The aged hematopoietic system promotes hippocampal‐dependent cognitive decline

Abstract The aged systemic milieu promotes cellular and cognitive impairments in the hippocampus. Here, we report that aging of the hematopoietic system directly contributes to the pro‐aging effects of old blood on cognition. Using a heterochronic hematopoietic stem cell (HSC) transplantation model (in which the blood of young mice is reconstituted with old HSCs), we find that exposure to an old hematopoietic system inhibits hippocampal neurogenesis, decreases synaptic marker expression, and impairs cognition. We identify a number of factors elevated in the blood of young mice reconstituted with old HSCs, of which cyclophilin A (CyPA) acts as a pro‐aging factor. Increased systemic levels of CyPA impair cognition in young mice, while inhibition of CyPA in aged mice improves cognition. Together, these data identify age‐related changes in the hematopoietic system as drivers of hippocampal aging.

pro-aging effects of old blood on the brain (Das et al., 2019;Smith et al., 2015;Villeda et al., 2011;Yousef et al., 2019). Despite the fact that the pro-aging factors identified to date are immune-related molecules (Smith, White, & Villeda, 2018), whether the aging hematopoietic system promotes hippocampal aging has not yet been investigated.
In mice and humans, the hematopoietic system undergoes many functional and structural changes during aging, characterized by myeloid expansion, decreased immunity, and chronic low-grade inflammation (Van Zant & Liang, 2012). We hypothesized that these cellular changes contribute to hippocampal aging through the accumulation of pro-aging immune factors in old blood. Many of the age-related changes observed in old blood have roots in hematopoietic stem cell (HSC) aging (Kim, Moon, & Spangrude, 2003;Pang et al., 2011).
Therefore, we employed a heterochronic HSC transplantation model to test how exposure to an aged hematopoietic system contributes to hippocampal aging ( Figure 1a). Young (2 months) recipient mice were sublethally irradiated (9 Gy) and transplanted with HSCs isolated from young (2 months) or old (24 months) donors, generating isochronic (Iso) and heterochronic (Het) HSC-reconstituted young mice. Blood chimerism was assessed by measuring the proportion of CD45.2 donor cells in CD45.1 recipient mouse blood by flow cytometry ( Figure S1a,b). Blood derived from old HSCs exhibited characteristic age-related myeloid bias 4.5 months post-transplantation (Figure S1c,d). Animals showed no signs of illness or weight loss regardless of treatment ( Figure S1e).
To investigate whether exposure to an old hematopoietic system elicits age-related cognitive impairments, we assessed hippocampal-dependent associative fear memory and spatial memory using contextual fear conditioning (FC) and radial arm water maze (RAWM) behavioral paradigms, respectively. During contextual FC, Het HSCreconstituted young mice exhibited decreased freezing during testing compared with Iso controls (Figure 1b), indicating impaired hippocampal-dependent memory. No differences were observed in baseline freezing during the training portion of the task ( Figure   S1f), or in the amygdala-dependent cued FC paradigm ( Figure 1c).
During RAWM, all mice showed similar swim speeds ( Figure S1g) and learning capacity during the training portion of the task ( Figure   S1h). Moreover, exposure to an aged hematopoietic system did not result in robust impairments in spatial memory during the testing portion of the task ( Figure S1h). These data indicate that the aged hematopoietic system drives impairments in hippocampal-dependent associative fear memory.
We next sought to assess cellular changes in the hippocampi of Het HSC-reconstituted young mice that might contribute to cognitive decline. Hippocampal neurogenesis has been shown to decline with aging and after exposure to an aged systemic milieu (Villeda et al., 2011). Therefore, we tested whether exposure to an old hematopoietic system impairs neurogenesis by immunohistochemical analysis 4.5 months post-HSC transplantation (Figure 1d,e). Although no difference was observed between groups in the number of Sox2+/ GFAP+neural stem cells in the dentate gyrus (DG), young mice exposed to an old hematopoietic system had decreased numbers of doublecortin (DCX)-positive immature neurons compared with controls (Figure 1d,e). To assess neuronal differentiation and survival, we employed a long-term bromodeoxyuridine (BrdU) incorporation paradigm, in which mature neurons derived from proliferating stem cells retain BrdU labeling (Figure 1a). We observed decreased numbers of BrdU+cells co-labeled with the mature neuronal marker NeuN in the DG of Het HSC-reconstituted young mice compared with Iso controls (Figure 1d,e). These data indicate that exposure to an old hematopoietic system impairs hippocampal neurogenesis.
We next investigated whether exposure to an aged hematopoietic system elicits synaptic changes in the hippocampus at a molecular and structural level. Activation of the transcription factor, CREB, via phosphorylation, has been implicated in age-related cognitive decline and rejuvenation (Villeda et al., 2014;Yu, Curlik, Oh, Yin, & Disterhoft, 2017). Correspondingly, we assessed the levels of CREB phosphorylation (pCreb) and observed a decrease in Het HSC-reconstituted young mice compared with Iso controls (Figure 1f,g). Moreover, we observed decreased expression of the synaptic markers AMPAR and the NMDA receptor subunit NR2B ( Figure 1f,g), both of which have previously been shown to decline with hippocampal aging (Shi et al., 2007;Wheatley et al., 2019). At a structural level, we examined dendritic spine density in granule cell neurons and observed a decrease in Het HSC-reconstituted young mice (Figure 1h,i). Together, these data indicate that the old hematopoietic system leads to an age-related decrease in synaptic density in the hippocampus.
Immunohistochemical identification of CD45.2+ hematopoietic cells in the DG of CD45.1 recipient mice revealed low and equivalent levels of immune cell infiltration in Het and Iso HSCreconstituted young mice ( Figure S2a,b). While we cannot exclude the possible contribution of these small numbers of peripheral immune cells, we hypothesized that the pro-aging effects of the old hematopoietic system are predominantly mediated through peripheral changes in circulating blood factors.
We performed unbiased proteomic analysis on blood plasma collected from Het and Iso HSC-reconstituted young mice 4.5 months post-transplantation. Using label-free mass spectrometry, we identified 22 factors that were differentially expressed between Het and Iso HSC-reconstituted young mice ( Figure 1j). To identify potential pro-aging factors, we focused analysis on proteins that were upregulated ≥1.5-fold in the plasma of Het HSC-reconstituted young mice (dotted line, Figure 1j). Of these, the most significantly upregulated cytokine was cyclophilin A (CyPA, encoded by Ppia)-an intracellular protein-containing peptidyl-prolyl cis-trans isomerase activity that is secreted in response to inflammatory stimuli (Nigro, Pompilio, & Capogrossi, 2013). Elevated CyPA plasma levels were confirmed in Het HSC-reconstituted mice by Western blot analysis ( Figure S3a,b).
In an independent cohort of naïve mice, we detected an age-related increase in CyPA expression in blood cells by qPCR ( Figure 1k). No age-related changes in CyPA expression were observed in other tissues ( Figure 1k). Examination of a previously published human RNAseq dataset from the GTEx project (Yang et al., 2015) demonstrated a similar increase in CyPA in human blood cells during aging ( Figure 1l).
We next analyzed the relationship between CyPA plasma levels and contextual FC performance in Iso and Het HSC-reconstituted mice.
We found an inverse correlation between CyPA levels and cognitive performance ( Figure 1m) positing CyPA as a potential pro-aging factor with relevance to cognition.
To test whether increasing systemic CyPA levels are sufficient to elicit age-related cognitive or cellular impairments, we utilized a hydrodynamic tail vein injection (HDTVI) in vivo transfection approach, wherein young (2 months) mice were intravenously injected with overexpression constructs encoding either CyPA or GFP control ( Figure 2a). The HDTVI-mediated increase in plasma CyPA levels was confirmed using a HiBiT-tagged CyPA and luminescent detection approach ( Figure S3c). Animals showed no signs of illness or weight loss regardless of treatment ( Figure S4a). One month after HDTVI, we assessed hippocampal-dependent object recognition and associative fear memory using novel object recognition (NOR) and contextual These data indicate that targeting extracellular CyPA at old age is sufficient to improve object recognition memory, promote neurogenesis, and increase the levels of key proteins important for synapse function.

F I G U R E 2 Cyclophilin
Cumulatively, our data demonstrate that age-related changes in the hematopoietic system promote molecular, cellular, and cognitive hallmarks of hippocampal aging. We identify CyPA as a pro-aging factor whose expression is elevated in the blood of het  (Ohtsuki et al., 2017;Satoh et al., 2013).
In these studies, CyPA plasma levels were also found to be elevated with aging (Ohtsuki et al., 2017;Ramachandran et al., 2014;Satoh et al., 2013). While little is known about the role of CyPA in aging, recent proteomic analysis using mass spectrometry has identified CyPA as part of the senescence-associated secretory phenotype (SASP) (Wiley et al., 2019). Ultimately, our data identify the aged hematopoietic system, and downstream circulating immune factors, as potential therapeutic targets to restore cognitive function in the elderly.

| Flow cytometry
HSCs were isolated as previously described (Ho et al., 2017). Bone marrow was obtained from leg, arm, and pelvic bones, isolated in

| BrdU administration
For long-term BrdU labeling studies, 50 mg/kg of BrdU (Sigma-Aldrich) was injected intraperitoneally into mice daily for 5 days.
Mice were euthanized 33 days after first BrdU injection.

| Immunohistochemistry
Tissue processing and immunohistochemistry were performed on free-floating sections according to standard published techniques (Smith et al., 2015). Mice were anesthetized with 87.5 mg/ kg ketamine and 12.5 mg/kg xylazine and transcardially perfused with 1× phosphate-buffered saline (PBS

| Golgi staining and analysis
Mice were anesthetized with 87.5 mg/kg ketamine and 12.5 mg/ kg xylazine, and brains were removed without perfusion. Golgi  Horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence (ECL) kit (GE Healthcare) were used to detect protein signals. Membranes were imaged using a ChemiDoc system (Bio-Rad) and quantified using ImageJ software (Version 1.52a). GAPDH bands were used for normalization for hippocampal lysates. Equal loading of plasma was confirmed using Ponceau S solution (Sigma-Aldrich). HiBiT was detected using the Nano-Glo

| qPCR
Tissue was dissected, snap-frozen, and total RNA was ex-  Yang et al. (2015) identified age-related gene expression changes in human RNA-seq data collected by the Genotype-Tissue Expression (GTEx) project (Ardlie et al., 2015). They examined tissue expression changes in nine tissues: subcutaneous adipose, tibial artery, left ventricle heart, lung, skeletal muscle, tibial nerve, skin, thyroid, and whole blood (n/tissue = 83-156 subjects, ages 20-70 years). Using a linear regression model, Yang et al. generated an aging coefficient for each gene detected in each tissue. Aging coefficients that are significantly >0 after false discovery rate adjustments (q < 0.05) are upregulated with age. We examined how

| Tissue expression of CyPA (human)
CyPA expression changes with age in their dataset by graphing the aging coefficient of CyPA against the -log10 adjusted q-value.

| Plasmid generation
RNA was isolated from adult mouse spleen tissue using TRIzol reagent (Thermo Fisher Scientific) and PureLink™ RNA Mini Kit following the manufacturer's instructions. The RNA concentration was determined via NanoDrop, and RNA was reverse-transcribed using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific) and oligodT primers (Promega). The following primers were used for PCR amplification of the PPIA-coding sequences and partial 3′ and 5′ untrans- Endotoxin-free plasmid kits were used for plasmid preparation prior to in vivo use.

| Hydrodynamic tail vein injection
The hydrodynamic tail vein injection protocol was adapted from a previously described protocol (Kovacsics & Raper, 2014). Endotoxinfree plasmids were prepared using the Qiagen Maxi-Prep Plus Kit (VWR). CyPA, CyPA-HiBiT, or GFP plasmid DNA (50 μg) was suspended in 10% body weight saline and injected in the tail vein in 5-7 s in young mice. For detection of HiBiT-tagged CyPA, blood was drawn 24 hr post-HDTVI. Following a similar timeline as previous pro-aging studies (Smith et al., 2015), behavioral testing was started 33 days post-HDTVI.

| Radial arm water maze
Spatial learning and memory were assessed using the radial arm water maze (RAWM) paradigm according to established protocol (Alamed, Wilcock, Diamond, Gordon, & Morgan, 2006). In this task, the location of the goal arm, which contains a platform, remains constant throughout the training and testing phase, while the start arm is changed during each trial. On day 1 during the training phase, mice are trained for 15 trails, with trials alternating between a visible and hidden platforms. On day 2 during the testing phase, mice are tested for 15 trials with a hidden platform. Entry into an incorrect arm is scored as an error, and errors are averaged over training blocks (three consecutive trials).

| Contextual fear conditioning
In this task, mice learned to associate the environmental context (fear-conditioning chamber) with an aversive stimulus (mild foot shock; unconditioned stimulus, US) enabling testing for hippocampal-dependent contextual fear conditioning. As contextual fear conditioning is hippocampal-and amygdala-dependent, the mild foot shock was paired with a light and tone cue (conditioned stimulus, CS) in order to also assess amygdala-dependent cued fear conditioning. Conditioned fear was displayed as freezing behavior.
Specific training parameters are as follows: Tone duration is 30 s; level is 70 dB, 2 kHz; shock duration is 2 s; and intensity is 0.6 mA.
This intensity is not painful and can easily be tolerated but will generate an unpleasant feeling. On day 1, each mouse was placed in a fear-conditioning chamber and allowed to explore for 2 min before delivery of a 30-s tone (70 dB) ending with a 2-s foot shock (0.6 mA). Two minutes later, a second CS-US pair was delivered.
On day 2, each mouse was first placed in the fear-conditioning chamber containing the same exact context, but with no CS or foot shock. Freezing was analyzed for 2 min. One hour later, the mice were placed in a new context containing a different odor, cleaning solution, floor texture, chamber walls, and shape. Animals were allowed to explore for 2 min before being re-exposed to the CS.
Freezing was analyzed for 30 s after re-exposure to CS. Freezing was measured using a FreezeScan video tracking system and software (Clever Sys, Inc) or EthoVision XT 11.5 tracking software (Noldus) and a Ugo Basile FC system.

| Novel object recognition
The novel object recognition task was adapted from a previously described protocol (Dubal et al., 2015). During the habituation phase (day 1), mice could freely explore an empty arena for 10 min. During the training phase (day 2), mice were exposed to two identical objects (either two striped scintillation vials or two lego constructions).
Mice were allowed to explore the objects for 5 min. Mice that did not explore the familiar objects for more than 3 s during the training phase were excluded from analysis. During testing (day 3), one familiar object was replaced with a novel object (vial replaced with lego or lego replaced with vial), and mice could explore for 5 min. Time spent exploring each object was quantified using the Smart Video Tracking Software (Panlab; Harvard Apparatus).

| Plasma collection
At time of euthanasia, mouse blood was collected via intracardial bleeds into EDTA-coated tubes. Plasma was generated by centrifugation of freshly (<30 min) isolated blood at 1,000 g. Aliquots were stored at −80°C until use.

| Label-free mass spectrometry
For proteomic analysis of old hematopoietic system associated factors, plasma was collected from young Iso and Het HSCreconstituted that had exhibited >60% or <50% freezing in con-  (Tyanova, Temu, & Cox, 2016), which incorporates the Andromeda search engine. Using this program, the MS data were recalibrated, protein/peptide identification was made using the Andromeda database search engine, the database search results were filtered at the 1% protein and peptide FDR, and protein levels were quantified. The resulting MaxQuant output was further processed using Perseus (V 1.6.0.7; Max Planck Institute for Biochemistry). Two samples exhibiting clotting changes as described in Geyer et al. (2016) were excluded from analysis.
Differentially expressed genes were identified by comparing LFQ intensities for each protein detected in Heterochronic HSCreconstituted mice compared with isochronic HSC-reconstituted controls.

| Data and statistical analysis
Mice were randomized prior to treatment. Researchers were blinded throughout histological, biochemical, and behavioral assessments.
Groups were un-blinded at the end of each experiment upon statistical analysis. Data are expressed as mean + SEM. The distribution of data in each set of experiments was tested for normality using the D'Agostino-Pearson omnibus test or Shapiro-Wilk test. No significant differences in variance between groups were detected using an F test.
Statistical analysis was performed with Prism 6.0 software (GraphPad Software). Means between two groups were compared with twotailed, unpaired Student's t test. Comparisons of means from multiple groups with each other or against one control group were analyzed with one-way ANOVA followed by appropriate post hoc tests (indicated in figure legends). The data that support the findings of this study are available from the corresponding author upon reasonable request. Cytometry Core, supported in part by Grant NIH P30 DK063720 and by the NIH S10 Instrumentation Grant S10 1S10OD021822-01.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing financial interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.