Author contributions: J.S.B.: designed all experiments; J.K.L.: performed the experiments and wrote the manuscript; H.K.J. and J.S.B.: supervised the project; E.H.S. and J.S.B.: edited and reviewed the manuscript.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS May 2, 2012.
Microglia have the ability to eliminate amyloid β (Aβ) by a cell-specific phagocytic mechanism, and bone marrow (BM) stem cells have shown a beneficial effect through endogenous microglia activation in the brains of Alzheimer's disease (AD) mice. However, the mechanisms underlying BM-induced activation of microglia have not been resolved. Here we show that BM-derived mesenchymal stem cells (MSCs) induced the migration of microglia when exposed to Aβ in vitro. Cytokine array analysis of the BM-MSC media obtained after stimulation by Aβ further revealed elevated release of the chemoattractive factor, CCL5. We also observed that CCL5 was increased when BM-MSCs were transplanted into the brains of Aβ-deposited AD mice, but not normal mice. Interestingly, alternative activation of microglia in AD mice was associated with elevated CCL5 expression following intracerebral BM-MSC transplantation. Furthermore, by generating an AD-green fluorescent protein chimeric mouse, we ascertained that endogenous BM cells, recruited into the brain by CCL5, induced microglial activation. Additionally, we observed that neprilysin and interleukin-4 derived from the alternative microglia were associated with a reduction in Aβ deposition and memory impairment in AD mice. These results suggest that the beneficial effects observed in AD mice after intracerebral SC transplantation may be explained by alternative microglia activation. The recruitment of the alternative microglia into the brain is driven by CCL5 secretion from the transplanted BM-MSCs, which itself is induced by Aβ deposition in the AD brain. Stem Cells201230:1544–1555;
The activation of microglial cells is consistently detected in the brains of Alzheimer's disease (AD), but the precise role of these cells in the pathogenesis of AD has not been resolved [ 1–3]. While some studies have demonstrated that in some environments microglia can promote disease [ 4, 5], many other studies indicate that they are beneficial and protective [ 6, 7]. For example, microglia can clear amyloid β (Aβ) plaques via phagocytosis and secrete neurotrophic factors [ 6, 8, 9]. However, as AD progresses microglia display dysfunctional morphology and exhibit a significant reduction of Aβ degradation factors [ 10, 11].
Activation of the systemic innate immune system and/or bone marrow (BM) stem cell transplantation may therefore be powerful approaches to treat AD and to eliminate Aβ from the central nervous system (CNS) [ 12]. Accumulating evidence implicates disruption within the hematopoietic system as an initiating event and/or contributing to the progression of AD [ 13, 14], and a few studies have shown that the activation of hematopoietic progenitor cells and the recruitment of these cells into the brain could be a viable therapeutic strategy for AD. For example, in specific pathological conditions, BM-derived hematopoietic cells can migrate into the brain to become new microglia [ 15, 16], and these microglia are more effective at Aβ phagocytosis than resident microglia in AD mice [ 6, 17, 18].
Mesenchymal stem cells (MSCs) initially attracted interest for their ability to differentiate into multiple cellular phenotypes. However, recent concepts regarding the therapeutic effects of MSCs have been broadened to include the secretion of biologically active molecules that exert beneficial effects on other cells [ 19–21]. Such paracrine effects of MSCs include the release of trophic, immunomodulatory, and chemoattractant factors [ 22]. Therefore, we have suggested that critical influences of secreted bioactive and immunomodulatory factors from transplanted MSCs in the AD brain microenvironment will produce a significant therapeutic benefit in this disorder. This hypothesis has particular clinical significance since it might broaden the opportunity for MSC replacement therapy in AD, where the widespread neuropathological changes have been previously considered a limiting factor to success. These and related observations have focused attention on evaluating the therapeutic, paracrine effects of BM-derived MSCs after transplantation into AD mice. Our published data have also shown that transplanted BM-MSCs did not undergo transdifferentiation despite their beneficial effects in AD mouse models, including microglial activation and Aβ phagocytosis, improvement of memory impairment [ 23, 24], and promotion of neuronal survival through the release of soluble factors in primary cultured Purkinje neurons [ 25]. These results further emphasize the possible positive, paracrine effects of BM-MSCs following transplantation. Although the positive effects of BM-MSC transplantation in AD mice were associated with endogenous microglia activation, it was not clear how BM-MSCs enhanced microglia recruitment and which soluble bioactive factors derived from BM-MSCs could activate these cells in AD.
Based on these concepts and findings, we speculated that microglia, activated by BM-MSCs, might in fact derive from endogenous BM, and that soluble factors released by transplanted BM-MSCs in the brains of amyloid precursor protein (APP) and presenilin 1 (PS1) AD mice might recruit these cells into the brain. This study was designed to determine whether BM-MSCs have the capacity to produce biological molecules recruiting BM-derived cells, and if so, how these molecules might lead to a beneficial effect in AD.
MATERIALS AND METHODS
Transgenic mouse lines overexpressing the hAPP695swe (APPswe) and presenilin-1M146V (PS1) mutations, respectively, were generated at GlaxoSmithKline (Harlow, U.K., http://www.gsk.com/research/index.html) by standard techniques on a C57BL/6 background (Charles River, Margate, U.K., http://criver.com). APPswe mice were backcrossed onto a pure C57BL/6 background before crossing with PS1 mice to produce double heterozygous mutant mice (APP/PS1) [ 24]. Green fluorescent protein (GFP) transgenic (C57BL/6-Tg (ACTB-EGFP)1Osb/J) mice were purchased from the Jackson Laboratory (Bar Harbor, ME; http://www.jax.org). All experiments were approved and performed in accordance with institutional guidelines.
APP/PS1-GFP Chimeric Mice
APP/PS1 mice (recipients, 6-month-old) were exposed to 10 Gy whole body irradiation (2 × 5 Gy) except in the brain, where they received 5 Gy head irradiation (1 × 5 Gy) [ 26]. Donor BM cells (1 × 107 per mouse) derived from GFP mice were administrated via tail vein to each recipient. Transplanted mice were given drinking water complemented with 0.2 mg/ml trimethoprim and 1 mg/ml sulfamethoxazole for 2 weeks. Chimeric mice were confirmed by blood smears from tail clippings for the presence of GFP 5 weeks after the BM transplantation. Also, BM cells and peripheral blood were harvested and red blood cells and debris were examined after density centrifugation using Ficoll (Stem cell Technology, Vancouver, BC, Canada, http://www.stemcell.com). The engraftment of donor-derived BM cells and peripheral blood mononuclear cells was evaluated by flow cytometric analysis.
Transwell Chamber Migration Assay
The transwell chamber migration assay was carried out using transwell cell culture inserts (Corning, NY, http://www.corning.com, 5-μM pore size) or the Cytoselect 24-well cell migration assay kit (Cell Biolabs, Inc., San Diego, CA, http://www.cellbiolabs.com). Briefly, the bottom chamber contained either conditioned medium (CM) from BM-MSCs or aggregated Aβ 42-treated (10 μM) BM-MSC CM. For some experiments, recombinant murine CCL5 (R & D systems, Minneapolis, MN, http://rndsystems.com) or CM derived from NIH 3T3 cells was included in the bottom chamber. 1 × 105 microglia or monocytes were placed in the top chamber in serum-free Dulbecco's modified Eagle's medium or RPMI 1640 (Gibco, Grand Island, NY, http://www.invitrogen.com). Following exposure for 4 hours, cells that had migrated across the insert membrane were quantified. CCL5 small interfering RNA (siRNA)-treated BM-MSC CM also was used in the bottom chamber and examined following the above protocol.
Three days before the first injection with BM-MSCs, the mice were anesthetized with a combination of 100 mg/kg ketamine and 10 mg/kg xylazine, and a stainless steel cannula was implanted in the animal's hippocampus using a stereotaxic frame (David Kopf Instrument, Tujunga, CA, http://www.kopfinstruments.com). The guide cannula was fixed in the hippocampal brain region according to the following coordinates: 1.6 mm posterior to the bregma, 1.7 mm bilateral to the midline, and 1.2 mm ventral to the skull surface. In the absence of treatment, the guide cannula contains an obturator that prevents the aperture from being filled in by surrounding tissue. BM-MSC suspensions, CCL5 knockdown BM-MSCs, or phosphate buffered saline (PBS) were transplanted biweekly for 1 month (n = 10 per group). Mice were transplanted starting at 7 months 2 week of age and finished at 8 months 2 week of age (three times biweekly by intracerebral injection). Three microliters of the cell suspension (approximately 1 × 104 cells) was injected into the hippocampus bilaterally. The cell suspension was delivered at a rate of 0.3 μl/minute. After surgery, each mouse was kept in an individual cage to prevent the removal of the guide cannula by other animals. APP/PS1-GFP chimeric mice (n = 10 per group) were treated by the same protocol.
The Morris water maze (MWM) task was used to assess spatial memory performance. The water maze was a white tank (1.0 m diameter, 30 cm height) filled to a depth of 20 cm (22°C–24°C). White opaque nontoxic paint was added to the water to hinder visibility. A submerged Plexiglas platform (10 cm diameter; 6–8 mm below the surface of the water) was located at a fixed position throughout the training session. The position of the platform was varied from mouse to mouse while being counterbalanced across experiment groups. All mice were habituated to the maze 1 day before training. The animals were subjected to four trials per day. A training session consisted of a series of four trials per day for 10 consecutive days (total 40 trials). In each of the four trials, the animals were placed at different starting positions equally spaced around the perimeter of the pool in a random order. The mouse was given 60 seconds to find the submerged platform. If the mouse did not mount the platform within 60 seconds, it was guided to the platform. The time to mount the platform was recorded as the latency for each trial. Mice were allowed to remain on the platform for 10 seconds before being returned to a home cage. A single probe trial, in which the platform was removed, was performed after the hidden platform task had been completed (day 11). Each mouse was placed into one quadrant of the pool and allowed to swim for 60 seconds. All trials were recorded using a charge-coupled device camera connected to a video monitor and a computer. The test was run using Image J software. All apparatus used in this study were made by O'Hara & Company (Tokyo, Japan, http://www.ohara-time.co.jp).
CM from BM-MSCs Stimulated by Aβ Induces Migration of Microglia
To examine whether soluble factors released from BM-MSCs exhibited chemoattractive effects following exposure to Aβ, a transwell migration assay was performed using BM-MSC CM. The migration assay protocol is described in Figure 1A. We first found that BV2 microglia migrated to wells containing both BM-MSC CM as well as to (non-Aβ-stimulated) CM from NIH 3T3 cells. Of note, the BM-MSC CM significantly enhanced microglia migration compared to NIH 3T3 CM and control media (prepared without cells) (Supporting Information Fig. S1). To examine this effect further, we stimulated BM-MSCs with 10 μM aggregated Aβ 42 for 24 hours and then used the CM from these cells for the migration assays (Fig. 1). Non-Aβ-stimulated BM-MSC CM significantly induced migration of BV2 cells compared to control media containing 10 μM pure Aβ 42 alone (Fig. 1B). Moreover, Aβ-stimulated BM-MSC CM further induced migration of BV2 microglial cells compared with the nonstimulated BM-MSC CM (Fig. 1B). Dose-dependent migration of BV2 microglia in response to Aβ-treated BM-MSC CM was observed (Fig. 1C).
Previous studies have suggested that formation of actin-filled projections is a critical step during cell migration [ 27]. Therefore, we examined the effect of Aβ-treated BM-MSC CM on cytoskeletal reorganization in BV2 microglia. Actin stress fiber formation was significantly increased in BV2 cells when they were stimulated with Aβ-treated BM-MSC CM compared to control media (Fig. 1D, 1E).
CCL5 Derived from BM-MSCs and Activated by Aβ is a Critical Factor that Promotes Microglia Migration
In order to identify the chemotactic cytokines that were upregulated in BM-MSCs after Aβ stimulation, we screened and compared the CM of nonstimulated and Aβ-stimulated BM-MSCs for 40 different secreted cytokines using an antibody-based mouse cytokine array (Supporting Information Fig. S2). The cell-free supernatant of Aβ-stimulated BM-MSCs induced stronger signals in six array spots in comparison to the supernatant of nonstimulated BM-MSCs. CM derived from BM-MSCs exposed to Aβ showed higher levels of CXCL1, macrophage colony-stimulating factor (M-CSF), macrophage inflammatory protein (MIP)-2, MIP-1β, CCL5, and tumor necrosis factor (TNF)-α (Fig. 2A). Of the selected cytokines, only CCL5 levels were significantly elevated in the mRNA of BM-MSCs after Aβ stimulation (Fig. 2B).
To confirm the secretion of CCL5 in BM-MSCs exposed to Aβ, we performed ELISA. The results showed that the CCL5 protein levels were higher in the Aβ-stimulated BM-MSC CM compared to nonstimulated BM-MSC CM (Fig. 2C). Although the CCL5 expression patterns of NIH 3T3 cell CM after Aβ treatment were similar to BM-MSCs, the total amount of CCL5 secretion from NIH 3T3 cells was significantly less than from BM-MSCs (Supporting Information Fig. S3). We subsequently examined whether CCL5 derived from the BM-MSCs could function as a chemoattractant and might be responsible for the microglia migration we had observed. At 100 ng/ml, recombinant murine CCL5 significantly promoted BV2 cell migration when compared with control media (Supporting Information Fig. S4).
To further confirm that CCL5 derived from BM-MSCs was important in promoting microglia migration, we used siRNA to knockdown CCL5 expression in BM-MSCs. After 48 hours of transfection with a construct expressing CCL5 siRNA, the CCL5 mRNA and protein content in BM-MSCs were decreased 90% and 39%, respectively, compared to control siRNA-treated BM-MSCs (Supporting Information Fig. S5A, S5B). We also observed less CCL5 mRNA (80% decrease) and protein (58% decrease) after Aβ stimulation in CCL5 siRNA-treated BM-MSCs compared to control siRNA-treated BM-MSCs.
CM was collected from BM-MSCs with or without Aβ stimulation and CCL5 knockdown, and microglia migration assays were then performed. CM from non-Aβ-stimulated BM-MSCs induced significant BV2 microglia migration compared to the control media (Fig. 2D). This migration effect was lower when CM of CCL5 knockdown BM-MSCs was tested, although this did not reach statistical significance. Aβ-stimulated BM-MSC CM obtained from cells after CCL5 knockdown also showed a similarly reduced influence on cell migration as the nonstimulated cells (Fig. 2D). Similar effects were observed when primary microglia (as opposed to BV2 cells) were tested (Fig. 2E).
CCL5 Derived from BM-MSCs Following Transplantation is a Critical Factor to Recruit Endogenous Microglia
In order to examine the in vivo effects of BM-MSCs on microglia migration in AD, we used APP/PS1 mice with Aβ depositions. The treatment protocol is described in Figure 3A and in Materials and Methods section. At 2 weeks after the last BM-MSC transplantation, we observed that CCL5 mRNA was significantly increased in the transplanted mice compared with PBS-infused counterparts (Fig. 3B). BM-MSC-transplanted WT mice also showed slight elevation of CCL5 expression compared with PBS-transplanted mice, but this did not reach statistical significance and was less than that of the APP/PS1 mice (Fig. 3B). The increased expression of CCL5 observed in the BM-MSC-treated mice was more significant in the hippocampus than the cortex, which might be related to the fact that the BM-MSCs were transplanted into the hippocampus. For this reason, subsequent experiments and analysis were focused on the hippocampus.
To examine whether the increased CCL5 levels were associated with microglia activation, we first investigated recruitment of microglia by counting Iba-1-positive cells in the BM-MSC-treated and nontreated mice. In BM-MSC-treated mice, the number of microglia was significantly increased compared with PBS-treated mice (Fig. 3C). However, in mice transplanted with CCL5 knockdown BM-MSCs, microglia recruitment was significantly lower than that of control BM-MSC-infused mice (Fig. 3C).
To examine whether the increased microglia following BM-MSC treatment were derived from endogenous BM cells or transdifferentiation of the transplanted MSC, we constructed chimeric mice by irradiating 6-month-old APP/PS1 mice and intravenously injecting BM cells collected from GFP mice [ 26]. At 5 weeks after the BM transplantation, chimeric mice were confirmed by the presence of GFP in BM cells and peripheral blood monocytes using flow cytometric analysis (data not shown). At 2 weeks after the last BM-MSC injection, brain sections were taken and the number of GFP-positive cells in the hippocampal region was estimated using stereological analysis. As expected, injection of BM-MSCs led to a significant increase of GFP-positive cells (Fig. 3D), indicating that the BM-MSC treatment induced migration of endogenous BM cells across the blood-brain barrier. However, AD mice that were transplanted with CCL5 knockdown BM-MSCs exhibited significantly fewer GFP-positive cells compared with mice injected with non-knockdown BM-MSCs (Fig. 3D).
To assess the percentage of GFP-positive cells that were microglia, we separated CD11b-positive cells by magnetic-activated cell sorting (MACS) and by flow cytometry analysis using brain suspensions from AD-GFP chimeric mice 14 days after the last BM-MSC transplantation. The relative levels of CD45 can distinguish microglia (CD45 intermediate) from macrophages (CD45 high). Our results revealed that the percentage of GFP-positive CD45-intermediate (GFP+/CD45dim) microglia obtained from the separated CD11b-positive cells were increased in the BM-MSC-treated GFP chimeric mice compared with the PBS-treated group (data not shown). Notably, GFP chimeric mice treated by transplantation of BM-MSCs with siRNA CCL5 knockdown showed reduced GFP+/CD45dim microglia compared with the non-knockdown BM-MSC-treated group at 14 days after the last injection, although this did not reach statistical significance (data not shown). It is important to note that for these experiments we used BM-MSCs that transiently expressed CCL5 siRNA. Therefore, the knockdown effect of CCL5 siRNA might not be maintained for a long time (14 days) post-treatment. To investigate this point, we determined the content of CCL5 in the hippocampus at early time points (3 and 7 days) after the last BM-MSC or CCL5 knockdown BM-MSC treatment. The results revealed that the CCL5 content derived from the transplanted BM-MSCs showed a time-dependent decrease. Accordingly, the knockdown effect of CCL5 siRNA was more significant at 3 days after transplantation than at 7 and 14 days, although there remained an effect at 14 days as well (Supporting Information Fig. S6).
Thus, to further study the role of CCL5 in the AD mouse brain, we also examined early time points for microglia migration after BM-MSC treatment. At 3 days after BM-MSC injection, the GFP+/CD45dim microglia were significantly increased in the BM-MSC-treated chimeric mice compared with the PBS-treated group (Fig. 3E). GFP-positive and CD45-high macrophages (GFP+/CD45high) also were increased in the BM-MSC-transplanted mice compared to the PBS-infused mice. However, mice treated by BM-MSCs with CCL5 siRNA knockdown showed significantly reduced GFP+/CD45dim microglia and slightly decreased GFP+/CD45high macrophages (Fig. 3E).
We also analyzed the GFP chimeric mice receiving BM-MSC transplants 14 days after the last transplantation by histology (Fig. 3F). Microscopic investigation of the brains revealed that numerous endogenous cells (GFP positive) expressed Iba-1, a marker for microglia, confirming the differentiation of migrating BM-derived cells into microglia. Although this result does not rule out the possibility for self-renewal of microglia within the brain, collectively the data provides evidence that a large number of microglia derive from BM cells that migrate across the blood-brain barrier in response to CCL5 released by transplanted BM-MSCs in the AD microenvironment.
CCL5 Derived from BM-MSCs Modulates the Microglial Activation Status
At 3 days after intracerebral BM-MSC treatment, we found that the transplanted mice exhibited a fourfold decrease in TNF-α and a twofold decrease in IL-1β compared to PBS-treated APP/PS1 mice (Fig. 4A). However, treatment with CCL5 knockdown BM-MSCs did not lead to decreased expression of these cytokines in the hippocampus (Fig. 4A).
In addition, the hippocampal brains of BM-MSC-transplanted APP/PS1 mice had significantly increased levels of the alternative microglia markers, IL-4 and YM-1, and CCL5 knockdown significantly inhibited the induction of IL-4 expression (Fig. 4B) and slightly inhibited the induction of YM-1 expression compared with the BM-MSC-treated group, although this did not reach statistical significance (Fig. 4B).
To confirm these effects, we measured the TNF-α and IL-4 protein content in the hippocampi by ELISA. As shown in Figure 4C, TNF-α was lower and IL-4 higher in the BM-MSC-treated mice compared to the PBS-treated mice (Fig. 4C). When CCL5 was knocked down in the BM-MSCs by siRNA, these outcomes were changed (Fig. 4C). Immunofluorescent images in AD-GFP chimeric mice showed that the BM-derived microglia expressed IL-4 at 14 days after the last BM-MSC treatment (Fig. 4D).
To determine whether CCL5 released from BM-MSCs was directly affecting the resident microglia states, we compared CM derived from BM-MSCs or CCL5 knockdown BM-MSCs on Aβ-treated inflammatory BV2 microglia. TNF-α expression levels were increased, and IL-4 expression levels decreased in Aβ-exposed BV2 cells (control) compared with the non-Aβ-treated BV2 cells. However, both CM evoked beneficial effects on Aβ-exposed inflammatory BV2 cells, respectively (Supporting Information Fig. S7).
Microglia Recruited by BM-MSC-Derived CCL5 Can Reduce Aβ Deposition by Expression of Aβ-Degrading Enzymes
To analyze the effects of BM-MSC-derived CCL5 on the Aβ load in the brain, we first determined the Aβ profile using 6E10 immunostaining analysis in treated APP/PS1 mice. We found the deposition of Aβ to be markedly reduced following BM-MSC transplantation (Fig. 5A, 5B) in the hippocampus of APP/PS1 mice, consistent with other results recently reported by our group [ 24]. Although this effect was slightly less using CCL5 knockdown BM-MSCs, the difference did not reach statistical significance (Fig. 5A, 5B). We further confirmed the role of CCL5 derived from BM-MSCs in Aβ deposition by using Aβ 40 and 42 immunohistochemical analyses and ELISA assays in the APP/PS1 mice. The contents of both Aβ 40 and 42 were significantly reduced following BM-MSCs treatment (Fig. 5C, 5D), and these effects were partially negated by knockdown of the CCL5 gene prior to transplantation of the cells (Fig. 5C, 5D).
To determine the relationship between BM-derived cells and Aβ deposits after transplantation with BM-MSCs or BM-MSCs expressing CCL5 siRNA, BM-derived cells and Aβ deposits were immunostained and analyzed in AD-GFP chimeric mice. We clearly observed the colocalization of Aβ (6E10) and BM-derived cells (GFP positive). Recruitment of BM-derived cells to Aβ deposits (number of GFP positive cells per Aβ plaque) appeared to be enhanced in BM-MSC-treated mice compared with PBS-treated mice (Fig. 5E) but were significantly reduced in chimeric mice treated with BM-MSCs transduced with CCL5 siRNA.
We next analyzed the expression of Aβ degrading enzymes that are known to be released by microglia [ 10]. Notably, we observed significantly increased levels of neprilysin (NEP) and matrix metallopeptidase 9 (MMP9) in the hippocampal regions of AD mice treated with BM-MSCs compared to PBS-infused mice. However, the expression of these enzymes was significantly decreased in the AD mice treated with BM-MSCs transduced with CCL5 siRNA (Fig. 5F). The levels of insulin degrading enzyme (IDE) also showed similar change patterns as NEP and MMP9, but this did not reach statistical significance. To confirm these effects, we examined the levels of NEP in the hippocampus of AD mice by Western blot analysis. As shown in Figure 5G, NEP was significantly increased in the BM-MSC-treated mice compared to the PBS-treated mice (Fig. 5G). When CCL5 was knocked down in the BM-MSCs by siRNA, these outcomes were decreased (Fig. 5G). To know whether the increased expression of NEP after BM-MSC treatment was associated with the migration of BM-derived cells into the brain, we performed NEP immunostaining using brain sections of AD-chimeric mice. We found that the GFP-positive BM-derived cells expressed NEP at 14 days after the last BM-MSC transplantation (Fig. 5H).
Released CCL5 Following BM-MSC Transplantation Improves Behavioral Abnormalities
To address the role of BM-MSC-derived CCL5 in the cognitive function, we performed a MWM test on the treated mice. Notably, we found that APP/PS1 mice treated with BM-MSCs performed significantly better on the MWM test than PBS-treated counterparts (Fig. 6A). However, mice treated with BM-MSCs transduced with CCL5 siRNA did not show improved memory function (Fig. 6A). Figure 6B shows examples of the swimming traces in each mouse group analyzed by the MWM task on day 10. In the probe trial, APP/PS1 mice treated with CCL5 knockdown BM-MSCs showed a partial but significant decrease of crossing platform number compared to BM-MSC-treated mice (Fig. 6C). The time spent in the target quadrant did not differ among the groups (Fig. 6D).
Here, we describe cellular and molecular mechanisms underlying the BM-MSC-mediated beneficial effects. We show that bioactive, soluble factors from BM-MSCs significantly increased microglia migration in vitro and in mouse brains after transplantation. We further show that the migration of these cells in vitro were markedly elevated when the BM-MSCs were stimulated with Aβ, mimicking the microenviornment of the AD brain. We also suggest that the chemotactic activity of BM-MSCs is mediated by secretion of CCL5, and confirmation of this hypothesis was obtained by demonstrating that transfection of these cells with CCL5 siRNA led to a decreased effect on microglia migration.
Our in vivo results further showed that CCL5 expression is elevated after BM-MSC transplantation in APP/PS1 mice, but not normal mice. Moreover, BM-MSC-derived CCL5 released by Aβ stimulation was shown to play critical roles in the Aβ clearance and anti-inflammatory effects of BM-derived alternative microglia and led to cognitive improvement in APP/PS1 mice (Supporting Information Fig. S8). Overall, these results suggest that the beneficial effects of BM-MSCs in AD pathogenesis may be explained by BM-derived alternative microglia activation through CCL5 secretion of BM-MSCs. These alternative microglia are derived from migration of endogenous cells into the brain, rather than transdifferentiation of the transplanted MSC.
CCL5 is an 8-kDa protein classified as a chemotactic cytokine or chemokine. It is also known as RANTES (Regulated upon Activation, Normal T-cell Expressed, and Secreted). The chemotactic activity of CCL5 brings T-cells, dendritic cells, natural killer (NK) cells, granulocytes, macrophages, and monocyte to sites of inflammation, either alone or through triggering MCP-1 production [ 28–30]. CCL5 and its receptor CCR5 have been implicated in a wide array of pathological conditions in the brain and neurodegenerative diseases [ 31]. However, the exact role of CCL5 in the diseased CNS is unclear because its role in immune cell recruitment and activation has two sides. In this regard, increase of CCL5 by BM-MSCs in AD might exert a beneficial effect, at least, on AD pathogenesis.
It was therefore essential to examine whether CCL5 was increased in the AD brain microenvironment after BM-MSC treatment, and if so, whether it was a critical factor that mediated recruitment of microglia and amelioration of AD pathogenesis. In the mouse model of AD used here, BM-MSC transplantation led to a significantly increased expression of CCL5 in the brain. Concomitant with an elevated CCL5 level after BM-MSC treatment, the number of microglia was dramatically increased in the AD mouse brains (Fig. 3). These results suggest that within the degenerative microenvironment of Aβ deposition, the ability of BM-MSCs to release CCL5 is augmented and that the elevated CCL5 increased the number of microglia.
Previous studies have shown that BM-derived microglia were more effective in reducing AD pathogenesis than the resident microglia [ 32, 33]. Radiation BM-chimeric mice have been used extensively for studies of microglial turnover and replacement [ 2, 6]. A disadvantage of this model is that the recipient is usually subjected to whole-body irradiation before BM transplantation, which might induce microglial activation or cell death [ 34, 35], and could evoke aberrant microglia recruitment and transformation. To minimize these side effects of irradiation, such as abnormal microglial activation, blood-brain barrier destruction, and proinflammatory cytokine induction, we used partial head shielding methods [ 19, 26]. Although this method could not completely rule out irradiation-induced side effects, it minimized the likelihood of unusual immune cell activation. In addition, even though a BM-chimeric mouse is not a “normal” mouse, the microglial response to AD in BM-GFP chimeric and nonchimeric mice was comparable in our experiment (data not shown). In this study, we found that BM-MSC transplantation facilitated the migration of BM-derived cells to the brain and differentiation into microglia (Fig. 3). However, knockdown of CCL5 expression in BM-MSCs largely negated the microglia activation and recruitment of BM-derived cells to the brain induced by BM-MSC treatment (Fig. 3). Therefore, these results indicated that CCL5 derived from BM-MSCs after cell transplantation was a critical factor leading to increased BM-derived microglia in the brain.
It is widely accepted that resident microglia degenerate during AD progression and shift to an M1 activation state, with high levels of TNF-α and IL-1β [ 11, 36]. In specific conditions, BM-derived macrophage/microglia are recruited to compensate for the defective function of senescent resident microglia in the AD brain [ 37, 38]. Our previous and current studies showed that the expression of TNF-α and IL-1β was significantly increased, and the expression of IL-4 was markedly decreased in APP/PS1 mice [ 18, 24]. In addition, a remarkable decline of proinflammatory cytokines also was observed in the BM-MSC-treated mice [ 24], and markers of M2 macrophage/microglia activation, such as IL-4 and YM-1, were significantly upregulated (Fig. 4). The expression changes of these cytokines occurring after BM-MSC treatment in APP/PS1 mice also were abolished after knockdown of CCL5 expression in transplanted BM-MSCs (Fig. 4), suggesting that CCL5 was a chemokine mediating these effects. The beneficial effect of IL-4 in particular has been extensively probed in AD [ 39, 40]. For example, IL-4 promoted Aβ degradation in mononuclear phagocytes, attenuated the neuroinflammation, and restored the spatial learning function in AD mice [ 40, 41]. This neuroprotective cytokine, IL-4, tends to be released by BM-derived microglia in the AD brain [ 18]. Although this study does not establish a direct relationship between CCL5 derived from BM-MSCs and increased neuroprotective cytokines, the BM-derived alternative macrophage/microglia could contribute to the elevated expression of IL-4 (Fig. 4).
To determine whether CCL5 released from BM-MSCs was directly affecting the resident microglial states, we exposed CM derived from BM-MSCs or CCL5 knockdown BM-MSCs on Aβ-induced inflammatory microglia. Both CM evoked beneficial effects on Aβ-stimulated microglia (Supporting Information Fig. S7). Taken together, our results suggest that multiple factors expressed by BM-MSCs might assist in inducing resident microglia to the M2 microglia activation state, and that the main role of CCL5 derived from the BM-MSCs was recruiting BM-derived cells of the M2 type into the brain.
BM-derived cells can cross the blood-brain barrier, differentiate into microglia, and phagocytose Aβ [ 17, 20, 28]. In line with previous studies, we observed diminished Aβ deposits correlating with elevated accumulation of microglia [ 24] and BM-derived cells (Fig. 5) after BM-MSC transplantation. In BM-MSC-treated AD-GFP chimeric mice, colocalization of BM-derived cells and Aβ significantly increased. Also, secreting Aβ degrading enzymes by microglia were elevated in BM-MSC-transplanted AD mice compared with PBS-infused counterparts (Fig. 5F). These effects of BM-MSCs were significantly negated by knockdown of CCL5 in the BM-MSCs prior to injection. We also found that the increased NEP expression after BM-MSCs transplantation correlated with elevated BM-derived cells (Fig. 5G, 5H). Our results suggest that BM-derived cells (maybe microglia) that were recruited to the brain by CCL5 are secreting Aβ degrading enzymes, such as NEP, and these recruited cells mainly contribute to the Aβ clearance.
A recent study showed that deficiency of IL-4 resulted in severe cognitive impairment, and that the immune cells of these mice were skewed toward a proinflammatory phenotype [ 42]. Also, excessive Aβ accumulation was associated with disturbed cognitive function in an AD mouse model [ 43]. In our AD mouse model, decreased IL-4 expression, elevated proinflammatory cytokines, and increased Aβ accumulation were observed in the brain at 9 months of age [ 18]. In line with this, our APP/PS1 mice also showed impaired memory function compared with WT mice (Fig. 6). When APP/PS1 mice were treated with BM-MSCs, improved spatial memory was observed. However, mice infused with CCL5 knockdown BM-MSCs did not exhibit cognitive restoration. Although the exact role of CCL5 released from BM-MSCs on memory function remains to be determined, it may lead to several beneficial effects, such as increased IL-4 and Aβ degrading enzyme secretion and decreased proinflammatory cytokines release from recruited BM-derived macrophage/microglia.
In summary, the results shown here reveal that CCL5 is a novel factor affecting the beneficial outcomes of intracerebral BM-MSC transplantation in AD. Identification of CCL5 as a major mediator of the beneficial effects of BM-MSCs in AD enhances our understanding of how BM-MSCs work and supports the concept that the protective effects of BM-MSCs are largely mediated through production of paracrine factors. BM-MSCs respond to their microenvironment by secreting a variety of such factors. Elucidation of these BM-MSC releasing factors, including CCL5, will likely provide strategies to improve the therapeutic potential of BM-MSCs for AD treatment.
This work was supported by the Bio & Medical Technology Development Program (2010-0020234), Basic Science Research Program (2010-0003949), and World Class University Program (R32-10064) of the National Research Foundation (NRF) of Korea funded by the Ministry of Education, Science and Technology, Republic of Korea. This work was also supported by the Korea Healthcare technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (A09076109110000100).
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.