfer the benefits of exercise to the aged brain, independent of physical activity. Here we show that liver-derived exercise factor (exerkine) glycosylphosphatidylinositol (GPI)-specific phospholipase D1 (GPLD1), a GPI-degrading enzyme, reverses aging- and Alzheimer’s-related memory loss by targeting brain vasculature. GPLD1 has potential to cleave over 100 putative GPI-anchored proteins, necessitating identification of downstream targets that mediate cognitive rejuvenation for translational application. We identified GPI-anchored tissue-nonspecific alkaline phosphatase (TNAP) on brain vasculature as a GPLD1 substrate. Mimicking age-related increase in cerebrovascular TNAP impaired blood-brain transport and cognition in young mice and mitigated GPLD1-induced cognitive benefits in aged mice. Inhibiting TNAP recapitulated benefits of GPLD1 in old age, restoring youthful hippocampal transcriptional signatures and rescuing cognition. In an Alzheimer’s disease model, increasing GPLD1, or inhibiting TNAP, ameliorated Aβ pathology and improved cognitive deficits. We thus identify brain vasculature as a mediator of the cognitive benefits of a liver-to-brain exercise axis.

Keywords: aging, rejuvenation, Alzheimer’s disease, exercise, GPLD1, liver, vasculature, blood-brain barrier, blood factors, cognition, memory

In brief statement

Liver exercise factor GPLD1 rejuvenates blood-brain barrier integrity and reverses cognitive impairments in aging and Alzheimer’s disease models by targeting GPI-anchored proteins on brain endothelial cells.

Graphical Abstract

graphic file with name nihms-2148338-f0008.jpg

INTRODUCTION

Exercise can reverse broad cellular and molecular hallmarks of brain aging, challenging long-standing views of age-related cognitive dysfunction as a rigid process^1–3. In the hippocampus – a brain region important for memory and highly sensitive to the detrimental effects of aging – running in aged mice increases regenerative capacity, enhances synaptic plasticity, attenuates neuroinflammation and boosts associated cognitive function^2,4–7. In animal models of Alzheimer’s disease (AD) pathology, exercise has been shown to improve learning and memory⁸. In humans, increased physical activity is associated with reduced risk of dementia^9,10, slowed cognitive decline with age¹¹, and delayed dementia onset¹⁰, even in autosomal dominant forms of AD¹². Unfortunately, its consistent application is often impeded in the elderly by physical and medical limitations^13–15. Therefore, identifying alternative means to confer cognitive benefits of exercise without physical activity may provide a unique and unexplored therapeutic approach.

We and others recently demonstrated that systemic administration of blood plasma from exercised to sedentary mice can transfer the benefits of exercise on the aged brain, independent of physical activity^16–18. We identified Glycosylphosphatidylinositol (GPI) Specific Phospholipase D1 (GPLD1), a liver-derived GPI-degrading enzyme, as an exercise-induced circulating blood factor (exerkine)¹⁶. While we demonstrated striking cognitive benefit of GPLD1 administration in aged mice¹⁶, GPLD1 has potential to cleave over 100 putative GPI-anchored proteins (GPI-APs). Therefore, to pursue the therapeutic potential of liver-derived GPLD1 in brain aging, it is important to identify downstream cellular and molecular targets that mediate cognitive rejuvenation for translational application to dementia-related neurodegenerative disorders, such as AD. Interestingly, increasing systemic GPLD1 improved cognitive function without readily entering the brain¹⁶, highlighting yet unidentified peripheral mechanisms of action.

To investigate factors downstream of GPLD1, we used bioinformatics and targeted candidate-based testing to identify GPI-anchored tissue-nonspecific alkaline phosphatase (TNAP) as a GPLD1 substrate on brain vasculature that is strongly upregulated during aging. Mimicking the age-related increase in cerebrovascular TNAP expression impaired blood-brain barrier (BBB) function and cognition in young mice and mitigated GPLD1-induced cognitive benefits in aged mice, whereas inhibiting TNAP activity recapitulated the transcriptional and cognitive effects of GPLD1 administration in aged mice. We found liver GPLD1 increased following exercise in a 5xFAD mouse model of AD pathology and TNAP was elevated in the brains of older adults and AD individuals compared to healthy young humans. Moreover, increasing liver-derived GPLD1, or inhibiting TNAP activity, reversed AD-related hippocampal transcriptional signatures, ameliorated Aβ pathology, and improved cognitive deficits. These findings indicate that brain vasculature is an important mediator of the restorative cognitive benefits of liver-derived exercise blood factors.

RESULTS

GPI-anchored TNAP is a GPLD1 substrate on the aged hippocampal vasculature

To begin, we first validated increased GPLD1 expression in the liver of aged mice in response to a voluntary exercise intervention compared to sedentary controls (Figure S1A-E). We mimicked the exercise-induced increase of GPLD1 in the liver of aged mice using an in vivo hydrodynamic tail vein injection (HDTVI)-mediated overexpression approach. We assessed health metrics and hippocampal-dependent memory using novel object recognition (NOR) and Y-maze behavioral tasks (Figure S1F-L). Young control mice were biased towards a novel object and a novel arm relative to a familiar condition during NOR and Y-maze testing, respectively, while aged control mice showed no preference (Figure S1G,H). However, age-related cognitive impairments were reversed in GPLD1-treated aged mice (Figure S1G,H). To identify potential downstream peripheral cellular and molecular targets mediating these cognitive rejuvenating effects of liver-derived GPLD1, we focused our analysis on candidate GPI-APs. To begin, we leveraged publicly available single cell RNA sequencing (scRNAseq) datasets ¹⁹²⁰and surveyed cellular expression of 148 putative GPI-APs across tissues (Figure 1A, Table S1)¹⁹,²⁰. We observed a high concentration of GPI-APs expressed on endothelial and epithelial cells (Figure 1A), positing vasculature as a potential GPLD1 target at the interface of the blood and brain.

Figure 1. — A, Pie charts of murine GPI-anchored protein (GPI-AP) expression by major cell categories (left). Differential GPI-AP expression in mouse brain endothelial cells (BEC) (right; adapted from Yousef et al., 2019). B, Volcano plot of GPI-APs in young and aged mouse BECs (adapted from Yousef et al., 2019). C, Expression of the potential GPLD1 substrate ALPL/Tissue non-specific Alkaline phosphatase (TNAP) on BECs. D, Representative images and quantification of TNAP labelling (red) in the hippocampal dentate gyrus (DG) region of young (3 month) and aged mice (22–24 month) (n=4 mice/group). E, Representative images and quantifications of endogenous Alkaline phosphatase (AP) labelling in the DG (dashed line, stitched overview image) of young and aged mice (n=3 mice/group). F, Aged mice (19–21 month) underwent a 6-week-long voluntary exercise intervention. G, Representative images and quantification of TNAP immunostaining in the DG of aged exercised mice (n=5–6 mice/group). H, Representative images and quantifications of endogenous AP labelling in the DG of aged exercised mice (n=9–11 mice/group). **I,J**, Schematic illustration (I) and quantification (J) of TNAP cleavage and secreted AP (SEAP) assay in the culture media of TNAP reporter cells following treatment with mouse or human GPLD1, catalytically inactive GPLD1 with the histidine H133N (H133N) or histidine H158N (H158N) amino acid substitution, or GFP control (n=6 wells/condition). K, Aged mice (20–22 month) with liver overexpression of GPLD1, H133N or GFP control using a hydrodynamic tail-vein injection approach. L, Representative images and quantification of TNAP immunostaining in DG of aged mice following liver GPLD1, H133N or GFP overexpression (n=6–7 mice/group). M, Representative images and quantification of endogenous AP labelling in the DG of aged mice following liver GPLD1, H133N or GFP overexpression (n=12–14 mice/group). Scale bar 200um. Data shown as mean+/− s.e.m. Statistical analysis was performed using t-test (**D, E, G, H**), ANOVA with Šidák’s post hoc test (**J, L, M**); *,p<0.05, **,p<0.01 ***,p<0.001, ****,p<0.0001.

We next analyzed age-related GPI-AP expression changes on brain endothelial cells (BECs) using a publicly available RNA-Seq dataset of young and aged mice^21,22. We identified 11 up-regulated and 2 down-regulated genes – of which alkaline phosphatase, biomineralization associated (Alpl) (encoding TNAP) was among the most significantly increased with age (Figure 1A,B). While its canonical function is still incompletely understood, TNAP is one of four alkaline phosphatase isoenzymes whose canonical role is to promote tissue mineralization and vascular calcification^23–25. It was recently shown that age-related increased cerebrovascular TNAP impairs blood-brain transport, while TNAP inhibition restores youthful blood-brain transport²⁶. Consequently, we elected to investigate whether vascular TNAP is a GPI-anchored GPLD1 substrate.

In the hippocampus, we found an age-related increase in TNAP vascular expression and alkaline phosphatase (AP) activity staining (Figures 1C-E, S1M,N). Similar increases were observed in additional brain regions (Figure S2A,B). Additionally, we observed a decrease in TNAP expression and AP activity in the cerebrovasculature of exercised compared to sedentary aged mice (Figures 1F-H), while Alpl mRNA expression in the hippocampus remained unchanged (Figure S1O).

To investigate the possibility of TNAP being a GPI-anchored GPLD1 substrate, we opted to first leverage an in vitro approach. The catalytic activity of GPLD1 is dependent on His133²⁷, with His133→Asn (H133N) or His158→Asn (H158N) amino acid substitutions ablating its enzymatic activity, which we validated using a cleavage reporter assay of the known GPLD1 substrate ALPP (Figure S2F)^16,28. We next generated a constitutively expressing TNAP reporter cell line in which cleavage can be assessed by measuring the AP activity of TNAP in the supernatant upon its cleavage and release from the plasma membrane (Figure 1I). TNAP expressing cells were treated with mouse or human GPLD1, catalytically inactive H133N or H158N GPLD1 or GFP control (Figure 1I, S2G). We detected a significant increase in TNAP-mediated AP activity in the supernatant from cells treated with mouse GPLD1 or human GPLD1, but not catalytically inactive H133N or H158N GPLD1 or GFP controls (Figure 1J, S2G), establishing TNAP as a direct GPI-anchored GPLD1 substrate.

To investigate whether vascular TNAP is a GPI-anchored GPLD1 substrate in vivo, we assessed changes in vascular TNAP expression and AP activity in the hippocampus of aged mice following increased liver-derived GPLD1 (Figure 1K). Aged male mice were given HDTVI with expression constructs encoding GPLD1, catalytically inactive H133N GPLD1 or GFP and TNAP expression and AP activity were assessed on hippocampal vasculature (Figure 1K-M). Increased liver-derived GPLD1 resulted in decreased vascular TNAP expression and AP activity staining in the hippocampal vasculature compared to the catalytically inactive H133N GPLD1 or GFP control treatment groups (Figure 1L,M). Liver-derived GPLD1 also decreased AP activity staining in additional brain areas surveyed (Figure S2C,D). Of note, we observe that GPLD1 treatment does not alter the mRNA expression of Alpl in the aged hippocampus (Figure S1P), indicating that the decrease in TNAP protein expression observed in GPLD1-treated aged mice is not due to changes at the mRNA level. Additionally, we orthogonally measured levels of circulating TNAP in blood plasma at baseline and following acute expression of liver-derived GPLD1 in an independent cohort of aged mice and observed an increase in circulating TNAP (Figure S2H,I). Given that circulating TNAP by nature must be derived from the plasma membrane protein that was cleaved, increased blood levels following GPLD1 treatment provide further evidence of TNAP cleavage in vivo. Collectively these data indicate that cerebrovascular TNAP is a GPI-anchored GPLD1 substrate.

Increasing liver-derived GPLD1 rejuvenates BBB function in the aged hippocampus

Having identified TNAP as a GPLD1 substrate and given the role of increased cerebrovascular TNAP expression in blood-brain transport dysfunction in aging, we next examined the effect of increasing liver-derived GPLD1 on BBB function in the aged hippocampus (Figure 2A). Aged mice were given retro-orbital injections of liver-specific Adeno-associated virus (AAV) encoding GPLD1 or catalytically inactive H133N GPLD1 (Figure 2A, S2E). To establish a baseline for age-related changes in BBB function^26,29,30, an additional control group of young mice given retro-orbital injections with liver-specific AAV encoding H133N GPLD1 was included (Figure 2A). We assessed GPLD1 expression across several highly vascularized tissues and observed selective increase only in the liver, further validating the specificity of the AAV-delivery approach (Figure S2E).

To evaluate BBB permeability, a subset of animals was systemically administered the small molecule tracer NHS-biotin (Figure 2B). Consistent with age-related dysfunctional blood-brain transport³⁰, we detected numerous hotspots of NHS-biotin leakage outside of blood vessels in the hippocampi of aged compared to young control animals (Figure 2B). However, increased liver-derived GPLD1 resulted in significantly reduced NHS-biotin leakage from blood vessels broadly across the aged hippocampus (Figure 2B).

In the aging cerebrovasculature a shift from receptor-mediated transport to caveolar transcytosis has also been reported²⁶. To assess receptor-mediated transport, a subset of animals was systemically administered with fluorescently labeled canonical ligand transferrin (TF-647) (Figure 2A,C). An age-related decrease in TF-647 transport was observed in the hippocampus of aged compared to young control animals, however, this decrease was in part blunted in the aged GPLD1 treatment group (Figure 2C). Next, we examined ligand-non-specific caveolae by measuring expression of Caveolin-1 (Cav1), a structural protein of caveolae that increases with age^26,29. While increased Cav1 expression was detected on blood vessels from the hippocampi of aged compared to young control animals, Cav1 expression was significantly reduced in aged animals with increased liver-derived GPLD1 (Figure 2D). Collectively, these data demonstrate that increased liver-derived GPLD1 improves BBB function in the aged hippocampus.

To gain mechanistic insight into the molecular changes induced in BECs by increased liver-derived GPLD1, we used scRNAseq. BECs were isolated from hippocampi of aged mice expressing liver-derived GPLD1 or H133N GPLD1, or young mice expressing liver-derived H133N GPLD1 (Figure 2E). Arterial, capillary and venous clusters were identified (Figures 2F, S3A-E). We detected prominent transcriptional changes in BECs due to aging (Figure 2G), of which over thirty-eight percent were restored towards a youthful profile following increased liver-derived GPLD1 (Figure 2G). Subsequently, we focused analysis on DEGs that change in aging but are rescued following GPLD1 treatment (Figure 1G, S3F-H). Gene ontology (GO) analysis of DEGs identified biological processes related to inflammatory processes, energy metabolism, and proteostasis (Figure 2H). These single-cell transcriptomics data indicate that increasing liver-derived GPLD1, in part, restores a more youthful transcriptional signatures in BECs from the aged hippocampus.

Mimicking the age-related increase in cerebrovascular TNAP disrupts hippocampal BBB function and impairs cognition

While increased TNAP expression on the aging cerebrovasculature has been shown to impair proper blood-brain transport to the aged brain²⁶, its impact on cognition has yet to be explored. Therefore, we assessed the functional consequence of mimicking an age-related increase in cerebrovascular TNAP expression on BBB function and hippocampal-dependent memory in young mice using a cell-type-specific, viral-mediated overexpression approach³¹. Young mice were given retro-orbital injections of BEC-specific AAV encoding TNAP or mRuby2 (Ruby) as a control (Figure 3A, S4A-C). Increased TNAP expression and AP activity was confirmed on hippocampal vasculature following AAV-TNAP injection (Figure 3B, S4C-E).

BBB function was assessed using NHS-Biotin tracer and fluorescently labeled transferrin (Figure 3A). Young animals overexpressing cerebrovascular TNAP exhibited increased NHS-biotin leakage from blood vessels and decreased TF-647 transport in the hippocampus compared to young control animals (Figure 3C,D). Concurrently, we observed increased Cav1 expression on blood vessels following TNAP overexpression (Figure 3E).

Hippocampal-dependent memory was assessed using NOR, Y-maze, and radial arm water maze (RAWM) behavioral paradigms (Figure 3F). During NOR testing, young control mice were biased towards a novel object relative to a familiar object, while young mice overexpressing cerebrovascular TNAP showed no preference (Figure 3G). No differences were observed during Y-maze testing between treatment groups (Figure 3H). In the training phase of the RAWM paradigm, all mice showed similar spatial learning ability (Figure 3I). However, young mice overexpressing cerebrovascular TNAP demonstrated impaired learning and memory for the platform location committing more errors during the testing phase of the task compared with young control mice (Figure 3I). No difference in health metrics was observed between groups (Figure S4F-L). These data indicate that increased cerebrovascular TNAP impairs BBB function in the hippocampus and is a negative regulator of object and spatial memory.

Restoring the age-related increase in cerebrovascular TNAP mitigates cognitive benefits of liver-derived GPLD1 in aging

We next examined whether cognitive benefits of increased liver-derived GPLD1 in aged mice could be mitigated by restoring the age-related increase in cerebrovascular TNAP expression. Aged mice were given retro-orbital injections of liver-specific AAV encoding GPLD1 and BEC-specific AAV encoding either TNAP or Ruby control (Figure 4A). Aged mice given retro-orbital injections with liver-specific AAV encoding catalytically inactive H133N GPLD1 and BEC-specific AAV encoding Ruby were included as a negative control (Figure 4A). The observed GPLD1-mediated decrease in vascular TNAP AP activity was abrogated following viral-mediated overexpression of cerebrovascular TNAP in aged mice (Figure 4B). Hippocampal-dependent learning and memory was assessed using NOR, Y-maze and RAWM (Figure 4A-E). Aged mice with increased liver-derived GPLD1 and expressing vascular Ruby control exhibited a bias for the novel object and the novel arm, and committed fewer errors locating a hidden platform, compared to H133N GPLD1-treated aged mice expressing vascular Ruby (Figure 4C-E). However, GPLD1-mediated cognitive benefits observed in NOR and RAWM testing, but not those observed during Y maze testing, were abrogated in aged mice overexpressing cerebrovascular TNAP (Figure 4C-E). Similarly, TNAP expression abrogated some of the improvements in health metrics observed in GPLD1-treated mice (Figure S4M-S). These in vivo viral-mediated overexpression and behavioral data indicate that targeting GPI-anchored cerebrovascular TNAP regulates, in part, the cognitive benefits of increased liver-derived GPLD1 on object and spatial memory in aged mice.

Figure 4. — A, AAV-mediated liver overexpression of GPLD1 or inactive H133N GPLD1 in aged (22–24 month) mice in combination with brain endothelial cell (BEC)-targeted overexpression of TNAP or Ruby2. B, Representative images and quantification of endogenous Alkaline phosphatase (AP) activity labelling in the dentate gyrus region (DG, dashed line; stitched overview image) of the hippocampus (n=13–15 mice/group). C, Object recognition memory was measured using Novel object recognition (NOR) as percent time exploring the novel object (n=16–20 mice/group). D, Spatial working memory was measured in the Y-maze task as the discrimination index for the novel arm. (n=16–20 mice/group). E, Hippocampal-dependent spatial learning and memory was evaluated by Radial arm water maze (RAWM) as the number of errors committed while attempting to find a hidden escape platform (n=16–20 mice/group). F, BEC-TNAP abrogation in aged transgenic mice (20–22 month) following AVV-mediated delivery of Alpl or safe harbor control gRNAs. G, Representative images and quantification of endogenous Alkaline Phosphatase (AP) labelling in the DG (n=8 mice/group). H, Object recognition memory was assessed using NOR (n=14–15 mice/group) I, Spatial working memory was assessed using Y-maze (n=11–13 mice/group). J, Spatial learning and memory was evaluated by RAWM (n=14–15 mice/group). K, Aged mice (22–24 month) with liver overexpression of GPLD1, inactive H133N GPLD1 (Control), or systemic TNAP inhibition (TNAPi) via SBI-425 administered in specialized chow. L, Representative images of endogenous AP activity labelling in aged mouse brain sections treated with the TNAP inhibitor. M, Object recognition memory was assessed using NOR (n=15–17 mice/group) N, Spatial working memory was assessed using Y-maze (n=15–17 mice/group). O, Spatial learning and memory was evaluated by RAWM (n=15–17 mice/group). Data shown as mean+/− s.e.m. Scale bar 200um. Statistical analysis was performed using one-way ANOVA with Šidák’s post hoc test (B), t-test (G) one-sample t-test versus 50% (**C,H,M**) or 0 (**D, I, N**), two-way ANOVA with Šidák’s post hoc test (**E, J, O**); *,p<0.05, **,p<0.01, ***,p<0.001, ****,p<0.0001.

Targeting cerebrovascular TNAP reverses aging-related cognitive impairments

To determine the effect of selectively targeting the age-related increase in cerebrovascular TNAP expression on cognitive function, we generated aged temporally controlled BEC-specific conditional Alpl genetic knockout mice using a viral-mediated in vivo CRISPR-Cas9 approach. Aged inducible Cas9 transgenic mice were given retro-orbital injections with BEC-specific AAV encoding Cre and guide RNA sequences targeting Alpl or a safe harbor locus as a control (Figure 4F, S5A,B). Decreased AP activity was confirmed on hippocampal vasculature and additional brain regions following AAV injection (Figure 4G, S5C,D). Hippocampal-dependent learning and memory was assessed using NOR, Y-maze and RAWM (Figure 4F). Abrogation of cerebrovascular TNAP expression resulted in a bias for the novel object during NOR testing and fewer errors committed locating a hidden platform during RAWM testing compared to aged control mice (Figure 4H,J). While no differences were observed during Y-maze testing between groups, selective improvements in overall wellbeing metrics were observed in the conditional knockout mice (Figure 4I, Figure S5E-I).

Next, we compared the cognitive benefits of increased liver-derived Gpld1 in aged mice with the effect of inhibiting the age-related increase in vascular TNAP activity using a more therapeutically tractable pharmacological approach. Aged mice given HDTVI with expression constructs encoding catalytically inactive H133N GPLD1 were administered an orally bioavailable non-brain penetrant TNAP inhibitor (SBI-425)³² or vehicle control in the chow (Figure 4K). Aged mice injected with expression constructs encoding GPLD1 and given control food were included as a positive control for GPLD1-mediated cognitive benefits (Figure 4J). Decreased AP activity was confirmed on hippocampal vasculature following TNAP inhibitor (TNAPi) treatment (Figure 4L). Inhibition of TNAP activity resulted in cognitive improvements in object memory during NOR testing and spatial memory during RAWM testing compared to aged control mice, similar in magnitude to aged animals with increased liver-derived GPLD1 (Figure 4M, O). While aged GPLD1-treated mice exhibited improvements in working memory during Y-maze testing, no change was observed following TNAPi treatment (Figure 4N). Selective improvements in overall wellbeing metrics were detected in aged mice following inhibition of TNAP activity or with increased liver-derived GPLD1 (Figure S5J-P). Collectively, these behavioral data indicate that targeting vascular TNAP can in part rescue object and spatial memory at old age, comparable to the cognitive benefits of increased liver-derived GPLD1.

Inhibiting TNAP activity recapitulates hippocampal transcriptional signatures of liver-derived GPLD1 in aging

To gain mechanistic insight, we further compared the cellular and molecular changes induced in both hippocampal BECs and parenchyma of aged mice following vascular TNAP inhibition and increased liver-derived GPLD1 expression using scRNAseq and single nucleus RNA-sequencing (snRNAseq) (Figure 5A).

Figure 5. — A, Aged mice (22–24 month) with liver overexpression of GPLD1, inactive H133N GPLD1 (Control), or systemic TNAP inhibition (TNAPi) via SBI-425 administration in specialized chow. Hippocampal brain endothelial cells (BECs) were profiled using single cell RNA sequencing (scRNAseq). Parenchymal gene expression was assessed using single nucleus RNA sequencing (snRNAseq) of the hippocampus. B, Venn diagram (left) of differentially expressed genes (DEGs) in BECs for the GPLD1 (blue) and TNAPi comparisons (red). Pie chart (right) of 793 overlapping DEGs between GPLD1 and TNAPi comparisons separated by unidirectional versus bidirectional changes. C, Scatter plot of conserved DEGs in BECs for the GPLD1 (x-axis) and TNAPi comparisons (y-axis). D, GO terms of biological processes associated with conserved DEGs in BECs. E, Venn diagram (left) of number of DEGs in BECs including a young versus aged comparison. Pie chart (right) of the 530 rejuvenated DEGs which are restored towards a more youthful state in the GPLD1 and TNAPi comparisons. F, Dot plot of rejuvenated DEGs in BECs from GPLD1 and TNAPi treatment groups. G, UMAP projection of identified cell types in the hippocampal snRNAseq analysis. H, Venn diagrams (top left) and bar graphs (bottom) of parenchymal DEGs by cell type for GPLD1 (blue) and TNAPi comparison (red). Pie chart (top right) of 147 overlapping DEGs between GPLD1 and TNAPi comparisons (top) separated by unidirectional (purple) versus bidirectional changes. Overlapping unidirectionally changing DEGs between the GPLD1 and TNAPi comparisons referred to as the conserved signature. I, Scatter plot (top) of conserved parenchymal DEGs for the GPLD1 (x-axis) and TNAPi comparisons (y-axis). Bar graph (bottom) of conserved parenchymal DEGs by cell type. J, Top GO terms of biological processes associated with parenchymal DEGs for the GPLD1 comparison. K, Top GO terms of biological processes associated with parenchymal DEGs for the TNAPi comparison. L, Top GO terms of biological processes associated with conserved parenchymal DEGs. M, Heatmap of DEGs selected from synapse-related GO terms. N, Violin plots of conserved DEGs (Grik5, Camkv, Gabra2) between the GPLD1 or TNAPi comparisons. O, Representative images and quantification of C1q immunolabelling in the dentate gyrus (DG) region of the hippocampus (n=12–13 mice/group). P, Representative images and quantification of GFAP immunolabelling in the DG (n=12–13 mice/group). Data shown as mean+/− s.e.m. Statistical analysis was performed using linear regression (**C, I**), ANOVA with Šidák’s post hoc test (**O,P**), MAST analysis for Control versus GPLD1 treatment and Control versus TNAPi treatment comparisons (N); *,p<0.05, **,p<0.01, ****,p<0.0001.

First, we isolated BECs from hippocampi of aged mice following TNAPi treatment and detected prominent transcriptional changes compared to BECs from aged controls (Figure 5B, S3A-E). Subsequently, we focused our analysis on conserved DEGs following both GPLD1 and TNAPi treatments and observed a significant overlap, with 65% of DEGs in the GPLD1 treatment group conserved with the TNAPi group (Figure 5B, C). Gene ontology analysis of conserved DEGs identified inflammatory processes, energy metabolism, and proteostasis as shared biological processes between the two interventions (Figure 5D, S3H,I). Moreover, we observed an overlap of 530 DEGs between the aging, GPLD1 and TNAPi comparisons, of which 97.3% were restored to a more youthful level (Figure 5E). Several of the DEGs identified in our scRNAseq dataset, including Vcam1 and Ppia (Cyclophilin A) have previously been implicated in brain aging and vascular dysfunction (Figure 5F)^21,33,34. Collectively, these single-cell transcriptomics data indicate that increasing liver-derived GPLD1 as well as TNAP inhibition, in part, restores a more youthful transcriptional signatures in BECs from the aged hippocampus. Since interventions that restore brain vascular dysfunction have been shown to regulate neuroinflammatory, regenerative and cognitive function in the aged brain^21,29,30,35, we next surveyed gene expression changes of cell types residing in the brain parenchyma using a complementary snRNAseq approach.

Nuclei were isolated from hippocampi of aged mice expressing liver-derived GPLD1, aged mice treated with TNAPi, or control mice expressing H133N GPLD1 and analyzed by snRNAseq (Figure 5A,G). A total of 23 cell clusters were identified, and populations were compared across treatment groups (Figures 5H, S6A-D). We detected prominent transcriptional changes in neuronal populations and microglia in the hippocampus of aged mice following either TNAPi treatment or increased liver-derived GPLD1 expression compared to aged control treated mice expressing liver-derived H133N GPLD1 (Figure 5H,I). Of the transcriptional changes induced following TNAP inhibition, about 30% were conserved with changes induced by increased liver-derived GPLD1 (Figure 5H,I, S6E,F). We found that transcriptional responses in the hippocampus are encoded in different cell types with a combination of upregulated and downregulated genes with DEGs largely unique to cell type (Figure 5H, S6G-I). GO analysis of DEGs between TNAPi and control treatment groups identified synaptic plasticity-related biological processes, consistent with changes observed following increased liver-derived GPLD1 expression (Figure 5J,K, S6J,K).

Focused analysis on conserved DEGs that change following both TNAPi treatment and increased liver-derived GPLD1 compared to aged control treated mice expressing liver-derived H133N GPLD1 (Figure 5H,I) revealed that most of the conserved transcriptional responses were unique to cell type with microglia and excitatory neuronal populations (CA1/2/3) exhibiting largest changes (Figure 5I, S6E,I). GO analysis of DEGs identified synaptic plasticity- and behavior-related biological processes, further supporting the cognitive improvements we observed in the GPLD1 and TNAPi treatment groups (Figure 5J-N, Figure 4K-O). Since we observed significant gene expression changes in the microglia cluster, we next complemented transcriptomics data by assessing levels of C1q, a microglia-derived complement factor involved in synaptic pruning and known to increase with age³⁶. We observed reduced C1q expression in the aged hippocampus following both TNAPi treatment and increased liver-derived GPLD1 compared to the control group (Figure 5O). Due to technical limitations, astrocytes were not robustly captured in our snRNAseq approach; therefore, we assessed levels of astrogliosis in the aged hippocampus and detected decreased GFAP expression in both GPLD1 and TNAPi treatment groups (Figure 5P). Together, these data indicate that targeting TNAP activity can partially recapitulate the transcriptomic signature observed in the hippocampus of aged mice with increased liver-derived GPLD1.

Liver-derived GPLD1 increases with exercise and rescues hippocampal transcriptional signatures in a model of AD pathology

Aging drives vulnerability to dementia-related neurodegenerative disorders, such as AD³⁷. Correspondingly, we next explored the translational potential of increased liver-derived GPLD1 in ameliorating AD pathology and associated cognitive deficits using a transgenic AD mouse model. We selected the 5xFAD model expressing the human APP and PSEN1 transgenes with five AD-linked mutations within APP/PSEN1^8,38, which represents a transgenic model with relatively early onset of AD pathology and cognitive deficits observed prior to normal aging-related cognitive decline^8,38,39. We sought to utilize the fast-progressing nature of AD pathology in the 5xFAD model to delineate between the beneficial effects of Gpld1 that target aging-specific impairments in aged wildtype mice from those that target AD pathology in 5xFAD mice at young ages normally devoid of natural age-related cognitive impairments. We exercised mature 5xFAD mice for three months using voluntary wheel running (Figure 6A). Hippocampal-dependent cognitive function was assessed using NOR. Exercised 5xFAD mice spent significantly more time with the novel object compared to their sedentary counterparts (Figure 6B). We detected increased GPLD1 expression in the liver (the main source of circulating GPLD1)^16,40 of exercised versus sedentary 5xFAD mice (Figure 6C), of which GPLD1 expression was positively correlated with performance in the NOR task (Figure 6D). These data indicate exercise increases liver GPLD1 expression concurrent with cognitive improvements in mature 5xFAD mice.

Figure 6. — A, Adult 5xFAD mice underwent a 12-week-long voluntary exercise intervention, followed by behavioral testing (9–10 month). B, Object recognition memory was measured using Novel object recognition (NOR) as percent time exploring the novel object. C, Liver GPLD1 expression was measured by RT-qPCR analysis. D, Correlation of liver GPLD1 expression with cognitive performance in the NOR task in exercised and sedentary 5xFAD mice. E, Single-nucleus RNA-sequencing (snRNAseq) gene expression profiling of hippocampal parenchymal cells from 5xFAD mice or wildtype littermates with liver overexpression of GPLD1 or GFP control. F, UMAP projection of identified cell types in the hippocampal snRNAseq analysis. G, Venn diagrams, pie chart (top) and bar graphs (bottom) representations of differentially expressed genes (DEGs) across different cell types for the 5xFAD (WT versus 5xFAD control, navy blue) and GPLD1 comparison (5xFAD+GPLD1 versus 5xFAD+GFP; light blue). Bidirectionally changing DEGs (purple) reversed to the WT condition are referred to as the restoration signature. H, Scatter plot (top) of 504 overlapping DEGs between the 5xFAD (x-axis) and GPLD1 comparisons (y-axis). Bar graph (bottom) of the restoration DEGs by cell type. I, Top GO terms of biological processes for the 5xFAD comparison. J, Top GO terms of biological processes for the GPLD1 comparison. K, Top GO terms of biological processes associated with restored DEGs. L, Heatmap of DEGs selected from neurogenesis-related GO terms. M, Representative images and quantification of immunolabeling of the neural progenitor cell marker MCM2 (green) in the dentate gyrus (DG) region of the hippocampus (n=11–13 mice/group). N, Representative images and quantification of Doublecortin (DCX, red) immunolabelling in the DG (n=11–13 mice/group). O, Representative immunoblots and quantification of BDNF in hippocampal lysates (n=4–5 mice/group). P, Representative images and quantification of GFAP immunolabelling in the DG (n=6–13 mice/group). Data shown as mean+/− s.e.m. Scale bar 200um. Statistical analysis was performed using one-sample t-test versus 50% (B), t-test (C), linear regression (**D, H**), ANOVA with Šidák’s post hoc test (**M, N, O, P**); *,p<0.05, **,p<0.01.

To investigate the cellular and molecular changes elicited in the hippocampus of 5xFAD mice by increased liver-derived GPLD1, we performed snRNAseq analysis. Mature 5xFAD mice were given HDTVI with expression constructs encoding GPLD1 or GFP control (Figure 6E). Age-matched wildtype littermate control mice given HDTVI with expression constructs encoding GFP were included to assess baseline changes induced by AD pathology (Figure 6E). A total of 22 cell clusters were identified, and populations were compared across genotype and treatment groups (Figure 6F, S7A-D). We detected pronounced transcriptional changes in excitatory neuronal populations (DG and CA1) in the hippocampus of 5xFAD compared to wildtype control mice (Figure 6G), as well as following increased liver-derived GPLD1 expression (Figure 6G). Of the transcriptional changes detected in 5xFAD close to 30% were restored towards wildtype conditions in the GPLD1 treated 5xFAD group (Figure 6G,H, S7E,F). We found that transcriptional responses detected in the hippocampus are encoded in different cell types with a combination of upregulated and downregulated genes that are largely unique to cell type (Figure 6G, S7G,H). GO analysis of DEGs in 5xFAD mice and following GPLD1 treatment identified neurogenesis- and synaptic-related biological processes (Figure 6I-J, S7I,J).

Subsequently, we focused our analysis on DEGs that change in 5xFAD mice but are rescued following GPLD1 treatment compared to wildtype control conditions (Figure 6H, H, S7E,F). Most of the rescued transcriptional responses were unique to cell type with excitatory neuronal populations (DG and CA1) exhibiting largest changes (Figure 6H, S7E,K). GO analysis of DEGs identified neurogenesis- and synaptic-related biological processes (Figure 6K-L). Consistent with transcriptomics data, we observed increased adult neurogenesis (Mcm2-postive neural progenitors and Dcx-positive newly born neurons) and elevated levels of BDNF, a neurotrophic factor involved in synaptic plasticity, in the hippocampus of GPLD1 treated 5xFAD mice (Figure 6M-O). Additionally, we surveyed the microglia cluster and observed a decrease in the expression of several disease-associated microglia markers (DAMs) and inflammatory genes previously reported in mouse models of AD pathology⁴¹ in GPLD1 treated compared to control 5xFAD mice (Figure S7L). Lastly, we assessed levels of astrogliosis in the hippocampus and detected decreased GFAP expression (Figure 6P). These transcriptomics data indicate that increasing liver-derived GPLD1 rescues hippocampal transcriptional signatures induced in a mouse model of AD pathology.

Increasing liver-derived GPLD1 or inhibiting TNAP activity in mice ameliorates AD pathology and cognitive deficits

AD pathology in 5xFAD mice has been shown to be amenable to exercise, reducing Aβ burden and improving cognitive function following wheel running⁸. Therefore, we assessed hippocampal changes in Aβ pathology and cognitive deficits in 5xFAD mice overexpressing liver-derived GPLD1, 5xFAD mice expressing GFP, or wildtype littermate controls expressing GFP (Figure 7A). We observed decreased Thioflavin S-positive amyloid deposits and Aβ expression in GPLD1 treated 5xFAD mice compared to GFP control 5xFAD mice (Figure 7B-E). We examined APP processing in the hippocampus and cortex and observed a decrease in the C-terminal fragment with no changes in full length APP levels (Figure S8A-J), indicating changes in APP processing in GPLD1 treated 5xFAD mice. Additionally, we assessed wellbeing metrics, including nest-building scoring, and hippocampal-dependent memory by NOR and Y-maze and observed functional improvements in nest formation, object memory and working memory in 5xFAD mice following GPLD1 treatment (Figure 7F-H). In an independent cohort of mice, we assessed spatial learning and memory using the active place avoidance task and observed impairments in 5xFAD compared to WT littermate controls, that were, in part, restored in 5xFAD mice following increased liver-derived GPLD1 (Figure S8K-N). Functional benefits of increased liver-derived GPLD1 in 5xFAD mice were comparable to baseline levels observed in age-matched wildtype GFP control animals (Figure 7F-H, S8O-R).

Figure 7. — A, Cognitive behavioral testing of 5xFAD mice or wildtype littermates with liver overexpression of GPLD1 or GFP control (9–10 month). **B,C**, Representative images (B) and quantification (C) of Thioflavin S (ThioS) amyloid plaque staining in the dentate gyrus (DG) region of the hippocampus of 5xFAD mice overexpressing liver GPLD1 or GFP control (n=11–13 mice/group). D,E, Representative images (D) and quantification (E) of 6E10-immunolabelled amyloid plaques in the dentate gyrus (DG) region of the hippocampus (n=11–14 mice/group). F, Hippocampal-dependent Nest forming performance was scored from 1 (worst) to 5 (best) (n=11–15 mice/group) G, Object recognition memory was measured using Novel object recognition (NOR) as percent time spent exploring the novel object (n=24–27 mice/group). H, Spatial working memory was measured in the Y-maze task as the discrimination index for the novel arm (n=10–14 mice/group). I,J, Western blot analysis (I) and quantification (J) of TNAP in human cortical lysates from young, aged and donors with Alzheimer’s disease (AD) (n = 6 samples/group). J, Quantifications of TNAP normalized to GAPDH (left) or beta-Actin (ACTB, right). **K,L**, Representative images (K) and quantifications of TNAP immunostaining (L) and Alkaline Phosphatase (AP) activity labelling (M) in the DG (dashed line; stitched overview image) (n=5–15 mice/group). N, Cognitive behavioral testing and histological analysis of adult 5xFAD or WT littermate controls (11–12 month) following systemic TNAP inhibition (TNAPi) via SBI-425 administration in specialized chow. O,P, Representative images (O) and quantification (P) of ThioS staining in the DG (n=18–20). Q,R, Representative images (Q) and quantification (R) of 6E10 immunolabelled amyloid plaques in the DG (n=15–19 mice/group). S, Nest forming performance was assessed in TNAPi-treated 5xFAD mice (n=17–20 mice/group). T, Object recognition memory was assessed using the NOR task (n=19–20 mice/group). U, Spatial working memory was assessed using Y-maze (n=16–20 mice/group). Scale bar 200um. Data shown as mean+/− s.e.m. Statistical analysis was performed using t-test (**C, E, P, R**), ANOVA with Šidák’s post hoc test (**F, J, L, M, S**); one-sample t-test versus 50% (**G, T**) or 0 (**H, U**); *,p<0.05, **,p<0.01, ****,p<0.0001.

To determine whether GPI-anchored GPLD1 targets mediating cognitive rejuvenation could similarly ameliorate AD pathology, we first assessed TNAP expression in the brains of older adults and AD individuals and observed elevated levels compared to healthy young humans by Western blot analysis (Figure 7I,J). Next, we examined changes in TNAP expression and AP activity in the hippocampus of GPLD1 treated 5xFAD mice and observed a reduction following increased liver-derived GPLD1 compared to 5xFAD mice expressing GFP (Figure 7K-M). Consequently, we investigated the effect of targeting TNAP activity in 5xFAD mice on Aβ pathology and cognitive function. Mature 5xFAD mice were administered TNAPi or vehicle control in the food (Figure 7N). Age-matched wild type littermate control mice given control food were included to assess baseline changes due to AD pathology (Figure 7N). We assessed wellbeing metrics by nest building scoring and hippocampal-dependent memory by NOR and Y maze. Inhibition of TNAP activity in 5xFAD mice resulted in decreased Thioflavin S-positive amyloid deposits and Aβ expression (Figure 7O-R), and functional improvements in nest formation and object memory, but not working memory, compared to control conditions (Figure 7S-U; S8S-V). Collectively, these data indicate that increasing liver-derived GPLD1, or inhibiting activity of its GPI-anchored substrate TNAP, can ameliorate Aβ pathology and cognitive deficits in an AD mouse model.

DISCUSSION

Cumulatively, our data demonstrate that the liver-derived exercise factor GPLD1 reverses memory loss in aging and models of AD pathology by targeting GPI-anchored TNAP on the brain vasculature. From a translational perspective, these data further suggest significant therapeutic potential to ameliorate neurodegenerative disease pathology, similar to the benefits of exercise, but independent of physical activity.

Broad therapeutic approaches targeting cerebrovascular health during aging may be ideally situated to mitigate progression of AD-related cognitive dysfunction. The brain vasculature shows dramatic changes in aging, including reduction in vessel density and dysregulation of BBB function^26,42. Disruptions of cerebrovascular integrity are also now appreciated as cardinal features of AD as highlighted by whole genome, neuropathological, neuroimaging, and cerebrospinal fluid (CSF) studies in humans^43–46. In individuals along the AD continuum, CSF and neuroimaging markers of cerebrovascular and BBB injury are evidenced prior to overt dementia symptoms and correlate with cognitive and functional symptoms⁴⁴. Even autosomal dominant forms of AD that are thought to represent young-onset AD pathology show cerebrovascular injury up to a decade prior to symptom onset⁴⁷. Together, vascular disruption is an early phenomenon in aging and AD and strongly tracks with onset and progression of cognitive symptoms in AD patients. Excitingly, the brain vasculature is amenable to blood-based interventions. As such, treatment with young blood factors or targeting pro-geronic factors has been shown to promote vascular outgrowth and restore function without the need for direct delivery into the brain^{21,22,35,48,49}. In this context, GPLD1 emerges as an attractive therapeutic candidate to improve cognition, possibly via modulation of factors that influence brain vascular function. Indeed, increased activity of the GPLD1 substrate TNAP on the aging cerebrovasculature has been shown to promote disruptions in BBB function²⁶, which we now functional link to cognitive deficits in aging and mouse models of AD pathology. As such, cleavage of vascular GPI-anchored TNAP by GPLD1 points to proper transport across blood vessels to the brain as one potential cerebrovascular target by which to restore cognition in older adults and AD individuals. In fact, TNAP inhibition has been demonstrated to restore blood-brain transport in aged animals to youthful levels²⁶.

While current therapeutic approaches to treat AD are predominantly focused on targeting neuropathology within the central nervous system, our work investigating a liver-to-brain exercise axis bolsters the possibility of targeting complementary molecular factors in blood and peripheral tissues to treat dementia-related neurodegenerative diseases. Our studies have significant translational potential, identifying GPI-APs downstream of GPLD1 as potential therapeutic targets. However, it should be noted that while inhibition of vascular GPI-anchored TNAP in part recapitulates cognitive benefits of liver-derived GPLD1 in aging and models of AD pathology, the benefits of GPLD1 are likely the result of targeting multiple GPI-APs across different cell types. Indeed, GPLD1-mediated cognitive benefits on working memory observed in Y-maze testing were not mitigated following increased cerebrovascular TNAP expression in aged mice. Mechanistically, the benefits of downstream GPLD1 targets may work in concert with other known exercise-induced factors, such as irisin, IGF1, clusterin, PF4, and SEPP1^{2,3,17,50–56}. Therefore, it is important for future studies to elucidate whether each factor acts through convergent or divergent cellular targets and molecular pathways. Given that individuals appear to progress along different aging trajectories⁵⁷, it may be that successful development of future therapies aimed at restoring cognition at old age necessitate identification of convergent mechanisms whose activation may provide additive benefits of multiple exercise-induced blood factors.

More broadly, GPLD1 has recently been reported to increase in liver and plasma of mouse models known to extend lifespan, including growth hormone receptor knockout mice, Snell dwarfs, caloric restriction, and treatment with compounds such as rapamycin^58–60. In this context, our data raise the exciting possibility that the benefits of liver-derived GPLD1 may extend beyond cognitive rejuvenation and amelioration of neurodegenerative disease pathology in the aging brain to promote overall longevity at the organismal level.

Limitations of the study

Hippocampal snRNAseq and BEC scRNAseq analyses begin to dissect the mechanisms by which increased liver-derived GPLD1 promotes cognitive benefits in aging and AD mouse models. We note that, due to technical limitations, our nuclei isolation and sequencing approach did not capture sufficient numbers of astrocytes for snRNAseq analysis. In addition, transcriptional changes in certain cell populations, such as activated microglia, are often difficult to resolve with snRNA-seq⁶¹, and the pooling of samples limited our ability to evaluate inter-individual variability. Although we complement the transcriptomic data with immunohistochemical analysis and observe reduced astrogliosis across conditions, these data provide limited insight into the molecular changes occurring within these cell populations. Increased TNAP activity was observed in several brain regions beyond the hippocampus. Similarly, GPLD1 treatment as well as our viral-mediated expression approaches targeted cerebrovascular TNAP more broadly in the brain. While our analyses focused primarily on hippocampal functions, our data suggest that the effects of GPLD1 and TNAP manipulations may extend to additional brain regions that were not investigated. In the periphery, bioinformatics analysis also identified enrichment of GPI-APs on peripheral immune cells, implicating molecular changes involved in age-related increased systemic inflammation as additional potential downstream GPLD1 targets to improve cognition. We anticipate that these data will prompt mechanistic and translational investigations delineating the contribution of identified genes within the hippocampus and GPI-APs in the periphery in mediating cognitive benefits observed following GPLD1 treatment in both aging and disease.

RESOURCE AVAILABILITY

LEAD CONTACT

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Saul Villeda (saul.villeda@ucsf.edu).

MATERIALS AVAILABILITY

Plasmids generated in this study are available from the lead contact upon request.

DATA AND CODE AVAILABILITY

The single-nucleus and single-cell RNA-sequencing datasets are available at the Gene Expression Omnibus (GEO). Accession numbers: GSE304483 and GSE269061.
This paper does not report original code.
Additional details are available from the lead contact upon request.

STAR METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mouse strains.

The following mouse lines were used: C57BL/6J mice (Jackson Laboratory, strain # 000664), C57BL/6J aged mice (National Institutes of Aging), B6SJL-Tg(APPSwFlLon,PSEN1*M146L*L286V)6799Vas/Mmjax (5xFAD; Jackson Laboratory/MMRRC strain # 034840-JAX) and homozygous inducible Cas9 transgenic mice (B6;129-Gt(ROSA)26Sor^tm1(CAG-cas9^*,-EGFP)Fezh/J; Jackson Laboratory strain# 034840-JAX). All studies in wildtype C57Bl6 mice were done in young (3–6 month) or aged (18–24 month) mice. Young wildtype mice were acquired at 8 weeks and aged in-house under standard housing conditions. Male 5xFAD and WT litter mate controls were acquired at 6 weeks and aged in house or 5xFAD transgenic males were bred with wildtype female mice and aged in-house. 5xFAD mice were treated at 6–9 month with GPLD1 and 8–11 month with TNAP inhibitor followed by behavioral testing. Male homozygous Cas9 mice were bred and aged in-house and used for experimental purposes at 20–22 months. All sequencing studies were performed with male mice. For each age group and experiment, litter mates were randomly assigned to each treatment group. Mice were group housed until 7–10 days before the start of behavioral testing. Extra enrichment was added during single housing. Health was regularly assessed, and mice were weighed during the course of treatments and testing. The numbers of mice used to result in statistically significant differences were calculated using standard power calculations with a = 0.05 and a power of 0.8. We used an online tool to calculate power (https://www.stat.uiowa.edu/~rlenth/Power/index.html) and samples size based on experience with the respective tests, variability of the assays and inter-individual differences within groups. Mice were housed under specific pathogen-free conditions under a 12h light-dark cycle, and all animal handling and use was in accordance with institutional guidelines approved by the University of California San Francisco Institutional Animal Care and Use Committee (IACUC).

Human tissue.

Post-mortem brain tissues from young, confirmed AD and age-matched, non-demented, non-pathological controls were obtained from Duke ADRC in strict accordance with all ethical and institutional guidelines. Individuals were grouped by age and clinical diagnosis. All subjects were male and Caucasian. Autopsy brain tissue from the frontal cortex of 6 young donors, non-demented aged and neuropathologically confirmed AD case were studied. Age at death: Young: 27+/−5.2; Aged: 81+/−12; 85+/−6.6 years. The aged group had a mean MMSE score of 29+/−1.4, while AD was 18.17+/−5. The AD neuropathological changes in the AD group where classified as intermediate to high, with a Braak stage of III-V. Additional details in Table S1.

Cell lines.

293T and Neuro-2a cells were cultured under standard conditions using DMEM with 10% FBS in 37C incubators with 5% CO2.

METHOD DETAILS

Exercise intervention.

Exercised mice were single-housed and given continuous access to a running wheel (Med Associates Inc., Cat# ENV-047) in their home cage for 6 weeks. Distance ran per mouse was tracked using Wheel Manager software (Med Associates, Inc.). Sedentary control mice were single-housed and given a locked wheel, house and nestlet as alternative enrichments.

Liver GPLD1 expression.

GPLD1, catalytically inactive GPLD1 with a H133N amino acid substitution (H133N), or GFP control were overexpressed in the liver in vivo using a hydrodynamic tail vein injection or AAV-delivery approach. For hydrodynamic tail vein injections, endotoxin-free plasmids were prepared using an endotoxin-free Maxi-Prep Kit. All plasmid sequences were verified by whole plasmid sequencing and low endotoxin level validated prior to administration (Thermo Fisher cat# PIA39552). Mice were briefly placed in a restrainer and GPLD1, H133N or GFP plasmid DNA (50ug) was suspended in 3mL sterile saline and injected in the tail vein in 5–7 seconds. Alternatively, GPLD1, H133N or GFP were expressed using an AAV-based expression plasmid with a hepatocyte specific TBG promoter and packaged with the AAV8 capsid. Genomic titers were assessed using qPCR with ITR specific primers as viral genomes per volume (vg/mL). 10^15 vg/kg were injected retro-orbitally into each mouse.

TNAP inhibition.

The orally bioavailable TNAP inhibitor SBI-425 was sourced from Sigma-Aldrich (Cat# SML2935) and Cayman Chemicals (Cat# 34626). TNAP inhibitor and control chow were manufactured by Research Diets Inc using the open standard diet (Cat# D11112201) with a formulation of 200mg/kg SBI-425. All mice had ad libitum access to the chow.

AAV production –

AAV GPLD1, H133N and GFP plasmids were generated using the NEBuilder HiFi DNA Assembly Kit (NEB #E5520S), following the manufacturer’s recommended design considerations and protocol. Briefly, the AAV2 vector backbone containing a TBG promoter was digested with the restriction enzymes NotI and HindIII and gel purified. eGFP and murine GPLD1 coding sequences were PCR-amplified with primer sets that included a 20 nucleotide overlap with the backbone and preserved the restriction sites. The H133N mutation in GPLD1 was introduced using the QuikChange Lightning Site-directed Mutagenesis kit (Agilent Cat# 210518) in combination with the following mutagenesis primers: GCTGACGTGAGCTGGAATAGCCTGGGTATTG, CAATACCCAGGCTATTCCAGCTCACGTCAGC.

The TNAP/Alpl expressing AAV construct was based on the control pAAV-CAG-Ruby2 plasmid (a gift from Viviana Gradinaru; Addgene plasmid #99123)⁶² and pEMS1938 plasmid with the Ple261 promoter (a gift from E. Simpson; Addgene plasmid# 82563). The Alpl mRNA with partial 3’ and 5’ UTRs was PCR-amplified from a mouse hippocampal cDNA library and cloned using the pENTER/D-TOPO cloning kit (Thermo Fischer Scientific K240020). The Alpl coding sequence was further amplified with primers containing the restriction sites KpnI and EcoRI. Both Alpl and the AAV backbone were digested with KpnI and EcoRI followed by ligation. All plasmids were sequence-verified using whole-plasmid sequencing (Primordium Labs).

For large scale AAV production HEK293T cells were cultured in 10x T182 flasks per virus. HEK293T cells were transfected with PEI (1mg/mL; Polysciences Cat# 23966–1), a combination of the AAV backbone, capsid and helper plasmid and Opti-MEM (Thermo Fisher Scientific Cat# 31985–062). For liver-targeting viruses, the pAAV2/8 capsid was used (a gift from James M. Wilson; Addgene plasmid# 112864) and for the BEC-targeting viruses the PHP.V1 capsid was used (pUCmini-iCAP-PHP.V1 was a gift from Viviana Gradinaru; Addgene plasmid# 127847)³¹. Media change was performed after overnight incubation. Supernatant and cells were harvested after 72–96 hours for AAV viral particle isolation and concentration. Briefly, cells were mechanically detached and collected together with the viral particle containing culture media. The media was collected after 10 minute centrifugation at 1000g and the remaining pellet was resuspended in chilled PBS. Following 4 freeze-thaw cycles, the lysate was cleared using centrifugation at 10000g for 10 minutes. The supernatant was re-combined with the cell culture media and filtered through 0.45μm PES filter. AAV viral particles were purified and concentrated using Ultracentrifugation (Beckman Coulter). Genomic viral titers were measured using qPCR in combination with primer sets targeting the ITR sequence (FWD: GGAACCCCTAGTGATGGAGTT; REV CGGCCTCAGTGAGCGA)⁶³.

Abrogation of cerebrovascular TNAP in aged Cas9-inducible mice.

5 different Alpl-targeting gRNAs and a non-targeting safe harbor control guide were first cloned into the lentiCRISPRv2 plasmid (lentiCRISPR v2 was a gift from Feng Zhang; Addgene plasmid #52961)⁶⁴, packaged into lentiviral particles and used to transduce murine Neuro-2a cells. One week after puromycin selection (72 hours) and recovery, Alpl expression was assessed using RT-qPCR analysis. The U6 gRNA cassettes of Alpl gRNA1 (target sequence: ACGCGATGCAACACCACTCAGGG) and the control guide were then cloned into an AAV backbone expressing a CAG-CRE. AAV particles were packaged with BEC-targeting PhP.V1 capsid and administered retro-orbitally into aged transgenic mice (20–22 months) with a stop-lox-stop Cas9 transgene (JAX strain# 024857). Genomic titers were assessed using qPCR with ITR specific primers as viral genomes per volume (vg/mL). 10^15 vg/kg were injected retro-orbitally into each mouse. NHS-biotin, labelled Transferring, AP activity measurements and behavioral testing were performed 5–8 weeks after AAV administration.

Novel object recognition (NOR).

The NOR task was performed using an established protocol⁷. Specifically, on day one (the habituation phase), mice performed open field testing by exploring an empty arena (40cm × 40cm) for 10 min. Infrared photobeam breaks were recorded and movement metrics were analyzed using the MotorMonitor software (Kinder Scientific). On day two (the training phase), two identical objects were placed into the habituated arena, and the mice were allowed to explore for 5 min. On day three (the testing phase), one object was replaced with a novel object, and the mice were allowed to explore for 5 min. The time spent exploring each object was quantified using the Smart Video Tracking Software (Panlab; Harvard Apparatus). Two different sets of objects were used. To control for any inherent object preference, half of the mice were exposed to object A as their novel object and half to object B. To control for any potential object-independent location preference, the location of the novel object relative to the trained object was also counterbalanced. To determine the percentage of time with the novel object, we calculate (time with novel object)/(time with trained object + time with novel object) × 100. Mice that did not explore both objects during the training phase were excluded from the analysis.

Radial arm water maze (RAWM).

Spatial learning and memory were assessed using an established 8-arm radial water maze paradigm⁶⁵. In this task, the mouse was trained to the location of a constant goal arm throughout the training and testing phase. The start arm changed each trial. Entry into an incorrect arm was scored as an error, and errors were averaged over training blocks consisting of three consecutive trials. During training (day 1), the mice were trained for 12 trials (blocks 1–4), with trials alternating between a visible and hidden platform. After an hour break, learning was tested for 3 trials (block 5) using only a hidden platform. During testing (day 2), the mice were tested for 15 trials (blocks 6–10) with a hidden platform. When scoring, investigators were blinded to treatment and genotype.

Y-maze.

The Y-maze task was conducted using an established forced alternation protocol. During the training phase, mice were placed in the start arm facing the wall and allowed to explore the start and trained arm for 5 minutes, while entry to the 3rd arm (novel arm) was blocked. The maze was cleaned between each mouse to remove odor cues, and the trained arm was alternated between mice. The mouse was then returned to its home cage. After 30–45 minutes, the mouse was placed in the start arm and allowed to explore all 3 arms for 5 minutes. Time spent in each arm was quantified using the Smart Video Tracking Software (Panlab; Harvard Apparatus). Percent time in each arm was defined as time in arm divided by time spent in all arms during the first minute of the task. The learning index was calculated as (entries in novel arm − entries in familiar arm)/(entries in novel arm + entries in familiar arm).

Active Place Avoidance.

The mice were placed in a rotating cylindrical arena (Maze Engineers; 1 meter in diameter, rotation speed at 1 rotation/minute) and had to learn to avoid a stationary 60° area (shock zone) using visual cues placed around the arena. During the habituation day, the mice were placed in the rotating arena for a duration of 5 minutes during which the shock zone was inactive. During testing, mice were placed in the in the area diagonally away from the shock zone and recorded for 10 min/day for a total of 4 consecutive days. Between animals, the arena was cleaned with 70% ethanol. Shocks were delivered at 0.5 mA if the mice entered the shock zone for a minimum of 1 second, with an inter-shock latency of 1.5s. Trials were recorded, analyzed and shocks administered using Noldus Ethovision XT. Number of entries/day into the shock zone, the number of administered shocks/day and the cumulative number of shocks over all trials were calculated for each animal.

Health metrics.

The nestlet assay was assessed using a pre-defined scoring system⁶⁶. Mice were provided with two pressed cotton nestlets and given up to 48 hours to build nests. Nest forming behavior was measured based on a nestlet score with a scoring system of 1 = no nest built and a score of 5 = an enclosed nest, as preciously described. For each mouse, the nesting scores were averaged from two separate experiments performed at a one-week interval.

Motor coordination was evaluated with the RotarRod test using a standard mouse RotaRod apparatus with a 1-inch diameter rotating cylinder (Harvard Apparatus, PanLab). On the training day, mice were first habituated on the still cylinder for 30 seconds and then trained at a constant speed (5 rpm) for 5 minutes. The next day (testing day), mice were placed on the rod at a constant rotation (4 rpm) for 10 seconds, followed by an acceleration from 4 to 40 rpm in 300 seconds. The latency to fall was recorded for each mouse on 3 trials with a 15-minute intertrial interval.

Grip Strength was measured for all 4 paws using a standard Grip strength test apparatus with a wire grid (Bioseb). Mice were held by the tail, lowered towards the apparatus, and allowed to grab the metal grid before being pulled back horizontally. The maximum force applied to the grid was recorded and averaged for 6 trials per mouse.

Tissue collection.

Mice were anesthetized with 87.5 mg per kg ketamine and 12.5 mg per kg xylazine and transcardially perfused with 25ml ice-cold phosphate-buffered saline. For brain tissue used in histological analysis of the brain vasculature, transcardial perfusion with 25ml PBS was followed with 25ml ice-cold 4% PFA. For a subset of PBS-perfused animals the liver, heart, kidney, spleen, lung and tibialis anterior muscle were dissected and snap frozen. To process the brains, the whole brain was sectioned in half along the sagittal plane. The hippocampus and cortex from one hemisphere were subdissected and snap-frozen and the other was postfixed in phosphate-buffered 4% paraformaldehyde, pH 7.4 at 4 °C for 48 h before cryoprotection with 30% sucrose.

Single nucleus RNA sequencing.

Neuronal nuclei were isolated based on the demonstrated protocol by 10x Genomics with modifications and performed on nuclei isolated from two-four mice per group (pool of 3 mice/well – 1 well for 5xFAD experiment and H133N of Figure 5, 2 wells for GPLD1/TNAPi samples). Only male mice were used for snRNAseq analysis. Briefly, flash-frozen dissected hippocampi were dounce-homogenized (Wheaton, Cat# 357538) in 500μL of Nuclei EZ prep lysis buffer (Sigma-Aldrich, NUC101, 1x RNAse Inhibitor) with 20 strokes of the loose pestle and 20 strokes of the tight pestle. 500 μL of Nuclei EZ prep lysis buffer was added and samples incubated for 7 minutes on ice. Samples were filtered through a 40μm filter and centrifuged at 500 RCF for 5 min at 4°C. Samples were resuspended in 1ml of Wash Buffer (PBS, 1%BSA, 1xRNAse Inhibitor) and incubated for 5 min on ice. Samples were centrifuged at 500 RCF for 5 min at 4°C, supernatant was removed, and samples were resuspended in 400μL of Wash Buffer with 1:10,000 dilution of Hoescht 33342 and incubated for five minutes before filtering through a 35μm FACS tube filter and sorting. Nuclei were sorted on a BD FACSAria II with a 100μm nozzle and with a flow rate of 1–2.5. Nuclei were first gated by forward and side scatter, then gated for doublets with height and width. Nuclei that were Hoechst+ were sorted and samples were combined per group. Isolated nuclei were given to the UCSF-CoLab Genomics Core for preparation with the 10x Genomics Chromium Single Cell Expression Solution 3′ kit. The Genomics Core prepared cells for 10x Genomics Chromium single-cell capture. 30,000 nuclei were loaded per sample. cDNA libraries were prepared according to the standard 10x Genomics protocols. The final library pool was sequenced on the NovaSeq 6000 or the NovaSeq X 10B system at the UCSF CAT Core. The raw base sequence calls were demultiplexed into sample-specific cDNA files with bcl2fastq/mkfastq and converted to count matrices using Cell Ranger 7.1 (10x Genomics).

Brain endothelial cell isolation and single cell RNA sequencing.

Hippocampal brain endothelial cells were isolated using enzymatic tissue dissociation and fluorescent activated cell sorting⁶⁷ and single cell analysis was performed on pooled hippocampal BECs from a pool of hippocampi from eight mice per group (Young Control, Aged control, Aged GPLD1- and TNAP inhibitor treated). Brain endothelial cells derived from male mice were used for scRNAseq analysis. Mice were anesthetized with 87.5 mg per kg ketamine and 12.5 mg per kg xylazine and transcardially perfused with 30ml ice-cold phosphate-buffered saline. The entire brain was removed, then the hippocampus was sub-dissected. Single-cell suspensions were generated by enzyme-mediated (papain) and mechanical dissociation using Miltenyi Neural Dissociation Kits (P) (Miltenyi, 130–092-628) according to the manufacturer’s instructions. The papain dissociation was done at 37°C for 10 minutes with three trituration steps. All other processing steps were performed at 4°C. Myelin was depleted from the suspensions using 22% Percoll solution (Cytiva, 17089101). Cell Fc receptors were blocked using a purified anti-mouse CD16/32 antibody (BD Biosciences, 553142) and stained for 30 minutes at 4°C with CD45-PE (BD Biosciences, 553081), CD41-PE (BD Biosciences, 558040) and CD31-APC (BD Biosciences, 551262) and DAPI (Thermo Scientific, 62248 1:1000). Cells were sorted on a BD FACSAria Fusion with a 100 μm nozzle and with a flow rate of 1–2.5. Cells were first gated by forward and side scatter, then gated for doublets with area and height. Cells per sample that were DAPI−, CD45−, CD41−, CD31+ were collected for sequencing analysis. Isolated cells were given to the UCSF-CoLab Genomics Core for analysis with the 10x Genomics Chromium Single Cell Expression Solution 3′ kit. The Genomics Core prepared cells for 10x Genomics Chromium single-cell capture. 22,000–30,000 cells were loaded per sample. cDNA libraries were prepared according to the standard 10x Genomics protocols. The final library pool was sequenced on the NovaSeq 6000 or the NovaSeq X 10B system at the UCSF CAT Core. The raw base sequence calls were demultiplexed into sample-specific cDNA files with bcl2fastq/mkfastq and converted to count matrices using Cell Ranger 7.1 (10x Genomics).

Single cell and single nucleus RNA sequencing analysis.

Raw FASTQ files were processed and aligned to mm10 using the Cell Ranger software package (10x Genomics) for the RNA expression matrix, including introns for single nucleus analysis. For the aging GPLD1/TNAP treatment, a total of 54,597 nuclei were sequenced at a depth of approximately 70,000 reads per cell and 60% sequencing saturation. For the 5xFAD sequencing, a total of 52,783 cells were sequenced at a depth of approximately 40,000 reads per cell and 31% sequencing saturation. Further QC and analysis were performed on R 4.2.2. For all experiments, downstream analysis and data visualization was performed using Seurat, as well as packages DropletUtils, ggplot2, knitr, WriteXLS, RColorBrewer, data.table, stringr, ggplot2, forcats, dplyr, Nebulosa, and htmlwidgets. Data were processed to remove doublets and unwanted sources of variation by removing nuclei with more than 6,000 and fewer than 200 genes per cell, number of counts more than 40,000 per nuclei, and regressing on number of UMIs. Genes expressed in fewer than three cells or nuclei were filtered out. Nuclei with a percentage of mitochondrial genes higher than 0.2% were removed. Final nuclei counts of 11062 (Aged CTRL), 20,713 (Aged GPLD1), 19581(Aged TNAP inhibitor), 17350 (WT GFP), 14404 (5xFAD GFP), and 18000 (5xFAD GPLD1) were used for gene expression analysis. Final cell count of 16594 (Young Control BECs), 20576 (Aged Control BECs, 12605 (Aged GPLD1 BECs), 20892 (Aged TNAPi BECs). The matrices of data were log-normalized in a sparse data matrix, scaled, integrated with CCA integration, and PCA was applied to reduce dimensionality. The first 20 PCA components were used to cluster cells by Louvain clustering at a resolution of 0.4, implemented in Seurat while UMAP plots were independently generated to aid in 2D representation of multidimensional data independent of the clustering. Cell types were identified using known markers, as well as the Allen Mouse Brain Atlas^4,5, and clusters containing more than one cell type specific marker were removed. Differential gene expression was performed on each cell type down-sampled to 800 nuclei/cluster/condition for the aging dataset and 550 nuclei/cluster/condition for the 5xFAD dataset and 12,000 cells/condition for the BEC dataset. Differential gene expression was determined using MAST statistical testing with a minimum of 10% of nuclei expressed, log fold change threshold of 0.15, and a pseudocount of 0.1. Log-normalized gene expression data were used for visualizations with violin plots and UMAP feature plots. Average expression matrices were used for heatmap visualization. Volcano plots were created using the EnhancedVolcano package and UpSet plots with the ComplexHeatmap package.

RNA extraction, cDNA synthesis and RT-qPCR analysis.

Total RNA was isolated from subdissected hippocampi, cortex, liver, kidney, spleen, lung, heart and tibialis anterior muscle tissue using TRI Reagent (Sigma-Aldrich, Cat#T9424) in combination with the PureLink RNA Mini Kit (Thermo Fisher Scientific Cat# 12183025) or Direct-zol^™ RNA Purification Miniprep Plus Kit (Zymo Research Cat# R2072) following the manufacturer’s protocol. 50–100mg tissue was dissociated using a Bead Ruptor Elite and Ceramic beads (Omni International Cat# 19–645-3). Lysates were centrifuged at 8,000 g for 10 min at 4 °C to remove cellular debris. Cell lines were lysed in the tissue culture plate using the Lysis reagent provided in the PureLink RNA Mini kit. The RNA concentrations were determined via Nanodrop and RNA was reverse-transcribed using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Cat# 4368813). To quantify mRNA expression levels, equal amounts of cDNA were synthesized using the High-Capacity cDNA Reverse Transcription kit (ThermoFisher Scientific, Cat# 4368813), then mixed with SYBR Fast mix (Kapa Biosystems) and primers. Gapdh mRNA was amplified as an internal control. Quantitative RT-PCR was carried out in a CFX384 Real Time System (Bio-Rad). Each sample and primer set were run in triplicates and relative expression levels were calculated using the 2−ΔΔct method. The following primer sets were used:

Gapdh: GGGTGTGAACCACGAGAAAT, ACTGTGGTCATGAGCCCTTC

Gpld1: GGAAGCAGAGAGGAATTGTGGC, TCCAAACCACGAGAAGTCCTCC

Alpl Pair 1: GAACAGACCCTCCCCACGAG, GTGCCGATGGCCAGTACTAA

Alpl Pair 2: TAACACCAACGCTCAGGTCC, TGGATGTGACCTCATTGCCC

Western blot analysis.

Brain tissue was lysed in chilled RIPA lysis buffer (Abcam Cat# ab156034) with complete protease inhibitor (Sigma-Aldrich Cat# 4693116001) and phosphatase inhibitor (Thermo-Fisher Cat# 78420). Tissue was dissociated using a Bead Ruptor Elite and Cermic beads (Omni International Cat# 19–645-3). Crude lysates were centrifuged at 10,000 g for 10 min at 4 °C to remove cellular debris. Protein concentrations in clarified lysates were quantified with a Pierce BCA protein assay. Tissue lysates were mixed with 4x NuPage LDS loading buffer (Invitrogen Cat# NP0008), loaded on a 4–12% SDS polyacrylamide gradient gel (BioRad Cat# 11346–02) and transferred onto a nitrocellulose or PVDF membrane using the Trans-Blot turbo transfer system (BioRad). For plasma samples, plasma was mixed with 4x NuPage LDS, water and 2-Mercaptoethanol and the 1 ul of plasma was loaded into each well. After transfer, some membranes were cut into strips at approximately 25kDa and 75kDa to probe for proteins with distinct molecular weight in parallel. The blots were blocked in 5% milk in Tris-Buffered Saline with 0.1% Tween (TBST) and incubated with anti-GAPDH (Abcam Cat# ab8245), anti-ALPL (R&D System Cat# MAB29092) and anti-ACTIN-Beta/ACTB (Abcam Cat# ab49900). APP protein and processing components were assessed using the following antibodies: anti-FL APP (Synaptic Systems Cat# 127 005), anti-APP 6E10 (human FL-APP and CTF detection; BioLegend Cat# 803001), anti-PS1 (Cell Signal Technology Cat# 5643S); anti-BACE1 (Sigma/Millipore Cat# MABN2640), anti-BIII-Tubulin (BioLegend Cat# 801201) primary antibodies⁶⁸. Horseradish peroxidase-conjugated secondary antibodies and an ECL kit (BioRad Cat# 1705060) were used to detect protein signals. Membranes were imaged with a ChemiDoc Imaging System (BioRad). Selected images were exported and quantified using the built-in gel analysis tool in FIJI/ImageJ. GAPDH, ACTB and BIII-Tubulin bands were used for normalization.

Immunohistochemistry.

Tissue processing and immunohistochemistry were performed on free-floating sections according to standard published techniques⁶. Cryoprotected brains were sectioned coronally at 40 μm with a freeze-stage microtome (Leica Camera). Free-floating sections were permeabilized with pre-treatment buffer (0.2% TritonX-100 in TBST) for 30 minutes, then washed 3x with TBST and blocked with TBST + 3% Normal Donkey serum (NDS). Sections were then incubated overnight at 4C with anti-ALPL (R&D Systems Cat# AF2910), anti-CD31 (BioLegend Cat# 160202 and R&D Systems Cat# AF3628), anti-AQP4 (Cell Signal cat# 59678S and Millipore cat# AB2218 or 59678S), anti-APP 6E10 (BioLegend Cat# 803001), anti-Caveolin1 (Cell Signal Technology Cat# 3267S) in TBST + 3% NDS. Labelling was revealed using secondary antibodies at 1:500 in TBST + 3% NDS for 1 hour room temperature. Nuclei were labelled with Hoechst (Thermo Fisher Scientific, Cat# H3570). Sections were mounted on Superfrost Plus microscopy slides and coverslipped with Prolong Gold. Sections were imaged using confocal microscopy (Zeiss LSM800 or Zeiss LSM900) or bright-field microscopy (Keyence). Labelling intensity and thresholded areas were quantified and averaged for 3–5 hippocampal sections for each mouse using FIJI and Zeiss Zen image analysis tools.

Sulfo-NHS-biotin BBB leakage assay.

Sulfo-NHS-LC-biotin (Thermo Fisher Cat# 21335) was reconstituted in sterile PBS pH7.4, stored on ice and used for up to 30 minutes before being discarded. Mice were injected retro-orbitally at a dose of 0.25mg/g BW (~7.5mg/mouse for a 30g mouse) and the biotin tracer was allowed to circulate for 5 min before perfusion and tissue collection. Mice were anesthetized with 87.5 mg per kg ketamine and 12.5 mg per kg xylazine and transcardially perfused with 25ml ice-cold PBS followed by 25ml 4% PFA. To process the brains, the whole brain was sectioned in half along the sagittal plane and postfixed in 4%PFA at 4 °C for 48 h before cryoprotection with 30% sucrose. Brains were then sectioned coronally at 40 μm with a freeze-stage microtome (Leica Camera) and stored in cryoprotective media. Brain sections were stained for vascular markers CD31 and AQP4 using standard immunohistochemistry approaches. A 1:1000 dilution of Streptavidin-AlexaFluor 647 was added during the secondary antibody incubation for 1 hour at room temperature. Sections were mounted on Superfrost Plus slides and coverslipped using ProlongGold. Hippocampal images were acquired using confocal microscopy (LSM900) and analyzed using Zeiss Zen software and ImageJ. Vascular area masks were generated by measuring the area covered by the vascular labelling. The NHS-biotin/Streptavidin labelling were measured, and the permeability index was determined as the area covered by NHS-biotin/Streptavidin signal outside of the vascular area mask.

Labelled Transferring (TF647) uptake assay.

50mg holo-Transferrin (Sigma Cat# T4132) was reconstituted in 4ml sterile PBS. Atto 647N NHS ester (Sigma Cat# 18373–1MG-F) was reconstituted in DMSO to generate a 30mM stock solution. Transferrin and Atto 647N NHS ester were mixed at a molar ratio of 1:1.8 and incubated at room temperature under constant agitation for 90 minutes. 50mM Tris (pH8) was added to quench the reaction and incubated at room temperature for 10 minutes. Reaction cleanup and unbound Atto 647N NHS ester was removed using two Zeba spin desalting columns (7k MWCO; Thermo Fisher Cat# 89892). 2 columns were used for each clean-up round. Protein concentration of the elution was determined using NanoDrop, adjusted to 10mg/ml in PBS and stored at −80C. Successful labeling was validated using Western blot analysis. 1μg of labelled and cold (unlabeled) Transferrin was separated on a 4–12% NuPAGE^™ Bis-Tris Mini Protein Gel, transferred to a nitrocellulose membrane and fluorescent signal visualized using a ChemiDoc imager (BioRad). For in vivo Transferrin uptake assay, 1mg of labelled TF-647 was injected retro-orbitally 20h before tissue collection. Mice were anesthetized with 87.5 mg per kg ketamine and 12.5 mg per kg xylazine and transcardially perfused with 25ml ice-cold PBS followed by 25ml 4% PFA. Brains were postfixed in 4%PFA at 4 °C for 48 h before cryoprotection with 30% sucrose. Brains were then sectioned coronally at 40 μm with a freeze-stage microtome (Leica Camera) and stored in cryoprotective media. Brain sections were stained for vascular markers CD31 and AQP4 using standard immunohistochemistry approaches. Hippocampal images were acquired using confocal microscopy (LSM900 and LSM800) and analyzed using Zeiss Zen software and ImageJ. Vascular area masks were generated, and TF-647 intensity was selectively measured in the area occupied by vascular labeling in the hippocampus.

AP activity labelling.

The Alkaline phosphatase activity labelling was performed on PFA-fixed free-floating sections (40um). Sections were washed 1x in TBST, followed by one wash in 0.1% Tris-HCL pH 8.5. The VectorRed Alkaline phosphatase substrate staining solution (Vector Laboratories Cat# SK-5100) was prepared following the manufacturer’s instruction in 0.1M Tris-HCL pH 8.5. Brain sections were incubated in the AP staining solution for 30 minutes at room temperature, followed by 3 washes in Tris-HCL pH 8.5. Sections were mounted on Superfrost Plus microscopy slides, air-dried overnight, cleared in Xylene and coverslipped with Permount. Sections treated with the selective TNAP inhibitor SBI-425 were used as a control to assess the specificity of the AP activity labelling. Images of the hippocampal dentate gyrus region were acquired on a Zeiss Epifluorescent microscope or Keyence bright-field microscopy for 3–5 hippocampal sections per mouse and quantified using FIJI. For co-labeling experiments, brain sections were first labelled for AP-activity staining, followed by standard immunohistochemistry for CD31/Aqp4 as outlined in the immunohistochemistry section.

Thioflavin S labelling.

Sections were rinsed three times in PBS, mounted on Superfrost Plus slides and air dried overnight⁶⁹. The sections were then incubated in freshly prepared and filtered 0.1% thioflavin S (ThioS) solution in 20% ethanol for 30 min at room temperature. The samples were rinsed twice with 20% ethanol for 2 min, followed by 2 washes in water. Following the Thioflavin S staining, sections were treated with 1:100 dilution of TrueBlack (Biotium cat# 23007) in 70% ethanol for 2 minutes and coverslipped with Prolong Gold. Images of the hippocampal dentate gyrus region were acquired on a Zeiss LSM800 or LSM900 confocal microscope and quantified using FIJI. On representative images, ThioS was detected in both the blue and green channels and appears as a merge on the representative images. The thresholded area occupied by the ThioS labelling in the dentate gyrus region of the hippocampus was quantified and averaged for 3–5 hippocampal sections per mouse.

TNAP cleavage assay.

A lentiviral Alpl/TNAP expression plasmid was first cloned under the control of the CMV promoter. Lentiviral particles were generated by co-transfecting HEK293T cells with the lentiviral Alpl/TNAP expression plasmid and lentiviral packaging plasmids (Addgene Plasmids# 12259 and 12260 were gifts from Didier Trono) using Lipofeamine 3000 (ThermoFisher, Cat# L300015)⁷⁰. Viral solution was collected 48 hours after transfection using 10 minute centrifugation at 1000g. The supernatant was collected and lentiviral particles purified and concentrated using 90 minute ultracentrifugation at 24’000 RPM (Beckman Coulter).TNAP reporter cells were generated by infecting HEK293T cells with the lentiviral particles, followed by subcloning and testing individual colonies. TNAP expression in reporter cells was validated using Western blot analysis. Reporter cells were plated at a confluency of 75–90% for transfection experiments and cultured in DMEM + 10% FBS. Lipofectamine 3000 was used as the transfection reagent. The TNAP reporter cells were transfected with GFP, mouse and human GPLD1, or enzymatically inactive mouse H133N GPLD1 expression constructs. The supernatant was collected at 48 hours for downstream analysis. A SEAP reporter assay kit (Abcam, Cat# 133077) was used to measure alkaline phosphatase activity in the media. Cells treated with the TNAP inhibitor SBI-425 (1uM) were used as a negative baseline control.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data, statistical analyses, and reproducibility.

All experiments were randomized and blinded by an independent researcher. Researchers remained blinded throughout histological, biochemical and behavioral assessments. Groups were unblinded at the end of each experiment on statistical analysis. Data are expressed as mean ± s.e.m. The distribution of data in each set of experiments was tested for normality using the D’Agostino–Pearson omnibus test or Shapiro-Wilk test. Statistical analysis was performed using Prism 8–10 (GraphPad). Means between two groups were compared using two-tailed unpaired Student’s t-tests. Comparisons of means from multiple groups with each other were analyzed using one-way ANOVA followed by the appropriate post hoc test, as indicated in the figure legends. Additional statistical details are indicated in the respective figure legends. All data generated or analyzed in this study are included in this article.

Supplementary Material

NIHMS2148338-supplement-1.xlsx^{(28.6KB, xlsx)}

NIHMS2148338-supplement-2.pdf^{(7.3MB, pdf)}

Key resources table.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti ALPL	R&D Systems	Cat#AF2910; RRID:AB_664062
Mouse monoclonal anti ALPL	R&D Systems	Cat#MAB29092; RRID:AB_2924405
Rat monoclonal anti CD31 (PECAM-1)	BioLegend	Cat#160202; RRID:AB_2876566
Goat polyclonal anti CD31 (PECAM-1)	R&D System	Cat#AF3628; RRID:AB_2161028
Rabbit polyclonal anti Aqp4	Millipore	Cat#AB2218; RRID:AB_11210366
Rabbit monoclonal anti Aqp4 [D1F8E]	Cell Signal Technology	Cat#59678; RRID:AB_2799571
Mouse monoclonal anti-β-Amyloid, 1-16 Antibody [6E10]	BioLegend	Cat#803001; RRID:AB_2564653
Mouse monoclonal anti Gapdh [6C5]	Abcam	Cat#ab8245; RRID:AB_2107448
Mouse monoclonal anti beta Actin-HRP [AC-15]	Abcam	Cat#ab49900; RRID:AB_867494
Rabbit monoclonal anti Caveolin1 (Cav1)	Cell Signal Technology	Cat#3267; RRID:AB_2275453
Rabbit monoclonal anti C1q	Abcam	Cat#ab182451; RRID:AB_2732849
Rabbit polyclonal anti GFAP	Dako	Cat#Z0334; RRID:AB_10013382
Rabbit monoclonal anti DCX	Cell Signal Technology	Cat#40619; RRID:AB_3696702
Mouse monoclonal anti MCM2/BM28	BD Biosciences	Cat#610700; RRID:AB_2141952
Guinea Pig polyclonal anti full length (FL) APP	Synaptic Systems	Cat#127 005; RRID:AB_2832229
Mouse monoclonal anti BACE1 (3D5)	Sigma-Aldrich/Millipore	Cat#MABN2640
Rabbit monoclonal anti PS1	Cell Signal Technology	Cat#5643S; RRID:AB_10706356
Mouse monoclonal anti BIII-Tubulin/TUBB3	BioLegend	Cat# 801201; RRID:AB_2313773
Rat anti CD16/32	BD Biosciences	Cat#553142; RRID:AB_394657
Rat anti CD45-PE	BD Biosciences	Cat#553081; RRID:AB_394611
Rat anti CD41-PE	BD Biosciences	Cat#558040; RRID:AB_397004
Rat anti CD31-APC	BD Biosciences	Cat#551262; RRID:AB_398497
Bacterial and virus strains
NEB^® 5-alpha Competent E. coli	New England Biolabs (NEB)	Cat#C2987U
Biological samples
Mouse brain tissue	This paper	N/A
Mouse liver tissue	This paper	N/A
Human cortex tissue block	This paper	N/A
Chemicals, peptides, and recombinant proteins
SBI-425	Sigma-Aldrich; Cayman Chemicals	Cat#SML2935; Cat#34626
PEI	Polysciences	Cat#23966-1
Hoechst 33342	Thermo Fisher	Cat#H3570
Thioflavine S	Sigma/Millipore	Cat#T1892-25G
EZ-Link^™ Sulfo-NHS-LC-Biotin	Thermo Fisher	Cat#21335
Atto 647N NHS ester	Sigma-Aldrich	Cat#18373-1MG-F
Holo-Transferrin	Sigma-Aldrich	Cat#T4132
Streptavidin-AlexaFluor647	Thermo Fisher	Cat#S32357
Percoll solution	Cytiva	Cat#17089101
TrueBlack	Biotium	Cat#23007
Critical commercial assays
Vector^® Red Substrate Kit, Alkaline Phosphatase	Vector Laboratories	Cat#SK-5100
SEAP Reporter Gene Assay Kit	Abcam	Cat#ab133077
NEBuilder HiFi DNA Assembly Kit	NEB	Cat#E5520S
PureLink RNA Mini kit	Thermo Fisher	Cat#12183025
Direct-zol^™ RNA Purification Kit, Miniprep Plus	Zymo Research	Cat#R2072
High Capacity cDNA Reverse Transcription kit	Thermo Fisher	Cat#4374966
PowerUp SYBR Green Master Mix	Thermo Fisher	Cat#A25742
Zeba^™ Spin Desalting Columns, 7K MWCO, 5 mL	Thermo Fisher	Cat#89892
Miltenyi Neural Dissociation Kits (P)	Miltenyi	Cat#130-092-628
Deposited data
Single-nucleus RNA sequencing data	This paper	GSE304483
BEC single-cell RNA sequencing data	This paper	GSE269061
Experimental models: Cell lines
293T	ATCC	Cat#CRL-3216; RRID:CVCL_0063
Neuro-2a	ATCC	Cat#CCL-131; RRID:CVCL_0470
Experimental models: Organisms/strains
C57BL/6J	The Jackson Laboratory	Strain#000664
C57BlL6/J NIA Aging mouse colony	NIA	N/A
5xFAD (B6SJL-Tg(APPSwFlLon,PSEN1M146LL286V)6799Vas/Mmjax)	The Jackson Laboratory/MMRRC	MMRRC Strain#034840-JAX
B6;129-Gt(ROSA)26Sor^tm1(CAG-cas9*^,-EGFP)Fezh/J	The Jackson Laboratory	Strain#024857
Oligonucleotides
Gpld1 qPCR FWD: GGAAGCAGAGAGGAATTGTGGC	Horowitz et al.	N/A
Gpld1 qPCR REV: TCCAAACCACGAGAAGTCCTCC	Horowitz et al.	N/A
Gapdh qPCR FWD: GGGTGTGAACCACGAGAAAT	This paper	N/A
Gapdh qPCR REV: ACTGTGGTCATGAGCCCTTC	This paper	N/A
GPLD1 qPCR FWD (human)	This paper	N/A
GPLD1 qPCR REV (human)	This paper	N/A
Alpl qPCR FWD: GAACAGACCCTCCCCACGAG	This paper	N/A
Alpl qPCR REV: GTGCCGATGGCCAGTACTAA	This paper	N/A
Alpl qPCR FWD2: TAACACCAACGCTCAGGTCC	This paper	N/A
Alpl qPCR REV2: TGGATGTGACCTCATTGCCC	This paper	N/A
Recombinant DNA
pAAV2/8	Addgene	Plasmid#112864
pUCmini-iCAP-PHP.V1	Addgene	Plasmid#127847
pAAV-CAG-mRuby2	Addgene	Plasmid#99123
pEMS1938	Addgene	Plasmid#82563
pAdDeltaF6	Addgene	Plasmid#112867
pMD2.G	Addgene	Plasmid#12259
psPAX2	Addgene	Plasmid#12260
lentiCRISPR v2	Addgene	Plasmid#52961
pTB CMV Gpld1 IRES eGFP	Horowitz et al.	N/A
pTB CMV Gpld1 H133N IRES eGFP	Horowitz et al.	N/A
pTB CMV Gpld1 H158N IRES eGFP	Horowitz et al.	N/A
pTB CMV IRES eGFP	Horowitz et al.	N/A
pTB CMV Alpl IRES eGFP	This paper	N/A
pTB CMV GPLD1 (human) IRES eGFP	This paper	N/A
pAAV-CAG-Alpl	This paper	N/A
pAAV-CRE-hU6-gRNA	This paper	N/A
pAAV TBG eGFP	This paper	N/A
pAAV TBG Gpld1	This paper	N/A
pAAV TBG Gpld1 H133N	This paper	N/A
Software and algorithms
FIJI/ImageJ2	ImageJ	https://imagej.net/software/fiji/
Prism 10	GraphPad	https://www.graphpad.com
Zeiss Zen 3.7/3.3	Zeiss
Cell Ranger version 7.1.0.	10X Genomics	https://support.10xgenomics.com/single-cell-gene-expression/software/downloads/latest
RStudio	Posit	https://posit.co/downloads/
Seurat v5	Satija Lab	https://satijalab.org/seurat/
Wheel Manager Software	MedAssociates	Cat# SOF-860
Smart Video Tracking Software	Panlab/Harvard Apparatus	http://www.panlab.com/en/products/smart-video-tracking-software-panlab
ChemiDoc Image Lab Software 6.1	BioRad	N/A
EthoVision XT	Noldus	N/A
Other

Open in a new tab

Highlights.

Liver exercise factor GPLD1 targets GPI-anchored proteins on the aged brain vasculature
GPI-anchored TNAP on brain endothelial cells disrupts BBB and impairs cognition
Increased GPLD1 or TNAP inhibition rejuvenate BBB function and cognition in aging
Increased GPLD1 or TNAP inhibition ameliorate Alzheimer’s disease pathology

ACKNOWLEDGMENTS

We thank Dr. Param Singh and Dr. Andrew Brack for critically reading manuscript and Dr. Elizabeth Crouch and Dr. Sophia M. Shi for critical input on brain vascular analyses. This work was funded by Simons Foundation (S.A.V.), Bakar Family Foundation (S.A.V.), Cure Alzheimer’s Fund (S.A.V.), Hillblom Foundation (G.B.), Glenn Foundation (T.A.), JSPS (Y.F.), Japanese Biochemistry Postdoctoral Fellowship (Y.F.), Multiple Sclerosis Foundation (A.R.P.), Frontiers in Medical Research fellowship (K.J.B.P.), American Federation for Aging Research (T.A.), National Science Foundation (J.S.), Bakar Aging Research Institute (S.A.V.), gift from Marc and Lynne Benioff, and National Institute on Aging (AG081038 (G.B.), AG086042 (J.S.), AG082414 (K.B.C., S.A.V.), AG077770 (S.A.V.), AG067740 (S.A.V.)). We acknowledge the UCSF Parnassus Flow Core (RRID:SCR_018206) and support by the DRC Center Grant NIH P30 DK063720 for assistance with Flow Cytometry. We thank the Genomics CoLabs at the UCSF Institute for Human Genetics for assistance with snRNAseq and the UCSF CAT core for sequencing.

Footnotes

DECLARATION OF INTERESTS STATEMENT

The Regents of the University of California have applied for a provisional patent application arising from this work, “Exercise-induced circulatory factors for amelioration of cognitive, neurological, and regenerative dysfunction during aging” (PCT/US2020/016549 Inventor S.A.V.). S.A.V. consulted for The Herrick Company, Inc. and is a cofounder of Ceiba Bio, Inc. All other authors declare no competing interests.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

1.Fan X, Wheatley EG, and Villeda SA (2016). Mechanisms of Hippocampal Aging and the Potential for Rejuvenation. Annu Rev Neurosci 40, 1–22. 10.1146/annurev-neuro-072116-031357. [DOI] [PubMed] [Google Scholar]
2.Bieri G, Schroer AB, and Villeda SA (2023). Blood-to-brain communication in aging and rejuvenation. Nat. Neurosci. 26, 379–393. 10.1038/s41593-022-01238-8. [DOI] [PubMed] [Google Scholar]
3.Chow LS, Gerszten RE, Taylor JM, Pedersen BK, Praag H. van, Trappe S, Febbraio MA, Galis ZS, Gao Y, Haus JM, et al. (2022). Exerkines in health, resilience and disease. Nat. Rev. Endocrinol. 18, 273–289. 10.1038/s41574-022-00641-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Praag H. van, Shubert T, Zhao C, and Gage FH. (2005). Exercise Enhances Learning and Hippocampal Neurogenesis in Aged Mice. J Neurosci 25, 8680–8685. 10.1523/jneurosci.1731-05.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Kumar A, Rani A, Tchigranova O, Lee W-H, and Foster TC (2012). Influence of late-life exposure to environmental enrichment or exercise on hippocampal function and CA1 senescent physiology. Neurobiol Aging 33, 828.e1-828.e17. 10.1016/j.neurobiolaging.2011.06.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Soto I, Graham LC, Richter HJ, Simeone SN, Radell JE, Grabowska W, Funkhouser WK, Howell MC, and Howell GR (2015). APOE Stabilization by Exercise Prevents Aging Neurovascular Dysfunction and Complement Induction. Plos Biol 13, e1002279. 10.1371/journal.pbio.1002279. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Speisman RB, Kumar A, Rani A, Foster TC, and Ormerod BK (2013). Daily exercise improves memory, stimulates hippocampal neurogenesis and modulates immune and neuroimmune cytokines in aging rats. Brain Behav Immun 28, 25–43. 10.1016/j.bbi.2012.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Choi SH, Bylykbashi E, Chatila ZK, Lee SW, Pulli B, Clemenson GD, Kim E, Rompala A, Oram MK, Asselin C, et al. (2018). Combined adult neurogenesis and BDNF mimic exercise effects on cognition in an Alzheimer’s mouse model. Science 361, eaan8821. 10.1126/science.aan8821. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Barnes DE, and Yaffe K (2011). The projected effect of risk factor reduction on Alzheimer’s disease prevalence. Lancet Neurology 10, 819–828. 10.1016/s1474-4422(11)70072–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Hörder H, Johansson L, Guo X, Grimby G, Kern S, Östling S, and Skoog I (2018). Midlife cardiovascular fitness and dementia: A 44-year longitudinal population study in women. Neurology 90, e1298–e1305. 10.1212/wnl.0000000000005290. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Yaffe K, Barnes D, Nevitt M, Lui LY, and Covinsky K (2001). A prospective study of physical activity and cognitive decline in elderly women: women who walk. Arch Intern Med 161, 1703–1708. 10.1001/archinte.161.14.1703. [DOI] [PubMed] [Google Scholar]
12.Müller S, Preische O, Sohrabi HR, Gräber S, Jucker M, Ringman JM, Martins RN, McDade E, Schofield PR, Ghetti B, et al. (2018). Relationship between physical activity, cognition, and Alzheimer pathology in autosomal dominant Alzheimer’s disease. Alzheimer’s Dementia 14, 1427–1437. 10.1016/j.jalz.2018.06.3059. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Cunningham C, Sullivan RO, Caserotti P, and Tully MA (2019). Consequences of physical inactivity in older adults: A systematic review of reviews and meta-analyses. Scand J Med Sci Spor 30, 816–827. 10.1111/sms.13616. [DOI] [PubMed] [Google Scholar]
14.Kohl HW, Craig CL, Lambert EV, Inoue S, Alkandari JR, Leetongin G, Kahlmeier S, and Group LPASW (2012). The pandemic of physical inactivity: global action for public health. Lancet Lond Engl 380, 294–305. 10.1016/s0140-6736(12)60898-8. [DOI] [PubMed] [Google Scholar]
15.Gomes M, Figueiredo D, Teixeira L, Poveda V, Paúl C, Santos-Silva A, and Costa E (2016). Physical inactivity among older adults across Europe based on the SHARE database. Age Ageing 46, 71–77. 10.1093/ageing/afw165. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Horowitz AM, Fan X, Bieri G, Smith LK, Sanchez-Diaz CI, Schroer AB, Gontier G, Casaletto KB, Kramer JH, Williams KE, et al. (2020). Blood factors transfer beneficial effects of exercise on neurogenesis and cognition to the aged brain. Science 369, 167–173. 10.1126/science.aaw2622. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Miguel ZD, Khoury N, Betley MJ, Lehallier B, Willoughby D, Olsson N, Yang AC, Hahn O, Lu N, Vest RT, et al. (2019). Exercise plasma boosts memory and dampens brain inflammation via clusterin. Nature 600, 494–499. 10.1038/s41586-021-04183-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Norevik CS, Huuha AM, Røsbjørgen RN, Bergersen L, Jacobsen K, Miguel-dos-Santos R, Ryan L, Skender B, Moreira JBN, Kobro-Flatmoen A, et al. (2023). Exercised blood plasma promotes hippocampal neurogenesis in the Alzheimer’s disease rat brain. J. Sport Heal. Sci. 10.1016/j.jshs.2023.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Schaum N, Karkanias J, Neff NF, May AP, Quake SR, Wyss-Coray T, Darmanis S, Batson J, Botvinnik O, Chen MB, et al. (2018). Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris. Nature 562, 367–372. 10.1038/s41586-018-0590-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Karlsson M, Zhang C, Méar L, Zhong W, Digre A, Katona B, Sjöstedt E, Butler L, Odeberg J, Dusart P, et al. (2021). A single–cell type transcriptomics map of human tissues. Sci. Adv. 7, eabh2169. 10.1126/sciadv.abh2169. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Yousef H, Czupalla CJ, Lee D, Chen MB, Burke AN, Zera KA, Zandstra J, Berber E, Lehallier B, Mathur V, et al. (2019). Aged blood impairs hippocampal neural precursor activity and activates microglia via brain endothelial cell VCAM1. Nat Med 25, 988–1000. 10.1038/s41591-019-0440-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Chen MB, Yang AC, Yousef H, Lee D, Chen W, Schaum N, Lehallier B, Quake SR, and Wyss-Coray T (2019). Brain Endothelial Cells Are Exquisite Sensors of Age-Related Circulatory Cues. Cell Reports 30, 4418–4432.e4. 10.1016/j.celrep.2020.03.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Millán JL, and Whyte MP (2016). Alkaline Phosphatase and Hypophosphatasia. Calcif Tissue Int 98, 398–416. 10.1007/s00223-015-0079-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Claudia G, Agnieszka S-K, Laurence B, Thibaut Q, Laura M, Slawomir P, Emmanuelle C-S, Luis MJ, Caroline F, and David M (2020). TNAP as a therapeutic target for cardiovascular calcification – a discussion of its pleiotropic functions in the body. Cardiovasc Res 118, cvaa299-. 10.1093/cvr/cvaa299. [DOI] [Google Scholar]
25.Millán JL (2006). Alkaline Phosphatases. Purinerg Signal 2, 335. 10.1007/s11302-005-5435-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Yang AC, Stevens MY, Chen MB, Lee DP, Stähli D, Gate D, Contrepois K, Chen W, Iram T, Zhang L, et al. (2020). Physiological blood–brain transport is impaired with age by a shift in transcytosis. Nature 583, 425–430. 10.1038/s41586-020-2453-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Raikwar NS, Bowen RF, and Deeg MA (2005). Mutating His29, His125, His133 or His158 abolishes glycosylphosphatidylinositol-specific phospholipase D catalytic activity. Biochem J 391, 285–289. 10.1042/bj20050656. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Scallon BJ, Fung W-JC, Tsang TC, Li S, Kado-Fong H, Huang K-S, and Kochan JP (1991). Primary Structure and Functional Activity of a Phosphatidylinositol-Glycan-Specific Phospholipase D. Science 252, 446–448. 10.1126/science.2017684. [DOI] [PubMed] [Google Scholar]
29.Park MH, Lee JY, Park KH, Jung IK, Kim K-T, Lee Y-S, Ryu H-H, Jeong Y, Kang M, Schwaninger M, et al. (2018). Vascular and Neurogenic Rejuvenation in Aging Mice by Modulation of ASM. Neuron 100, 167–182.e9. 10.1016/j.neuron.2018.09.010. [DOI] [PubMed] [Google Scholar]
30.Shi SM, Suh RJ, Shon DJ, Garcia FJ, Buff JK, Atkins M, Li L, Lu N, Sun B, Luo J, et al. (2025). Glycocalyx dysregulation impairs blood–brain barrier in ageing and disease. Nature 639, 985–994. 10.1038/s41586-025-08589-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Kumar SR, Miles TF, Chen X, Brown D, Dobreva T, Huang Q, Ding X, Luo Y, Einarsson PH, Greenbaum A, et al. (2020). Multiplexed Cre-dependent selection yields systemic AAVs for targeting distinct brain cell types. Nat. Methods 17, 541–550. 10.1038/s41592-020-0799-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Brichacek AL, Benkovic SA, Chakraborty S, Nwafor DC, Wang W, Jun S, Dakhlallah D, Geldenhuys WJ, Pinkerton AB, Millán JL, et al. (2019). Systemic inhibition of tissue-nonspecific alkaline phosphatase alters the brain-immune axis in experimental sepsis. Sci Rep-uk 9, 18788. 10.1038/s41598-019-55154-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Smith LK, Verovskaya E, Bieri G, Horowitz AM, Ungern‐Sternberg S.N.I. von, Lin K, Seizer P, Passegué E, and Villeda SA. (2020). The aged hematopoietic system promotes hippocampal‐dependent cognitive decline. Aging Cell 19, e13192. 10.1111/acel.13192. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Nakada-Honda N, Cui D, Matsuda S, and Ikeda E (2021). Intravenous injection of cyclophilin A realizes the transient and reversible opening of barrier of neural vasculature through basigin in endothelial cells. Sci. Rep. 11, 19391. 10.1038/s41598-021-98163-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Katsimpardi L, Litterman NK, Schein PA, Miller CM, Loffredo FS, Wojtkiewicz GR, Chen JW, Lee RT, Wagers AJ, and Rubin LL (2014). Vascular and Neurogenic Rejuvenation of the Aging Mouse Brain by Young Systemic Factors. Science 344, 630–634. 10.1126/science.1251141. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Stephan AH, Madison DV, Mateos JM, Fraser DA, Lovelett EA, Coutellier L, Kim L, Tsai H-H, Huang EJ, Rowitch DH, et al. (2013). A Dramatic Increase of C1q Protein in the CNS during Normal Aging. J. Neurosci. 33, 13460–13474. 10.1523/jneurosci.1333-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Hou Y, Dan X, Babbar M, Wei Y, Hasselbalch SG, Croteau DL, and Bohr VA (2019). Ageing as a risk factor for neurodegenerative disease. Nat. Rev. Neurol. 15, 565–581. 10.1038/s41582-019-0244-7. [DOI] [PubMed] [Google Scholar]
38.Oakley H, Cole SL, Logan S, Maus E, Shao P, Craft J, Guillozet-Bongaarts A, Ohno M, Disterhoft J, Eldik LV, et al. (2006). Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: potential factors in amyloid plaque formation. J Neurosci Official J Soc Neurosci 26, 10129–10140. 10.1523/jneurosci.1202-06.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Kimura R, and Ohno M (2008). Impairments in remote memory stabilization precede hippocampal synaptic and cognitive failures in 5XFAD Alzheimer mouse model. Neurobiol Dis 33, 229–235. 10.1016/j.nbd.2008.10.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Maguire GA, and Gossner A (1994). Glycosyl Phosphatidyl Inositol Phospholipase D Activity in Human Serum. Ann Clin Biochem 32, 74–78. 10.1177/000456329503200107. [DOI] [PubMed] [Google Scholar]
41.Keren-Shaul H, Spinrad A, Weiner A, Matcovitch-Natan O, Dvir-Szternfeld R, Ulland TK, David E, Baruch K, Lara-Astaiso D, Toth B, et al. (2017). A Unique Microglia Type Associated with Restricting Development of Alzheimer’s Disease. Cell 169, 1276–1290.e17. 10.1016/j.cell.2017.05.018. [DOI] [PubMed] [Google Scholar]
42.Ungvari Z, Tarantini S, Kiss T, Wren JD, Giles CB, Griffin CT, Murfee WL, Pacher P, and Csiszar A (2018). Endothelial dysfunction and angiogenesis impairment in the ageing vasculature. Nat Rev Cardiol 15, 555–565. 10.1038/s41569-018-0030-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Zlokovic BV (2011). Neurovascular pathways to neurodegeneration in Alzheimer’s disease and other disorders. Nat Rev Neurosci 12, 723–738. 10.1038/nrn3114. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Nation DA, Sweeney MD, Montagne A, Sagare AP, D’Orazio LM, Pachicano M, Sepehrband F, Nelson AR, Buennagel DP, Harrington MG, et al. (2018). Blood-brain barrier breakdown is an early biomarker of human cognitive dysfunction. Nat Med 25, 270–276. 10.1038/s41591-018-0297-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Sweeney MD, Sagare AP, and Zlokovic BV (2018). Blood–brain barrier breakdown in Alzheimer disease and other neurodegenerative disorders. Nat Rev Neurol 14, 133–150. 10.1038/nrneurol.2017.188. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Arvanitakis Z, Capuano AW, Leurgans SE, Bennett DA, and Schneider JA (2016). Relation of cerebral vessel disease to Alzheimer’s disease dementia and cognitive function in elderly people: a cross-sectional study. Lancet Neurology 15, 934–943. 10.1016/s1474-4422(16)30029-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Lee S, Viqar F, Zimmerman ME, Narkhede A, Tosto G, Benzinger TLS, Marcus DS, Fagan AM, Goate A, Fox NC, et al. (2015). White matter hyperintensities are a core feature of Alzheimer’s disease: Evidence from the dominantly inherited Alzheimer network. Ann Neurol 79, 929–939. 10.1002/ana.24647. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Ozek C, Krolewski RC, Buchanan SM, and Rubin LL (2018). Growth Differentiation Factor 11 treatment leads to neuronal and vascular improvements in the hippocampus of aged mice. Sci Rep-uk 8, 17293. 10.1038/s41598-018-35716-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Park MH, Lee JY, Park KH, Jung IK, Kim K-T, Lee Y-S, Ryu H-H, Jeong Y, Kang M, Schwaninger M, et al. (2018). Vascular and Neurogenic Rejuvenation in Aging Mice by Modulation of ASM. Neuron 100, 762. 10.1016/j.neuron.2018.10.038. [DOI] [PubMed] [Google Scholar]
50.Leiter O, Brici D, Fletcher SJ, Yong XLH, Widagdo J, Matigian N, Schroer AB, Bieri G, Blackmore DG, Bartlett PF, et al. (2022). Platelet-derived exerkine CXCL4/platelet factor 4 rejuvenates hippocampal neurogenesis and restores cognitive function in aged mice. Nat. Commun. 14, 4375. 10.1038/s41467-023-39873-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Leiter O, Zhuo Z, Rust R, Wasielewska JM, Grönnert L, Kowal S, Overall RW, Adusumilli VS, Blackmore DG, Southon A, et al. (2023). Selenium mediates exercise-induced adult neurogenesis and reverses learning deficits induced by hippocampal injury and aging. Cell Metab. 35, 1085. 10.1016/j.cmet.2023.04.019. [DOI] [PubMed] [Google Scholar]
52.Lourenco MV, Frozza RL, Freitas G.B. de, Zhang H, Kincheski GC, Ribeiro FC, Gonçalves RA, Clarke JR, Beckman D, Staniszewski A, et al. (2019). Exercise-linked FNDC5/irisin rescues synaptic plasticity and memory defects in Alzheimer’s models. Nat. Med. 25, 165–175. 10.1038/s41591-018-0275-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Islam MR, Valaris S, Young MF, Haley EB, Luo R, Bond SF, Mazuera S, Kitchen RR, Caldarone BJ, Bettio LEB, et al. (2021). Exercise hormone irisin is a critical regulator of cognitive function. Nat. Metab. 3, 1058–1070. 10.1038/s42255-021-00438-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Trejo JL, Carro E, and Torres-Aleman I (2001). Circulating insulin-like growth factor I mediates exercise-induced increases in the number of new neurons in the adult hippocampus. J. Neurosci. : Off. J. Soc. Neurosci. 21, 1628–1634. [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Schroer AB, Ventura PB, Sucharov J, Misra R, Chui MKK, Bieri G, Horowitz AM, Smith LK, Encabo K, Tenggara I, et al. (2023). Platelet factors attenuate inflammation and rescue cognition in ageing. Nature 620, 1071–1079. 10.1038/s41586-023-06436-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Park C, Hahn O, Gupta S, Moreno AJ, Marino F, Kedir B, Wang D, Villeda SA, Wyss-Coray T, and Dubal DB (2023). Platelet factors are induced by longevity factor klotho and enhance cognition in young and aging mice. Nat. Aging 3, 1067–1078. 10.1038/s43587-023-00468-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Oh HS-H, Rutledge J, Nachun D, Pálovics R, Abiose O, Moran-Losada P, Channappa D, Urey DY, Kim K, Sung YJ, et al. (2023). Organ aging signatures in the plasma proteome track health and disease. Nature 624, 164–172. 10.1038/s41586-023-06802-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Li X, Shi X, McPherson M, Hager M, Garcia GG, and Miller RA (2022). Cap-independent translation of GPLD1 enhances markers of brain health in long‐lived mutant and drug‐treated mice. Aging Cell 21, e13685. 10.1111/acel.13685. [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Li X, Hager M, McPherson M, Lee M, Hagalwadi R, Skinner ME, Lombard D, and Miller RA (2023). Recapitulation of anti-aging phenotypes by global, but not by muscle-specific, deletion of PAPP-A in mice. GeroScience 45, 931–948. 10.1007/s11357-022-00692-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Hager M, Chang P, Lee M, Burns CM, Endicott SJ, Miller RA, and Li X (2024). Recapitulation of anti-aging phenotypes by global overexpression of PTEN in mice. GeroScience 46, 2653–2670. 10.1007/s11357-023-01025-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Thrupp N, Frigerio CS, Wolfs L, Skene NG, Fattorelli N, Poovathingal S, Fourne Y, Matthews PM, Theys T, Mancuso R, et al. (2020). Single-Nucleus RNA-Seq Is Not Suitable for Detection of Microglial Activation Genes in Humans. Cell Rep. 32, 108189. 10.1016/j.celrep.2020.108189. [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Chan KY, Jang MJ, Yoo BB, Greenbaum A, Ravi N, Wu W-L, Sánchez-Guardado L, Lois C, Mazmanian SK, Deverman BE, et al. (2017). Engineered AAVs for efficient noninvasive gene delivery to the central and peripheral nervous systems. Nat. Neurosci. 20, 1172–1179. 10.1038/nn.4593. [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Aurnhammer C, Haase M, Muether N, Hausl M, Rauschhuber C, Huber I, Nitschko H, Busch U, Sing A, Ehrhardt A, et al. (2012). Universal Real-Time PCR for the Detection and Quantification of Adeno-Associated Virus Serotype 2-Derived Inverted Terminal Repeat Sequences. Hum. Gene Ther., Part B: Methods 23, 18–28. 10.1089/hgtb.2011.034. [DOI] [PubMed] [Google Scholar]
64.Sanjana NE, Shalem O, and Zhang F (2014). Improved vectors and genome-wide libraries for CRISPR screening. Nat. Methods 11, 783–784. 10.1038/nmeth.3047. [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Alamed J, Wilcock DM, Diamond DM, Gordon MN, and Morgan D (2006). Two-day radial-arm water maze learning and memory task; robust resolution of amyloid-related memory deficits in transgenic mice. Nat Protoc 1, 1671–1679. 10.1038/nprot.2006.275. [DOI] [PubMed] [Google Scholar]
66.Stevenson ME, Bieri G, Kaletsky R, Ange J.St., Remesal L, Pratt KJB, Zhou S, Weng Y, Murphy CT, and Villeda SA. (2023). Neuronal activation of Gαq EGL-30/GNAQ late in life rejuvenates cognition across species. Cell Rep. 42, 113151. 10.1016/j.celrep.2023.113151. [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Crouch EE, and Doetsch F (2018). FACS isolation of endothelial cells and pericytes from mouse brain microregions. Nat. Protoc. 13, 738–751. 10.1038/nprot.2017.158. [DOI] [PubMed] [Google Scholar]
68.Sasmita AO, Ong EC, Nazarenko T, Mao S, Komarek L, Thalmann M, Hantakova V, Spieth L, Berghoff SA, Barr HJ, et al. (2025). Parental origin of transgene modulates amyloid-β plaque burden in the 5xFAD mouse model of Alzheimer’s disease. Neuron 113, 838–846.e4. 10.1016/j.neuron.2024.12.025. [DOI] [PubMed] [Google Scholar]
69.Bieri G, Lucin KM, O’Brien CE, Zhang H, Villeda SA, and Wyss-Coray T (2018). Proteolytic cleavage of Beclin 1 exacerbates neurodegeneration. Mol Neurodegener 13, 68. 10.1186/s13024-018-0302-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Lin K, Bieri G, Gontier G, Müller S, Smith LK, Snethlage CE, White CW, Maybury-Lewis SY, and Villeda SA (2021). MHC class I H2-Kb negatively regulates neural progenitor cell proliferation by inhibiting FGFR signaling. Plos Biol 19, e3001311. 10.1371/journal.pbio.3001311. [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

NIHMS2148338-supplement-1.xlsx^{(28.6KB, xlsx)}

NIHMS2148338-supplement-2.pdf^{(7.3MB, pdf)}

Data Availability Statement

The single-nucleus and single-cell RNA-sequencing datasets are available at the Gene Expression Omnibus (GEO). Accession numbers: GSE304483 and GSE269061.
This paper does not report original code.
Additional details are available from the lead contact upon request.

[R1] 1.Fan X, Wheatley EG, and Villeda SA (2016). Mechanisms of Hippocampal Aging and the Potential for Rejuvenation. Annu Rev Neurosci 40, 1–22. 10.1146/annurev-neuro-072116-031357. [DOI] [PubMed] [Google Scholar]

[R2] 2.Bieri G, Schroer AB, and Villeda SA (2023). Blood-to-brain communication in aging and rejuvenation. Nat. Neurosci. 26, 379–393. 10.1038/s41593-022-01238-8. [DOI] [PubMed] [Google Scholar]

[R3] 3.Chow LS, Gerszten RE, Taylor JM, Pedersen BK, Praag H. van, Trappe S, Febbraio MA, Galis ZS, Gao Y, Haus JM, et al. (2022). Exerkines in health, resilience and disease. Nat. Rev. Endocrinol. 18, 273–289. 10.1038/s41574-022-00641-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R4] 4.Praag H. van, Shubert T, Zhao C, and Gage FH. (2005). Exercise Enhances Learning and Hippocampal Neurogenesis in Aged Mice. J Neurosci 25, 8680–8685. 10.1523/jneurosci.1731-05.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R5] 5.Kumar A, Rani A, Tchigranova O, Lee W-H, and Foster TC (2012). Influence of late-life exposure to environmental enrichment or exercise on hippocampal function and CA1 senescent physiology. Neurobiol Aging 33, 828.e1-828.e17. 10.1016/j.neurobiolaging.2011.06.023. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Soto I, Graham LC, Richter HJ, Simeone SN, Radell JE, Grabowska W, Funkhouser WK, Howell MC, and Howell GR (2015). APOE Stabilization by Exercise Prevents Aging Neurovascular Dysfunction and Complement Induction. Plos Biol 13, e1002279. 10.1371/journal.pbio.1002279. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] 7.Speisman RB, Kumar A, Rani A, Foster TC, and Ormerod BK (2013). Daily exercise improves memory, stimulates hippocampal neurogenesis and modulates immune and neuroimmune cytokines in aging rats. Brain Behav Immun 28, 25–43. 10.1016/j.bbi.2012.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Choi SH, Bylykbashi E, Chatila ZK, Lee SW, Pulli B, Clemenson GD, Kim E, Rompala A, Oram MK, Asselin C, et al. (2018). Combined adult neurogenesis and BDNF mimic exercise effects on cognition in an Alzheimer’s mouse model. Science 361, eaan8821. 10.1126/science.aan8821. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.Barnes DE, and Yaffe K (2011). The projected effect of risk factor reduction on Alzheimer’s disease prevalence. Lancet Neurology 10, 819–828. 10.1016/s1474-4422(11)70072–2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] 10.Hörder H, Johansson L, Guo X, Grimby G, Kern S, Östling S, and Skoog I (2018). Midlife cardiovascular fitness and dementia: A 44-year longitudinal population study in women. Neurology 90, e1298–e1305. 10.1212/wnl.0000000000005290. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] 11.Yaffe K, Barnes D, Nevitt M, Lui LY, and Covinsky K (2001). A prospective study of physical activity and cognitive decline in elderly women: women who walk. Arch Intern Med 161, 1703–1708. 10.1001/archinte.161.14.1703. [DOI] [PubMed] [Google Scholar]

[R12] 12.Müller S, Preische O, Sohrabi HR, Gräber S, Jucker M, Ringman JM, Martins RN, McDade E, Schofield PR, Ghetti B, et al. (2018). Relationship between physical activity, cognition, and Alzheimer pathology in autosomal dominant Alzheimer’s disease. Alzheimer’s Dementia 14, 1427–1437. 10.1016/j.jalz.2018.06.3059. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] 13.Cunningham C, Sullivan RO, Caserotti P, and Tully MA (2019). Consequences of physical inactivity in older adults: A systematic review of reviews and meta-analyses. Scand J Med Sci Spor 30, 816–827. 10.1111/sms.13616. [DOI] [PubMed] [Google Scholar]

[R14] 14.Kohl HW, Craig CL, Lambert EV, Inoue S, Alkandari JR, Leetongin G, Kahlmeier S, and Group LPASW (2012). The pandemic of physical inactivity: global action for public health. Lancet Lond Engl 380, 294–305. 10.1016/s0140-6736(12)60898-8. [DOI] [PubMed] [Google Scholar]

[R15] 15.Gomes M, Figueiredo D, Teixeira L, Poveda V, Paúl C, Santos-Silva A, and Costa E (2016). Physical inactivity among older adults across Europe based on the SHARE database. Age Ageing 46, 71–77. 10.1093/ageing/afw165. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.Horowitz AM, Fan X, Bieri G, Smith LK, Sanchez-Diaz CI, Schroer AB, Gontier G, Casaletto KB, Kramer JH, Williams KE, et al. (2020). Blood factors transfer beneficial effects of exercise on neurogenesis and cognition to the aged brain. Science 369, 167–173. 10.1126/science.aaw2622. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] 17.Miguel ZD, Khoury N, Betley MJ, Lehallier B, Willoughby D, Olsson N, Yang AC, Hahn O, Lu N, Vest RT, et al. (2019). Exercise plasma boosts memory and dampens brain inflammation via clusterin. Nature 600, 494–499. 10.1038/s41586-021-04183-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] 18.Norevik CS, Huuha AM, Røsbjørgen RN, Bergersen L, Jacobsen K, Miguel-dos-Santos R, Ryan L, Skender B, Moreira JBN, Kobro-Flatmoen A, et al. (2023). Exercised blood plasma promotes hippocampal neurogenesis in the Alzheimer’s disease rat brain. J. Sport Heal. Sci. 10.1016/j.jshs.2023.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Schaum N, Karkanias J, Neff NF, May AP, Quake SR, Wyss-Coray T, Darmanis S, Batson J, Botvinnik O, Chen MB, et al. (2018). Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris. Nature 562, 367–372. 10.1038/s41586-018-0590-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] 20.Karlsson M, Zhang C, Méar L, Zhong W, Digre A, Katona B, Sjöstedt E, Butler L, Odeberg J, Dusart P, et al. (2021). A single–cell type transcriptomics map of human tissues. Sci. Adv. 7, eabh2169. 10.1126/sciadv.abh2169. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] 21.Yousef H, Czupalla CJ, Lee D, Chen MB, Burke AN, Zera KA, Zandstra J, Berber E, Lehallier B, Mathur V, et al. (2019). Aged blood impairs hippocampal neural precursor activity and activates microglia via brain endothelial cell VCAM1. Nat Med 25, 988–1000. 10.1038/s41591-019-0440-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] 22.Chen MB, Yang AC, Yousef H, Lee D, Chen W, Schaum N, Lehallier B, Quake SR, and Wyss-Coray T (2019). Brain Endothelial Cells Are Exquisite Sensors of Age-Related Circulatory Cues. Cell Reports 30, 4418–4432.e4. 10.1016/j.celrep.2020.03.012. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R23] 23.Millán JL, and Whyte MP (2016). Alkaline Phosphatase and Hypophosphatasia. Calcif Tissue Int 98, 398–416. 10.1007/s00223-015-0079-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] 24.Claudia G, Agnieszka S-K, Laurence B, Thibaut Q, Laura M, Slawomir P, Emmanuelle C-S, Luis MJ, Caroline F, and David M (2020). TNAP as a therapeutic target for cardiovascular calcification – a discussion of its pleiotropic functions in the body. Cardiovasc Res 118, cvaa299-. 10.1093/cvr/cvaa299. [DOI] [Google Scholar]

[R25] 25.Millán JL (2006). Alkaline Phosphatases. Purinerg Signal 2, 335. 10.1007/s11302-005-5435-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] 26.Yang AC, Stevens MY, Chen MB, Lee DP, Stähli D, Gate D, Contrepois K, Chen W, Iram T, Zhang L, et al. (2020). Physiological blood–brain transport is impaired with age by a shift in transcytosis. Nature 583, 425–430. 10.1038/s41586-020-2453-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R27] 27.Raikwar NS, Bowen RF, and Deeg MA (2005). Mutating His29, His125, His133 or His158 abolishes glycosylphosphatidylinositol-specific phospholipase D catalytic activity. Biochem J 391, 285–289. 10.1042/bj20050656. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R28] 28.Scallon BJ, Fung W-JC, Tsang TC, Li S, Kado-Fong H, Huang K-S, and Kochan JP (1991). Primary Structure and Functional Activity of a Phosphatidylinositol-Glycan-Specific Phospholipase D. Science 252, 446–448. 10.1126/science.2017684. [DOI] [PubMed] [Google Scholar]

[R29] 29.Park MH, Lee JY, Park KH, Jung IK, Kim K-T, Lee Y-S, Ryu H-H, Jeong Y, Kang M, Schwaninger M, et al. (2018). Vascular and Neurogenic Rejuvenation in Aging Mice by Modulation of ASM. Neuron 100, 167–182.e9. 10.1016/j.neuron.2018.09.010. [DOI] [PubMed] [Google Scholar]

[R30] 30.Shi SM, Suh RJ, Shon DJ, Garcia FJ, Buff JK, Atkins M, Li L, Lu N, Sun B, Luo J, et al. (2025). Glycocalyx dysregulation impairs blood–brain barrier in ageing and disease. Nature 639, 985–994. 10.1038/s41586-025-08589-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] 31.Kumar SR, Miles TF, Chen X, Brown D, Dobreva T, Huang Q, Ding X, Luo Y, Einarsson PH, Greenbaum A, et al. (2020). Multiplexed Cre-dependent selection yields systemic AAVs for targeting distinct brain cell types. Nat. Methods 17, 541–550. 10.1038/s41592-020-0799-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] 32.Brichacek AL, Benkovic SA, Chakraborty S, Nwafor DC, Wang W, Jun S, Dakhlallah D, Geldenhuys WJ, Pinkerton AB, Millán JL, et al. (2019). Systemic inhibition of tissue-nonspecific alkaline phosphatase alters the brain-immune axis in experimental sepsis. Sci Rep-uk 9, 18788. 10.1038/s41598-019-55154-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R33] 33.Smith LK, Verovskaya E, Bieri G, Horowitz AM, Ungern‐Sternberg S.N.I. von, Lin K, Seizer P, Passegué E, and Villeda SA. (2020). The aged hematopoietic system promotes hippocampal‐dependent cognitive decline. Aging Cell 19, e13192. 10.1111/acel.13192. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R34] 34.Nakada-Honda N, Cui D, Matsuda S, and Ikeda E (2021). Intravenous injection of cyclophilin A realizes the transient and reversible opening of barrier of neural vasculature through basigin in endothelial cells. Sci. Rep. 11, 19391. 10.1038/s41598-021-98163-w. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R35] 35.Katsimpardi L, Litterman NK, Schein PA, Miller CM, Loffredo FS, Wojtkiewicz GR, Chen JW, Lee RT, Wagers AJ, and Rubin LL (2014). Vascular and Neurogenic Rejuvenation of the Aging Mouse Brain by Young Systemic Factors. Science 344, 630–634. 10.1126/science.1251141. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] 36.Stephan AH, Madison DV, Mateos JM, Fraser DA, Lovelett EA, Coutellier L, Kim L, Tsai H-H, Huang EJ, Rowitch DH, et al. (2013). A Dramatic Increase of C1q Protein in the CNS during Normal Aging. J. Neurosci. 33, 13460–13474. 10.1523/jneurosci.1333-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R37] 37.Hou Y, Dan X, Babbar M, Wei Y, Hasselbalch SG, Croteau DL, and Bohr VA (2019). Ageing as a risk factor for neurodegenerative disease. Nat. Rev. Neurol. 15, 565–581. 10.1038/s41582-019-0244-7. [DOI] [PubMed] [Google Scholar]

[R38] 38.Oakley H, Cole SL, Logan S, Maus E, Shao P, Craft J, Guillozet-Bongaarts A, Ohno M, Disterhoft J, Eldik LV, et al. (2006). Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: potential factors in amyloid plaque formation. J Neurosci Official J Soc Neurosci 26, 10129–10140. 10.1523/jneurosci.1202-06.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R39] 39.Kimura R, and Ohno M (2008). Impairments in remote memory stabilization precede hippocampal synaptic and cognitive failures in 5XFAD Alzheimer mouse model. Neurobiol Dis 33, 229–235. 10.1016/j.nbd.2008.10.006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R40] 40.Maguire GA, and Gossner A (1994). Glycosyl Phosphatidyl Inositol Phospholipase D Activity in Human Serum. Ann Clin Biochem 32, 74–78. 10.1177/000456329503200107. [DOI] [PubMed] [Google Scholar]

[R41] 41.Keren-Shaul H, Spinrad A, Weiner A, Matcovitch-Natan O, Dvir-Szternfeld R, Ulland TK, David E, Baruch K, Lara-Astaiso D, Toth B, et al. (2017). A Unique Microglia Type Associated with Restricting Development of Alzheimer’s Disease. Cell 169, 1276–1290.e17. 10.1016/j.cell.2017.05.018. [DOI] [PubMed] [Google Scholar]

[R42] 42.Ungvari Z, Tarantini S, Kiss T, Wren JD, Giles CB, Griffin CT, Murfee WL, Pacher P, and Csiszar A (2018). Endothelial dysfunction and angiogenesis impairment in the ageing vasculature. Nat Rev Cardiol 15, 555–565. 10.1038/s41569-018-0030-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R43] 43.Zlokovic BV (2011). Neurovascular pathways to neurodegeneration in Alzheimer’s disease and other disorders. Nat Rev Neurosci 12, 723–738. 10.1038/nrn3114. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R44] 44.Nation DA, Sweeney MD, Montagne A, Sagare AP, D’Orazio LM, Pachicano M, Sepehrband F, Nelson AR, Buennagel DP, Harrington MG, et al. (2018). Blood-brain barrier breakdown is an early biomarker of human cognitive dysfunction. Nat Med 25, 270–276. 10.1038/s41591-018-0297-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R45] 45.Sweeney MD, Sagare AP, and Zlokovic BV (2018). Blood–brain barrier breakdown in Alzheimer disease and other neurodegenerative disorders. Nat Rev Neurol 14, 133–150. 10.1038/nrneurol.2017.188. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R46] 46.Arvanitakis Z, Capuano AW, Leurgans SE, Bennett DA, and Schneider JA (2016). Relation of cerebral vessel disease to Alzheimer’s disease dementia and cognitive function in elderly people: a cross-sectional study. Lancet Neurology 15, 934–943. 10.1016/s1474-4422(16)30029-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] 47.Lee S, Viqar F, Zimmerman ME, Narkhede A, Tosto G, Benzinger TLS, Marcus DS, Fagan AM, Goate A, Fox NC, et al. (2015). White matter hyperintensities are a core feature of Alzheimer’s disease: Evidence from the dominantly inherited Alzheimer network. Ann Neurol 79, 929–939. 10.1002/ana.24647. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R48] 48.Ozek C, Krolewski RC, Buchanan SM, and Rubin LL (2018). Growth Differentiation Factor 11 treatment leads to neuronal and vascular improvements in the hippocampus of aged mice. Sci Rep-uk 8, 17293. 10.1038/s41598-018-35716-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] 49.Park MH, Lee JY, Park KH, Jung IK, Kim K-T, Lee Y-S, Ryu H-H, Jeong Y, Kang M, Schwaninger M, et al. (2018). Vascular and Neurogenic Rejuvenation in Aging Mice by Modulation of ASM. Neuron 100, 762. 10.1016/j.neuron.2018.10.038. [DOI] [PubMed] [Google Scholar]

[R50] 50.Leiter O, Brici D, Fletcher SJ, Yong XLH, Widagdo J, Matigian N, Schroer AB, Bieri G, Blackmore DG, Bartlett PF, et al. (2022). Platelet-derived exerkine CXCL4/platelet factor 4 rejuvenates hippocampal neurogenesis and restores cognitive function in aged mice. Nat. Commun. 14, 4375. 10.1038/s41467-023-39873-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R51] 51.Leiter O, Zhuo Z, Rust R, Wasielewska JM, Grönnert L, Kowal S, Overall RW, Adusumilli VS, Blackmore DG, Southon A, et al. (2023). Selenium mediates exercise-induced adult neurogenesis and reverses learning deficits induced by hippocampal injury and aging. Cell Metab. 35, 1085. 10.1016/j.cmet.2023.04.019. [DOI] [PubMed] [Google Scholar]

[R52] 52.Lourenco MV, Frozza RL, Freitas G.B. de, Zhang H, Kincheski GC, Ribeiro FC, Gonçalves RA, Clarke JR, Beckman D, Staniszewski A, et al. (2019). Exercise-linked FNDC5/irisin rescues synaptic plasticity and memory defects in Alzheimer’s models. Nat. Med. 25, 165–175. 10.1038/s41591-018-0275-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R53] 53.Islam MR, Valaris S, Young MF, Haley EB, Luo R, Bond SF, Mazuera S, Kitchen RR, Caldarone BJ, Bettio LEB, et al. (2021). Exercise hormone irisin is a critical regulator of cognitive function. Nat. Metab. 3, 1058–1070. 10.1038/s42255-021-00438-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] 54.Trejo JL, Carro E, and Torres-Aleman I (2001). Circulating insulin-like growth factor I mediates exercise-induced increases in the number of new neurons in the adult hippocampus. J. Neurosci. : Off. J. Soc. Neurosci. 21, 1628–1634. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R55] 55.Schroer AB, Ventura PB, Sucharov J, Misra R, Chui MKK, Bieri G, Horowitz AM, Smith LK, Encabo K, Tenggara I, et al. (2023). Platelet factors attenuate inflammation and rescue cognition in ageing. Nature 620, 1071–1079. 10.1038/s41586-023-06436-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] 56.Park C, Hahn O, Gupta S, Moreno AJ, Marino F, Kedir B, Wang D, Villeda SA, Wyss-Coray T, and Dubal DB (2023). Platelet factors are induced by longevity factor klotho and enhance cognition in young and aging mice. Nat. Aging 3, 1067–1078. 10.1038/s43587-023-00468-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R57] 57.Oh HS-H, Rutledge J, Nachun D, Pálovics R, Abiose O, Moran-Losada P, Channappa D, Urey DY, Kim K, Sung YJ, et al. (2023). Organ aging signatures in the plasma proteome track health and disease. Nature 624, 164–172. 10.1038/s41586-023-06802-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R58] 58.Li X, Shi X, McPherson M, Hager M, Garcia GG, and Miller RA (2022). Cap-independent translation of GPLD1 enhances markers of brain health in long‐lived mutant and drug‐treated mice. Aging Cell 21, e13685. 10.1111/acel.13685. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R59] 59.Li X, Hager M, McPherson M, Lee M, Hagalwadi R, Skinner ME, Lombard D, and Miller RA (2023). Recapitulation of anti-aging phenotypes by global, but not by muscle-specific, deletion of PAPP-A in mice. GeroScience 45, 931–948. 10.1007/s11357-022-00692-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R60] 60.Hager M, Chang P, Lee M, Burns CM, Endicott SJ, Miller RA, and Li X (2024). Recapitulation of anti-aging phenotypes by global overexpression of PTEN in mice. GeroScience 46, 2653–2670. 10.1007/s11357-023-01025-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R61] 61.Thrupp N, Frigerio CS, Wolfs L, Skene NG, Fattorelli N, Poovathingal S, Fourne Y, Matthews PM, Theys T, Mancuso R, et al. (2020). Single-Nucleus RNA-Seq Is Not Suitable for Detection of Microglial Activation Genes in Humans. Cell Rep. 32, 108189. 10.1016/j.celrep.2020.108189. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R62] 62.Chan KY, Jang MJ, Yoo BB, Greenbaum A, Ravi N, Wu W-L, Sánchez-Guardado L, Lois C, Mazmanian SK, Deverman BE, et al. (2017). Engineered AAVs for efficient noninvasive gene delivery to the central and peripheral nervous systems. Nat. Neurosci. 20, 1172–1179. 10.1038/nn.4593. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R63] 63.Aurnhammer C, Haase M, Muether N, Hausl M, Rauschhuber C, Huber I, Nitschko H, Busch U, Sing A, Ehrhardt A, et al. (2012). Universal Real-Time PCR for the Detection and Quantification of Adeno-Associated Virus Serotype 2-Derived Inverted Terminal Repeat Sequences. Hum. Gene Ther., Part B: Methods 23, 18–28. 10.1089/hgtb.2011.034. [DOI] [PubMed] [Google Scholar]

[R64] 64.Sanjana NE, Shalem O, and Zhang F (2014). Improved vectors and genome-wide libraries for CRISPR screening. Nat. Methods 11, 783–784. 10.1038/nmeth.3047. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R65] 65.Alamed J, Wilcock DM, Diamond DM, Gordon MN, and Morgan D (2006). Two-day radial-arm water maze learning and memory task; robust resolution of amyloid-related memory deficits in transgenic mice. Nat Protoc 1, 1671–1679. 10.1038/nprot.2006.275. [DOI] [PubMed] [Google Scholar]

[R66] 66.Stevenson ME, Bieri G, Kaletsky R, Ange J.St., Remesal L, Pratt KJB, Zhou S, Weng Y, Murphy CT, and Villeda SA. (2023). Neuronal activation of Gαq EGL-30/GNAQ late in life rejuvenates cognition across species. Cell Rep. 42, 113151. 10.1016/j.celrep.2023.113151. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R67] 67.Crouch EE, and Doetsch F (2018). FACS isolation of endothelial cells and pericytes from mouse brain microregions. Nat. Protoc. 13, 738–751. 10.1038/nprot.2017.158. [DOI] [PubMed] [Google Scholar]

[R68] 68.Sasmita AO, Ong EC, Nazarenko T, Mao S, Komarek L, Thalmann M, Hantakova V, Spieth L, Berghoff SA, Barr HJ, et al. (2025). Parental origin of transgene modulates amyloid-β plaque burden in the 5xFAD mouse model of Alzheimer’s disease. Neuron 113, 838–846.e4. 10.1016/j.neuron.2024.12.025. [DOI] [PubMed] [Google Scholar]

[R69] 69.Bieri G, Lucin KM, O’Brien CE, Zhang H, Villeda SA, and Wyss-Coray T (2018). Proteolytic cleavage of Beclin 1 exacerbates neurodegeneration. Mol Neurodegener 13, 68. 10.1186/s13024-018-0302-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R70] 70.Lin K, Bieri G, Gontier G, Müller S, Smith LK, Snethlage CE, White CW, Maybury-Lewis SY, and Villeda SA (2021). MHC class I H2-Kb negatively regulates neural progenitor cell proliferation by inhibiting FGFR signaling. Plos Biol 19, e3001311. 10.1371/journal.pbio.3001311. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

LIVER EXERKINE REVERSES AGING- AND ALZHEIMER’S-RELATED MEMORY LOSS VIA VASCULATURE

Gregor Bieri

Karishma JB Pratt

Yasuhiro Fuseya

Turan Aghayev

Juliana Sucharov

Alana M Horowitz

Amber R Philp

Karla Fonseca-Valencia

Rebecca Chu

Mason Phan

Laura Remesal

Shih-Hsiu J Wang

Andrew C Yang

Kaitlin B Casaletto

Saul A Villeda

SUMMARY

In brief statement

Graphical Abstract

INTRODUCTION

RESULTS

GPI-anchored TNAP is a GPLD1 substrate on the aged hippocampal vasculature

Figure 1. GPI-anchored TNAP is a GPLD1 substrate on the aged hippocampal vasculature.

Increasing liver-derived GPLD1 rejuvenates BBB function in the aged hippocampus

Figure 2. Increasing liver-derived GPLD1 rejuvenates BBB function in the aged hippocampus.

Mimicking the age-related increase in cerebrovascular TNAP disrupts hippocampal BBB function and impairs cognition

Figure 3. Mimicking the age-related increase in cerebrovascular TNAP disrupts BBB function and impairs cognition.

Restoring the age-related increase in cerebrovascular TNAP mitigates cognitive benefits of liver-derived GPLD1 in aging

Figure 4. Increasing cerebrovascular TNAP mitigates cognitive benefits of liver-derived GPLD1 while targeting TNAP reverses aging-related cognitive impairments.

Targeting cerebrovascular TNAP reverses aging-related cognitive impairments

Inhibiting TNAP activity recapitulates hippocampal transcriptional signatures of liver-derived GPLD1 in aging

Figure 5. Inhibiting TNAP activity recapitulates hippocampal transcriptional signatures of GPLD1 treatment in aging.

Liver-derived GPLD1 increases with exercise and rescues hippocampal transcriptional signatures in a model of AD pathology

Figure 6. Liver-derived GPLD1 increases with exercise and rescues hippocampal transcriptional signatures in models of AD pathology.

Increasing liver-derived GPLD1 or inhibiting TNAP activity in mice ameliorates AD pathology and cognitive deficits

Figure 7. Increasing liver-derived GPLD1 or inhibiting TNAP activity in mice ameliorates AD pathology and cognitive deficits.

DISCUSSION

Limitations of the study

RESOURCE AVAILABILITY

LEAD CONTACT

MATERIALS AVAILABILITY

DATA AND CODE AVAILABILITY

STAR METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mouse strains.

Human tissue.

Cell lines.

METHOD DETAILS

Exercise intervention.

Liver GPLD1 expression.

TNAP inhibition.

AAV production –

Abrogation of cerebrovascular TNAP in aged Cas9-inducible mice.

Novel object recognition (NOR).

Radial arm water maze (RAWM).

Y-maze.

Active Place Avoidance.

Health metrics.

Tissue collection.

Single nucleus RNA sequencing.

Brain endothelial cell isolation and single cell RNA sequencing.

Single cell and single nucleus RNA sequencing analysis.

RNA extraction, cDNA synthesis and RT-qPCR analysis.

Western blot analysis.

Immunohistochemistry.

Sulfo-NHS-biotin BBB leakage assay.

Labelled Transferring (TF647) uptake assay.

AP activity labelling.

Thioflavin S labelling.

TNAP cleavage assay.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data, statistical analyses, and reproducibility.

Supplementary Material

Key resources table.

Highlights.

ACKNOWLEDGMENTS

Footnotes

REFERENCES

Associated Data