Stroke-Induced Neurological Dysfunction in Aged Mice Is Attenuated by Preconditioning with Young Sca-1⁺ Stem Cells

Abstract

Aims

To date, stroke remains one of the leading causes of death and disability worldwide. Nearly three-quarters of all strokes occur in the elderly (>65 years old), and a vast majority of these individuals develop debilitating cognitive impairments that can later progress into dementia. Currently, there are no therapies capable of reversing the cognitive complications which arise following a stroke. Instead, current treatment options focus on preventing secondary injuries, as opposed to improving functional recovery.

Methods

We reconstituted aged (20-month old) mice with Sca-1⁺ bone marrow (BM) hematopoietic stem cells isolated from aged or young (2-month old) EGFP⁺ donor mice. Three months later the chimeric aged mice underwent cerebral ischemia/reperfusion by bilateral common carotid artery occlusion (BCCAO), after which cognitive function was evaluated. Immunohistochemical analysis was performed to evaluate host and recipient cells in the brain following BCCAO.

Results

Young Sca-1⁺ cells migrate to the aged brain and give rise to beneficial microglial-like cells that ameliorate stroke-induced loss of cognitive function on tasks targeting the hippocampus and cerebellum. We also found that young Sca-1⁺ cell-derived microglial-like cells possess neuroprotective properties as they do not undergo microgliosis upon migrating to the ischemic hippocampus, whereas the cells originating from old Sca-1⁺ cells proliferate extensively and skew toward a pro-inflammatory phenotype following injury.

Conclusions

This study provides a proof-of-principle demonstrating that young BM Sca-1⁺ cells play a pivotal role in reversing stroke-induced cognitive impairments and protect the aged brain against secondary injury by attenuating the host cell response to injury.

Graphical Abstract

Significance Statement.

Young Sca-1⁺ hematopoietic stem cells give rise to anti-inflammatory microglia-like cells in the aged brain after injury. Aged mice reconstituted with young Sca-1⁺ hematopoietic stem cells prior to transient ischemia reperfusion demonstrate significantly improved cognitive and motor function performance and recovery after injury. Aged donor cells, but not young, undergo microgliosis in the ischemic brain and contribute to secondary injury.

Introduction

Stroke is the second leading cause of death and third leading cause of disability worldwide.¹ Of all stroke occurring annually, the vast majority of them are classified as ischemic strokes, with nearly 80% of all strokes occurring in the elderly population.² Indeed, aging is a major risk factor for stroke, which doubles in incidence every 10 years after the age of 55.² This is particularly concerning as the past half a decade has for the first time seen a greater number of elderly individuals than children under the age of 5.³ Despite an increase in overall longevity, quality of life remains poor for the elderly. This is especially true regarding functional outcomes after stroke, as to date there is a lack of effective therapies to treat cognitive and motor impairments that arise in the elderly following this injury.^4,5

Although the mechanisms underlying greater stroke incidence and impaired recovery being more prominent among the elderly are unknown, leading theories suggest age-dependent immune dysfunction (ie, inflammaging) to be a main driver of this.⁶ Indeed, it is widely accepted that the immune system and brain-resident microglia are known to play a key role in aging and response to injury.⁷ Microglia are the resident macrophages in the brain, which during development originate from the yolk sac and establish themselves around the same time as neurons.⁸ They are highly plastic cells, enabling them to act as the first line of defense to the pathophysiological changes induced by ischemic stroke.⁹ Under normal conditions, they are in a “resting” state, where they use their processes to survey their micro-environment.¹⁰ Traditionally, it has been thought that microglial activation is deleterious in ischemic stroke. However, growing evidence suggests that microglia may play a dual role, as they can secrete both pro- and anti-inflammatory factors, which are critical for neurogenesis, angiogenesis, and synaptic remodeling; all of them can impact functional recovery after cerebral ischemia.^11,12 Activated microglia, characterized by their retraction of their processes and production of inflammatory factors, are generally categorized as M0, M1, or M2, based on their functional profiles. M0 represents the resting state, M1 classical activation, which promotes inflammatory responses, and M2 alternate activation, which promotes tissue remodeling, wound healing, and immune regulation.^13,14 Following ischemic injury, there is intense neuron-microglia crosstalk in response to neuronal damage, leading to early activation of microglia to engulf cellular debris and regulate neuronal function via various neurotransmitters and modulators.¹⁵ For example, ATP and CX3CL1 released from damaged neurons are able to bind to receptors to activate microglia, which in turn induced further neuro-cytotoxicity via promoting M1 classical activation.^16-19 Additionally, LCN2 released by injured neurons serves as a “help-me” signal which can polarize microglia into the M2 pro-tissue repair phenotype, via enhancing IL-10 expression and phagocytic activity. Therefore, in the aftermath of ischemic stroke, neuronal damage initiates numerous activation processes among microglia. Activated microglia, in turn, carry out a number of beneficial functions essential for neuron survival; on the other hand, microglial activation also give rise to progressive neurotoxic consequences, owing to the simultaneous excess production of a large array of cytotoxic factors.

Accumulating evidence also suggests that bone marrow (BM) stem and precursor cells possess the ability to migrate to the brain and ultimately give rise to functional microglial-like cells under normal and pathological conditions.^20,21 Although little is understood about the controversial role that BM-derived microglial-like cells play in the brain, transplantation of BM-derived mononuclear cells were shown to attenuate learning and memory deficits in injured and neurologically impaired rodents.^22,23 Despite these promising findings, the invasive nature of most cell therapies and limited therapeutic potential due to rapid cell death and dispersion following injection demands alternative avenues of treatment. To address this issue, we have previously shown that continuous migration of cells to the brain can be achieved by BM reconstitution via a minimally invasive tail vein injection following BM clearance.²⁴ Recently, it has also been demonstrated that aged recipient mice of young whole BM, but not old, showed improved recovery of locomotory function following a middle cerebral artery occlusion, whereas in contrast, aged cells injected into healthy young recipient mice acquired a senescent profile that is normally observed after injury and with aging.²⁵ Indeed, it is accepted that inflammatory responses in the brain become exacerbated after injury with age, and that this may often lead to further cognitive complications. Currently, it remains unclear as to how: (1) aged animals benefit from young and not old BM cells after injury and (2) whether young BM cells can ameliorate hippocampal-dependent learning and memory impairments in aged mice after stroke.

In this study, we have identified that reconstitution of aged BM using young Sca-1⁺ stem cells purified from whole BM improved recovery of cognitive function following cerebral ischemia in aged mice. Moreover, recipients of aged Sca-1⁺ stem cells displayed severe host and donor cell immune activation marked by extensive proliferation and hypertrophy. Our findings suggest that recovery of function in aged individuals after stroke was inhibited by over-activation of aged central- and peripheral-derived glial cells. This detrimental immune cell response and impaired functional outcome after stroke in aged recipients can be ameliorated by introducing young Sca-1⁺ stem cell-derived microglial-like cells.

Materials and Methods

Animals

Young (2-month old) and middle-aged (12-month old) C57BL/6-Tg(CAG-EGFP)1Osb/J and C57BL/6 wild-type (WT) mice were obtained from Jackson Laboratories. Mice were bred and aged to 20 months within our animal housing facility. All mice were kept on a 12-h light/dark cycle and provided with free access to both food and water.

Donor Cell Preparation

Young and old (20-month old) BM was obtained from the femurs and tibiae of C57BL/6-Tg-GFP mice. The bones were flushed with sterile PBS and then incubated in RBC lysis buffer (154.42 mM NH₄Cl, 11.9 mM NaHCO₃, 0.026 mM EDTA) for 5 minutes, followed by centrifugation at 1000 rpm for 5 minutes. The pellet was then suspended in Iscove’s Modified Dulbecco’s Medium (Thermo Fisher Scientific) and passed through a 40-µm filter. Cells were counted and Sca-1⁺ cells isolated using immunomagnetic activated cell sorting according to the manufacturer’s instructions (STEMCELL Technologies).

Bone Marrow Reconstitution

Female C57BL/6J mice (20-month old) were subjected to whole-body irradiation at 10 Gy (Cs-137 irradiator, Gammacell 40 Exactor), after which 2 × 10⁶ young or old GFP⁺/Sca-1⁺ cells were injected via the tail vein, generating Y⁺-O and O⁺-O chimeras, respectively. The irradiated mice were then used for the experiments at 3 months post-BM reconstitution.

Bilateral Common Carotid Artery Occlusion

All procedures carried out were performed as previously described.²⁶ Briefly, regional brain ischemia was induced by bilateral common carotid arteries occlusion (BCCAO) for 15 minutes under 2% isoflurane in wild-type mice and mice 3 months post-reconstitution. The animals were then placed in a sterile cage sitting on top of a heating blanket (37 °C) until they woke up. Control animals were sham-operated without occlusion. Buprenorphine was given every 12 h post-surgery, for up to 48 h. Brains from control and experimental animals were collected at either 3 days or 2 weeks after stroke.

Behavioral Tests

The open field test apparatus consisted of a 38 × 60 × 60 cm chamber with grey Plexiglass walls and transparent ceiling to allow for video recording. Mice were placed in the chamber for 10 minutes and their ambulatory distance and rearing count was tracked and analyzed by idTracker.²⁷ Additional analyses were performed using a custom python script, which transformed x- and y-coordinates associated to animal behavior into the various parameters reported in the results section. The novel object recognition task was performed as previously described.²⁸ In short, mice underwent 3 consecutive days of testing, each separated by a whole 24 h. The first day (habituation phase) involved the animals roaming freely in a novel environment for 10 minutes under video surveillance. The second day (familiarization phase) introduced 2 identical copies of the same object, while the third and final day (testing phase) introduced one copy of an object from the second day and one novel object that the mice have never seen before. For both days, the mice were allowed to freely interact with each object for a total duration of 10 minutes and the interaction time for each object was separately measured and recorded. Recognition memory performance was defined by calculating the discrimination index (ie, time spent interacting with novel object versus time spent interacting with familiar object). Objects were randomized between mice to determine which would be the familiar set and which would be the novel set, and were washed with 75% ethanol between animals in order to eliminate odor. Spatial reference learning and memory was carried out as described previously.²⁹ Briefly, the Barnes maze consisted of an elevator platform (30 cm above ground) with 20 equidistance holes and spatial cues around the circumference. Each mouse was given 90 s to find the correct hole leading to a dark chamber. The total time taken and the number of errors made before finding the correct hole was measured. Mice ran the maze under light stimulation for a total of 4 trials/day, for 5 consecutive days, with at least 20 min between each trial. The maze was cleaned with 75% ethanol between each trial and if the mice could not complete the maze within 90 s, they were gently guided to the escape hole by hand. Object location memory was performed as described.³⁰ In short, this behavioral test was performed similarly to novel object recognition mentioned previously; however, the testing phase involved shifting the location of similar objects, as opposed to introducing a novel object. Location memory was quantified by calculating the discrimination index derived from object interaction time. Finally, the rotarod performance and hanging limb test were performed, as previously described.^31,32

Immunofluorescence Staining

Brains were extracted 3 days and 2 weeks after ischemia/reperfusion and fixed in 4% paraformaldehyde overnight at 4 °C before preservation in increasing concentrations of 10%, 20%, and 30% sucrose in PBS (w/v). Brains underwent stereological sectioning coronally at a thickness of 20 µm on a freezing-sliding microtome, and sections were stored in cryoprotective medium (OCT, Dako) at −20 °C. Frozen tissue sections were blocked with 5% (w/v) donkey serum before incubation at room temperature for 2 hours with primary antibodies at the following concentrations: GFAP (1:200, Z0334, Dako), Iba1 (1:400, PA5-27436, Thermo Fisher Scientific), GFP (1:400, A21311, Life Technologies), iNOS (1:200, N32030/L19, BD Transduction Laboratories), Arg-1 (1:200, SC-18351, Santa Cruz Biotech), Ki67 (1:200, ab16667, Abcam), and CD45 (1:200, 550539, BD Pharmingen). Incubation with respective Alexa 488, 568, or 647 conjugated secondary antibodies (Life Technologies) was carried out at room temperature for 1 hour with light protection. Cell nuclei were identified with 4ʹ,6-diamidino-2-phenylindole (DAPI) at 1:2000 for 5 minutes. Stained tissue slides were stored at −4 °C under dark conditions.

Cell Quantification

Cell quantification was carried out on all positive cells of interest present within a specific region of the brain (ie, hippocampus), or within randomly selected regions of the cortex (1.0 mm² cross-sections), and were averaged across 3 replicates. Fluorescent images were obtained with a Nikon Eclipse Ti fluorescent microscope, or using a VS120 slide scanner (Olympus). Mean fluorescence intensity was carried out using ImageJ (image processing and analysis software in Java) by calculating mean gray values of fluorescent signal for each separate channel (FITC or TRITC) across identically sized cross-sectional areas of brain tissue. Prior to analysis, images were captured at an identical exposure time using the Olympus cellSens imaging software. Proportion and number of cells based on shape and size were quantified automatically using built-in cell counting tools in the Olympus cellSens imaging software.

In Situ End Labeling of DNA Fragmentation (TUNEL) and Propidium Iodide (PI) Staining

For TUNEL staining, brains were removed as before, and 20 µm coronal sections of the hippocampus, at 2 weeks after stroke, were obtained. Tissue sections were first incubated in permeabilization solution (0.1% Triton X-100, 0.1% sodium citrate, PBS), followed by incubation for 2 hours at 37 °C using the In Situ Cell Death Detection Kit, TMR Red (Sigma-Aldrich), according to the manufacturer’s instructions. Sections were then incubated for 5 minutes in 1:2000 DAPI followed by immediate application of mounting media and coverslip. For PI staining, a 1:1000 dilution of PI (P4864, Sigma-Aldrich) was applied to hippocampal sections at room temperature for 20 minutes, followed by incubation for 5 minutes in 1:2000 DAPI. Viewing of DNA fragmentation and cell viability was performed with a Nikon Eclipse Ti fluorescent microscope under the TRITC filter.

Statistical Analysis

All values are expressed as mean ± s.e.m. Analyses were performed using GraphPad Prism 6.0 software. Students two-tailed t-test was used for 2-group comparison, whereas one-way ANOVA was used for 3 or more group comparisons. Statistical analysis of cell quantification and behavioral data was carried out using an unpaired, 2-tailed t-test. Two-way ANOVA was carried out for 2-factor variables followed by Tukey post hoc tests. P < .05 indicated statistical significance.

Results

Aging Alters Immunological Response and Functional Recovery After Stroke

In this study, we investigated the impact that aging has on central nervous system repair and recovery, following ischemic injury induced by BCCAO. BCCAO has been shown to induce neuronal death in the hippocampus, leading to behavioral deficits in learning and memory.^26,33,34 Two weeks following BCCAO, both young and old WT mice exhibited pronounced behavior deficits (Fig. 1a). Representative tracking of animals in the open field test used for quantification is shown in Fig. 1b which revealed that BCCAO reduced overall ambulatory activity in young and old animals compared with age matched sham controls (Fig. 1c). However, aged animals showed the lowest levels of ambulatory activity following BCCAO (Fig. 1c). This was also associated with a marked increase in anxiety, as shown by decreased entries into the central region (Fig. 1d) and percentage of time spent inside the central zone during the open field test (Fig. 1e). Next, we assessed memory and learning associated behaviors, using the novel object recognition and Barnes maze tests. Aged animals also showed significant impairment in novel object recognition and were not able to distinguish between test objects (Fig. 1f). In contrast, young mice maintained novel object recognition capacity; suggesting greater preservation of cognitive function in these mice. Assessment of spatial learning and memory using the Barnes maze revealed both age and injury-dependent effects on maze performance, with old BCCAO mice exhibiting significantly impaired spatial memory compared to all other groups (Fig. 1g, h). Finally, motor coordination/grip strength, as assessed by rotarod, revealed significantly impaired motor coordination in old versus young BCCAO mice (Fig. 1i).

Figure 1.

Cognitive function declines with age and ischemic stroke. (a) Schematic describing the injury model and time course for behavioral testing. (b) Open-field test depiction of activity levels of young and old mice, for sham surgery and BCCAO. (c) Ambulatory distance quantified for the open-field test. (d) Thigmotaxis quantified by the number of entries made into the central zone (green square) and the (e) overall activity in this zone. (f) Novel object recognition task 2 weeks following BCCAO. (g) Barnes maze spatial learning and memory paradigm for total completion time and (h) average number of errors before finding the correct escape hole. (i) Rotarod performance test for identifying motor coordination, endurance, and grip strength. Analysis was performed 2 weeks following BCCAO.

Following ischemic injury in the brain, there is a coordinated cellular repair response, consisting of cell infiltration from the periphery and activation of the resident cell populations.³⁵ Therefore, we examined whether age-dependent differences in cognitive performance following BCCAO was associated with differences in cellular responses. To identify age- and injury-induced differences in glial cell numbers, we stained cross-sections of the hippocampus for microglia (Iba1⁺ cells) and astrocytes (GFAP⁺). In line with previous reports,^36,37 we found that microglial (Extended Data Fig. 1a, b) and astrocyte (Extended Data Fig. 1a, c) numbers significantly increased with age, as well as following ischemic stroke. Taken together, these findings demonstrate that 15 minutes BCCAO is sufficient to induce hippocampal- and cerebellar-dependent behavioral changes, as well as differential levels of glial cell response, after injury in an age-dependent manner.

Sca-1⁺ Stem Cell-Derived Cells Migrate from the BM to the Aged Ischemic Hippocampus

We previously demonstrated that young Sca-1⁺ stem cells promote functional recovery of the aged heart after myocardial infarction,^38-40 and rectify radiation-induced injury of the brain following transplantation.²⁴ Moreover, we and others have shown that transplanting young BM stem cells into aged mice leads to the recruitment of young cells from the BM to the brain, where they take on a microglia-like phenotype.^23-25 Therefore, the presence of young BM cells in aged mice may affect peripheral and central responses to injury. We next evaluated whether the transplantation of young BM stem cells into aged mice could lead to improved cognitive function after cerebral infarction. Aged mice were reconstituted with either young or old Sca-1⁺ BM stem cells expressing green fluorescent protein (GFP), generating respective chimeras (Fig. 2a). Three months after reconstitution animals were subjected to BCCAO (Fig 2a.). Two weeks following ligation, GFP⁺ cells were present in different regions of the hippocampus (CA1-CA3 and dentate gyrus) of both groups, with a significantly greater total number of cells seen in O⁺-O animals (Fig. 2b-f). Morphometric analysis revealed cells exhibiting branching morphologies in O⁺-O animals, whereas donor cells in Y⁺-O mice were primarily ovoid (Fig. 2b-e, g). Immunofluorescent labelling of GFP⁺ cells in the ischemic hippocampus revealed that these cells express immune (CD45⁺) and microglial-like (Iba1⁺) markers (Extended Data Fig. 2a, b). Interestingly, the opposite trend was observed at 72 h following injury, whereby Y⁺-O mice demonstrate a greater number of GFP⁺ cells in the hippocampus when compared to O⁺-O mice (Extended Data Fig. 3a, b). Moreover, ovoid cells predominate 72 h following injury in both groups (Extended Data Fig. 3c), rather than the branching morphologies seen 2 weeks after ligation. Taken together, these results demonstrate that cells from the BM can be directed to specifically home to the injured hippocampus and acquire a microglial-like morphology. Moreover, these data show that aging alters the temporal cellular responses following injury, as cellular response is dampened in O⁺-O mice compared with Y⁺-O mice at 3 days post-BCCAO, whereas at 2 weeks post-BCCAO, O⁺-O exhibited increased number of GFP⁺ cells, suggesting impaired resolution of injury responses.

Figure 2.

Young and old GFP⁺/Sca-1⁺ stem cells migrate from the BM to the hippocampus after cerebral ischemia. (a) Timeline of BM reconstitution, stroke, and behavioral analysis. (b-e) Immunofluorescence of entire hippocampus and dentate gyrus (DG) from aged recipients of young (Y⁺-O) and old (O⁺-O) Sca-1⁺ stem cells, at 2 weeks after BCCAO or sham surgery, respectively. (f, g) Quantification of total number and proportion of GFP⁺ cells in the hippocampus, according to different regions and morphology after stroke. Scale bars = 200 µm for enlarged regions and 20 µm for inserts.

Figure 3.

Young Sca-1⁺ stem cell reconstituted animals exhibit a decreased number of host immune and glial cells in the aged hippocampus following injury. Immunostaining of the hippocampus following ischemia for CD45⁺ myeloid immune cells (a); GFAP⁺ astrocytes (b); and Iba1⁺ microglia (c), as well as mean fluorescent intensity of the hippocampus for the different cell types. Scale bars are 100 µm. Analysis was performed 2 weeks following BCCAO.

Aged Mice Reconstituted with Young Sca-1⁺ Stem Cells Show Reduced Host Cell Response to Stroke in the Injured Hippocampus

Past studies show that repair and response to injury declines with age.⁴¹ Here, we found that aged mice treated with young Sca-1⁺ stem cells prior to cerebral infarction exhibited an attenuated immune and glial host cell response within the hippocampus, when compared with mice treated with old Sca-1⁺ stem cells (Fig. 3). This is shown by significantly lower intensity of host immune cells (Fig. 3a), astrocytes (Fig. 3b), and microglia (Fig. 3c) in the infarct region of Y⁺-O vs. O⁺-O, at 2 weeks following BCCAO. Interestingly, although CD45⁺ and Iba-1⁺ were different between sham animals, GFAP⁺ cells were only different between Y⁺-O and O⁺-O at 2 weeks following BCCAO (Fig 3b). This suggests that the donor cells impacted host responses following injury.

Young Sca-1⁺ Stem Cells Reduce Neuronal Death in the Aged Ischemic Hippocampus

Neuronal death in the hippocampus after stroke has been linked to poor outcomes in the elderly population, due to the limited reparative capacity of aged tissue.⁴² To determine whether neuronal death after stroke is age dependent, we examined cell viability (propidium iodide) and apoptosis (TUNEL) in the ischemic hippocampus of young and old wild-type mice. Quantification of PI⁺ (Extended Data Fig. 4a, b) and TUNEL⁺ (Extended Data Fig. 4a, c) cells in the hippocampus, at 2 weeks post-BCCAO, in young and old mice, confirmed an age-dependent effect on neuronal cell death following ischemic injury, as cell death was significantly greater in aged vs. young mice (Extended Data Fig 4d, e). Next, we examined whether BM reconstitution with young Sca-1⁺ stem cells can confer neuroprotective effects after injury. In support of this notion, we found that Y⁺-O mice had significantly lower PI⁺ and TUNEL⁺ intensity, compared to O⁺-O, at 2 weeks after BCCAO (Fig. 4a-d). Consistent with previous findings,²⁵ our results demonstrate potential neuroprotective properties exhibited by young cells on neuronal survivability. Importantly, we also found that Y⁺-O mice exhibited lower TUNEL⁺ signals in the hippocampus at 3 days following ischemia, compared to O⁺-O mice (Extended Data Fig. 5a, b). Collectively, these data support that young Sca-1⁺ cells limited cell death in aged recipients following ischemic injury.

Figure 4.

Young Sca-1⁺ stem cells reduce neuronal death in the hippocampus following stroke. (a, b) Fluorescent staining of cell viability (propidium iodide; PI) and apoptosis (TUNEL) for reconstituted animals before and after stroke. Mean fluorescence intensity quantified for PI (c) and TUNEL (d). Scale bars = 100 µm (a), 200 µm (b, c) and 20 µm for insets (b). Analysis was performed 2 weeks following BCCAO.

Figure 5.

Aged animals reconstituted with young Sca-1⁺ stem cells show improved motor and cognitive function following ischemic injury. (a) Visual illustration of open-field test with (b) quantified activity levels and (c) travelling speed, for aged animals reconstituted with young and old Sca-1⁺ stem cells. (d-f) Thigmotaxis quantified by number of entries into central zone (green square) and proportion of time spent active in inside and outside zones, respectively. (g) Rotarod performance test and (h) hanging limb test for motor coordination and grip strength. i, Exploratory initiative/curiosity measurements obtained during the open-field test. j, k, Barnes maze spatial learning and memory test, depicting maze completion time and average number of errors made prior to finding the correct escape chamber, respectively. l, m, Novel object recognition and object location spatial memory, respectively. Analysis was performed 2 weeks following BCCAO.

Mice Reconstituted with Young Sca-1⁺ Stem Cells Show Improved Motor and Cognitive Function After Stroke

Hemiparesis is the most common deficit experienced following stroke, affecting over 80% of patients suffering from this injury, with recovery lasting up to several years.^42-44 Since previous studies found that BM cells have the potential to migrate to the cerebellum in greater numbers than any other region of the brain,⁴⁵ and these cells have been shown to benefit muscle repair after injury,⁴⁶ we were curious to determine whether reconstitution with young Sca-1⁺ stem cells could benefit motor function recovery after stroke. We found that aged animals reconstituted with young Sca-1⁺ stem cells after stroke displayed greater locomotive activity (Fig. 5a-c), exploratory initiative (Fig. 5d-f), motor endurance/coordination (Fig. 5g), and grip strength (Fig. 5h). Taken together, aged mice after stroke demonstrate significantly improved mobility and fine motor control when reconstituted with young Sca-1⁺ stem cells, relative to aged mice reconstitution with old Sca-1⁺ stem cells.

Since cognitive complications after stroke primarily arise in elderly patients, and not young ones, we were also curious to examine the effects of the age of donor cells on hippocampal-dependent memory and learning abilities in healthy and injured animals. To study this, we examined learning and memory in animals reconstituted with either young or old Sca-1⁺ stem cells, using the novel object recognition task and Barnes maze 2 weeks after BCCAO or sham surgery. In line with previous data reported by our group and others, we found that Y⁺-O mice outperformed O⁺-O mice on hippocampal-dependent behavioral paradigms for sham surgery and BCCAO groups at curiosity/exploratory initiative (Fig. 5i), spatial learning and memory (Fig. 5j, k, m), and novel object recognition (Fig. 5l).

Young and Old Donor Sca-1⁺ Stem Cells Give Rise to Differentially Activated Cells in Response to Injury in the Aged Brain

Numerous studies have demonstrated that age can influence tissue responses to injury and dictate subsequent cellular phenotype and functions.^47-49 In order to gain mechanistic insight into why young cells better preserve hippocampal function following injury, we further classified GFP⁺ cells in the aged brain according to markers commonly expressed by anti- and pro-inflammatory macrophages, arginase-1 (Arg-1) and inducible nitric oxide synthase (iNOS), respectively. In line with past work,²⁴ we found that GFP⁺ cells in the brain originating from young and old BM Sca-1⁺ stem cells were primarily Arg-1⁺ and iNOS⁺ in the cortex, at baseline and 2 weeks following cerebral ischemia, respectively (Fig. 6a, b, f). Additionally, we found that old donor cells were primarily iNOS⁺ in the injured aged hippocampus, but were unable to detect significant differences in differentially activated states among young GFP⁺ cells, due to their limited number in the ischemic hippocampus at this time point (Fig. 6c-e).

Figure 6.

Young and old Sca-1⁺ stem cells give rise to anti- and pro-inflammatory cells in the injured brain. (a, b) Immunostaining of BM-derived GFP⁺ cells in the cortex and hippocampus (c, d) for Arg-1 (red) and iNOS (gray) in reconstituted mice subjected to sham surgery and BCCAO. (e, f) Quantified number of BM-derived GFP⁺ cells expressing pro- and anti-inflammatory markers in the hippocampus and cortex for experimental and control reconstituted mice. Scale bars are (a, b) 200 µm and 20 µm (insets) and (c, d) 100 µm. Analysis was performed 2 weeks following BCCAO.

Old Donor BM Sca-1⁺ Stem Cells, But not Young, Give Rise to Microglial-Like Cells that Undergo Extensive Gliosis in the Aged and Injured Hippocampus

Healthy aged cells are known to have an increased tendency of undergoing gliosis, known as microgliosis for microglia, and astrogliosis for astrocytes, during homeostasis and after injury.^50-52 To determine whether old Sca-1⁺ stem cells give rise to microglial-like cells that undergo microgliosis, which we previously demonstrated express the microglial marker Iba1, we examined their proliferation (Ki-67), cell size, and quantity in the ischemic hippocampus. We found that GFP⁺ cells originating from old BM Sca-1⁺ stem cells, but not young, were: (1) hypertrophic, (2) proliferated extensively at the site of injury, and (3) greater in cell number in the hippocampus (Fig. 7a, b). Taken together, these results demonstrate an important observation that BM cells can migrate to the brain under pathological conditions and influence the repair and recovery process. Moreover, donor BM cells of old origin which migrate to the brain exhibit signs of microgliosis, while young cells retain a pro-reparative phenotype, even when transplanted into the aged environment.

Figure 7.

Aged Sca-1⁺ stem cell-derived cells, but not young, undergo microgliosis in the ischemic hippocampus. (a) Immunostaining of the hippocampus after stroke for cellular proliferation (red, Ki-67) and BM-derived GFP⁺ cells. (b) Quantification of total number and proportion of GFP⁺-proliferating cells, and their size in the hippocampus, post-sham surgery and BCCAO. Scale bars = 200 and 20 µm (insets). Analysis was performed 2 weeks following BCCAO.

Discussion

In this study, we showed that BM reconstitution with either young or old Sca-1⁺ stem cells results in the migration of these cells from the BM to the hippocampus after ischemia and gives rise to CD45⁺ and Iba1⁺ microglial-like cells. Aged recipients of young Sca-1⁺ stem cells had preserved spatial and recognition memory, as well as significantly reduced neuronal death in the hippocampus after stroke. In contrast, aged recipients of old Sca-1⁺ stem cells did not recover stroke-induced loss of learning and memory, and the resulting BM-derived microglial-like cells acquired an inflammatory phenotype, which was also observed in host cells surrounding the ischemic hippocampus. Finally, cells originating from old Sca-1⁺ stem cells underwent extensive microgliosis after injury. This phenomenon was also observed in host astrocytes and microglia in recipients of old, but not young, Sca-1⁺ stem cells at 2 weeks after stroke. Comparatively, young BM-derived microglial-like cells displayed an acute response to injury, as they were present in larger numbers 72 h following injury, compared with their numbers in the hippocampus 2 weeks post-BCCAO.

Gliosis is a natural and fundamental process that normally occurs after injury and ensures proper tissue repair by mitigating the spread of insult to surrounding healthy cells and tissue.⁵³ However in the aged brain, gliosis could result in secondary injury that may confer greater downstream complications, compared with the initial trauma.⁵⁴ The main cause of this has yet to be elucidated; however, microglial priming has been recently reported as being one possible factor responsible for this phenomenon.^55,56 Here, injury leads to an over-exaggerated immune response that is typically characterized by cellular hypertrophy, immune activation, and extensive proliferation of brain resident glial cells. This in turn has detrimental effects on healthy neurons, as it leads to synaptic pruning, impeded synaptic plasticity, as well as cytokine-driven neuronal death and impairments of functional recovery following trauma or infection.⁵⁴ In this study, we report similar observations, as well as new insights demonstrating that aged BM cells, but not young, seemingly initiate a cascade of negative responses in host cells following their infiltration to the brain in response to injury.

Although it is not entirely clear in what way old Sca-1⁺ stem cell-derived microglial-like cells have an effect on host cells in the brain following injury, heterochronic parabiosis studies provide valuable insight into numerous beneficial and detrimental circulating factors present in young and old blood, which exert effects on rejuvenating global function and life span in aged recipients.^57-59 Since this discovery, many of these blood-borne factors have been identified as different chemokines and cytokines, such as CCL11 and GDF11, respectively, as well as MMP-9 and its inhibitor, TIMP-1, and have been implicated in aging and disease.^60,61 The latter is notably important as it plays a function during memory formation and its expression levels have been shown to increase with age in both mice and humans following stroke.^25,62 MMP-9 normally remodels components of the extracellular matrix, although after cerebral ischemia, it plays a role in blood-brain barrier disruption and excitotoxic neuronal loss.⁶³ One of the main sources of MMP-9 after stroke has been attributed to BM-derived neutrophils.⁶³ Interestingly, aged whole BM predominantly gives rise to greater neutrophil migration to the brain after injury, whereas young BM cells are mostly involves monocyte migration to the brain.²⁵ Moreover, stroke-responsive neutrophils have been found to produce reactive oxygen species,⁶⁴ and this finding may explain why the vast majority of cells originating from old Sca-1⁺ stem cells are positive for the pro-inflammatory marker, iNOS, and become activated upon migrating to ischemic tissue.

Indeed, nitric oxide is a known potent inducer of DNA damage that has been shown to ultimately lead to cellular apoptosis.⁶⁵ In contrast to this, cells in the brain after stroke that are derived from young Sca-1⁺ stem cell are positive for the anti-inflammatory Arg-1 marker.²⁴ Arg-1 has been extensively described for its many beneficial effects during tissue repair and neuroprotection following injury.⁶⁶ Recently, Arg-1 was shown to activate a pro-survival gene expression profile in surrounding cells by inhibiting oxidative stress induced death, such as those mediated by nitric oxide.⁶⁷ In addition, arginase deficiency has been described to result in neurological impairments, dementia, and axonal degeneration in both humans and animal models.^68-70 When taken together with previous findings, our data suggest that young BM-derived Sca-1⁺ stem cells migrate to the brain after stroke and differentiate into neuroprotective Arg-1⁺ cells that attenuate the response to injury and supposedly prevent further insult to the hippocampus. Although we have identified that Arg-1 polarization is the most likely mechanism underlying the beneficial effects described within this study, it is probable that several other BM-derived paracrine factors might also play pivotal roles in aiding brain recovery after injury. For example, we previously demonstrated that hypoxia-induced expression of VEGF and TGF-β1 was greater in young versus old Sca-1⁺ cells, indicating the presence of putative age-related differences among BM cell secretomes. Mechanistically, differences in secretome function are in part influenced by a decline in autophagy in old, compared with young, Sca-1⁺ cells, which negatively impacts the secretion of key bioactive factors involved in regulating numerous cellular processes.⁷¹ Future studies using genomic and protein profiling at the single-cell level are required to delineate mechanistic insights into the age-related decline in BM stem cell efficacy, which could be targeted to improve the function of old donor cells.

Several therapeutic approaches have been developed to improve repair following ischemic stroke, some of which focus on remodeling the injured brain by altering the activation phenotype of microglia. Currently, 4 broad approaches have been proposed, eg, small molecules, partial MHC class II constructs, stem cell therapy, and microglia transplantation, all of which have been applied for treating both acute and chronic effects of stroke.⁷² More specifically, all these methods aim to diminish the number of pro-inflammatory M1 microglia and increase microglial polarization toward the M2 phenotype.^72,73 With respect to stem cell therapy, neural and BM cell transplantation have been shown to benefit functional recovery in animal stroke models, through mechanisms involving greater angiogenesis, and endogenous neurogenesis.^74-76 However, further work is needed to determine how cell transplantation modulates microglia activation.

Aging also impacts the resting state of microglia, and can impact their responses to injury, as studies have shown that aged microglia exhibit alterations in overall transcriptional activation after cerebral ischemia, with marked impairment in inflammatory responses, immune cell chemotaxis, tissue remodeling, and cell-cell interactions.⁷⁷ Aged microglia also exhibit reduced interaction with neighboring neurons and diminished polarity toward the infarct lesion following cerebral ischemia.⁷⁷ These changes in gene expression and cell phenotype add another layer of complexity in the development of effective therapeutic approaches, which aim to modulate the activation state of microglia. In this study we demonstrate that young BM Sca-1⁺ cells can differentiate into microglia-like cells with anti-inflammatory and neuroprotective properties, which were able to ameliorate the stroke-induced loss of cognitive function. By contrast, old Sca-1⁺ cells favor a pro-inflammatory phenotype, leading to microgliosis. Future studies will focus on identifying the key factors, or epigenetic states, in young Sca-1⁺ cells favoring the microglia polarization toward the beneficial M2 state. The applications of these factors or induction of old cells into a younger state is a promising new avenue for manipulating microglial phenotype and improving repair following cerebral ischemia.

There are limitations to this study, one of which is the usage of whole-body radiation, as the brain ends up undergoing several drastic cellular and molecular changes, along with disrupting the blood-brain barrier and increasing its permeability, all of which could be confounding factors for evaluating the effectiveness of young Sca-1⁺ cells.⁷⁸ However, it is important to note that all old recipients were subject to whole-body radiation before receiving young or old Sca-1⁺ cells, but only the Y⁺-O chimeras exhibited neuroprotective effects, implying that the cells were most likely responsible for the ensuing cognitive improvements. Follow-up investigations should prioritize on developing a method which depletes aged HSCs, and allows engraftment of young HSCs, without requiring the usage of irradiation.

Conclusion

We demonstrate that BM reconstitution with young donor Sca-1⁺ stem cells gives rise to beneficial microglia-like cells in the aged brain that restore ischemic injury-induced loss of motor and cognitive function. Moreover, young donor cells attenuate their own and host cell responses to injury, whereas old donor cells undergo microgliosis and likely contribute to secondary injury.

Funding

This work was supported by a grant from the Canadian Institutes of Health Research (332652) to R.K.L.

Conflict of Interest

The authors declared no potential conflicts of interest.

Author Contributions

L.W.: conception and design, collection, and/or assembly of data, data analysis and interpretation, manuscript writing. F.J.A.: conception and design, manuscript writing, provision of study material or patients, data analysis and interpretation. J.W.: provision of study material or patients. S.-H.L.: manuscript writing, provision of study material or patients. R.-K.L.: financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.

Data Availability

All data are available upon reasonable request from the principal investigator.

Ethical Approval

The Animal Care Committee of the University Health Network approved all experimental procedures, which were carried out according to the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, 2011).