Distinct Effects of the Hippocampal Transplantation of Neural and Mesenchymal Stem Cells in a Transgenic Model of Alzheimer’s Disease

Alzheimer’s disease (AD) is a severe disabling condition with no cure currently available, which accounts for 60–70% of all dementia cases worldwide. Therefore, the investigation of possible therapeutic strategies for AD is necessary. To this end, animal models corresponding to the main aspects of AD in humans have been widely used. Similar to AD patients, the double transgenic APPswe/PS1dE9 (APP/PS1) mice show cognitive deficits, hyperlocomotion, amyloid-β (Αβ) plaques in the cortex and hippocampus, and exacerbated inflammatory responses. Recent studies have shown that these neuropathological features could be reversed by stem cell transplantation. However, the effects induced by neural (NSC) and mesenchymal (MSC) stem cells has never been compared in an AD animal model. Therefore, the present study aimed to investigate whether transplantation of NSC or MSC into the hippocampus of APP/PS1 mice reverses AD-induced pathological alterations, evaluated by the locomotor activity (open field test), short- and long-term memory (object recognition) tests, Αβ plaques (6-E10), microglia distribution (Iba-1), M1 (iNOS) and M2 (ARG-1) microglial phenotype frequencies. NSC and MSC engraftment reduced the number of Αβ plaques and produced an increase in M2 microglia polarization in the hippocampus of APP/PS1 mice, suggesting an anti-inflammatory effect of stem cell transplantation. NSC also reversed the hyperlocomotor activity and increased the number of microglia in the hippocampus of APP/PS1 mice. No impairment of short or long-term memory was observed in APP/PS1 mice. Overall, this study highlights the potential beneficial effects of transplanting NSC or MSC for AD treatment.


Introduction
Alzheimer's disease (AD) is the main form of dementia and accounts for more than 60% of cases [1]. Currently, AD has no cure and only palliative treatments are available, implying in a high economic burden due to the use of health care resources, treatment costs as well as indirect costs due to patients' inability to work [2]. AD is mainly characterized by gradual memory loss with cognitive impairment, personality and behavioral changes. Neuropathological features include the presence of amyloid-β (Aβ) plaques and neurofibrillary tangles, synaptic loss, neuronal death, and increased neuroinflammation found mainly in the cortex and hippocampus of AD patients and animal models [3].
Neuroinflammation is marked by microglial activation, which plays a key role in the pathogenesis of AD [4]. Activated microglia can assume two distinct phenotypes widely described in the literature as M1 (pro-inflammatory) and M2 (anti-inflammatory). On the one hand, M1 polarization leads to cytotoxicity due to high production and release of pro-inflammatory cytokines and increased expression of inducible nitric oxide synthase (iNOS) and consequent nitric oxide (NO) formation [5,6]. On the other hand, M2 polarization leads to the release of anti-inflammatory factors, neurotrophins, and enhanced expression of Arginase 1 (ARG-1), an enzyme responsible for tissue repair and reorganization of the extracellular matrix [6][7][8][9]. In view of that, iNOS is used as a marker of M1 polarization while ARG-1 is used as a marker of M2 microglial phenotype [9]. Both iNOS and ARG1 are intracellular proteins largely expressed in immune cells that compete for the substrate arginine [10] representing a stimulus-specific regulatory mechanism for limiting NO production. Accordingly, overexpression of ARG-1 leads to a reduction of the expression of iNOS [11], allowing the distinction of the M1/M2 states of microglia.
These neuropathological processes lead to behavioral disturbances that are associated with the memory deficits and cognitive decline observed during the AD progress [1]. Although several advances have been made in the understanding of AD, there is no effective treatment available to date. In this context, it is important to develop new therapeutic approaches to prevent or control the progression of neurodegeneration and neuroinflammation caused by Aβ peptide accumulation and deposition. Stem cell transplantation has shown potential therapeutical effects for several neurological diseases, including AD. Recent studies have shown that mesenchymal (MSC) and neural (NSC) stem cell transplantation in animal models of neurological diseases induced neuroprotective effects as well as significant functional recovery opening new perspectives for clinical applications [12][13][14][15][16][17].
The main effect of transplanted MSC is the immunomodulatory function through crosstalk with the immune cells or paracrine actions, which have been shown to reduce Aβ deposition in AD experimental models [10,11]. Protective mechanisms attributed to MSC include secretion of substances involved in anti-inflammatory, proliferative and antiapoptotic responses, as well as rapid migration and interaction with microglia, neurons, and astrocytes [18][19][20]. While MSC preferably have paracrine effects and trigger actions on resident cells, NSC are more likely to proliferate, selfrenew and differentiate into astrocytes, oligodendrocytes and neurons [21,22], as well as are capable of restoring working memory and short-term memory in AD [13,14]. Since these two cell populations act differently when transplanted into the brain, it is interesting to compare them in terms of their therapeutic effects in AD.
Aiming to resemble AD neuropathology, double mutant transgenic mice carrying genes associated with this neurodegenerative disease were generated (APPswe and PS1dE9) and are briefly referred as APP/PS1 [23]. The APPswe mutation favors APP processing by β-and γ-secretases leading to Αβ formation and aggregation [24]. PS1 is a core protein in the γ-secretase complex, and the deletion of exon 9 (dE9) in the PS1 gene has also been associated with increased Aβ deposition [3]. Histochemical studies in this in vivo model showed deposition of Aβ starting at 6 months of age (symptomatic middle stage) with a progressive increase of neuroinflammatory responses, plaques formation along with behavioral alterations [23,25]. The consistent deposition of Aβ in early APP/PS1 mice confirms its utility for behavioral and histological studies and makes it a valuable tool for the investigation of new therapeutic approaches [23]. Therefore, the aim of the present study was to compare the effects induced by MSC and NSC hippocampal transplantation on Aβ plaque formation, microglial cell frequency, and behavioral responses in a murine model (APP/PS1) of AD.

Animals and Maintenance
All animals were provided by CEDEME (Center for the Development of Animal Models in Biology and Medicine at the Universidade Federal de São Paulo). They were housed in polypropylene home cages (41 cm × 34 cm × 16.5 cm) in a pathogen-free facility, under controlled temperature (22-23ºC) and lighting (12 h light, 12 h dark; lights on at 6:45 a.m.) conditions. Appropriate food and water were available ad libitum. The Ethics Committee of the Federal University of São Paulo approved all experiments under the protocol #9268250618.

MSC Extraction and Cell Culture
MSC were obtained from male transgenic mice expressing GFP (n = 8), with 6 to 8 weeks old. Animals were euthanized by anesthetic overdose (thiopental 100 mg/kg, i.p.) after lidocaine administration (10 mg/kg, i.p.). Both femurs and tibias were removed, then muscles and cartilage were dissected. In a vertical laminar flux, the largest epiphyses of both bones were cut, MSC were removed by flushing with Dulbecco's Modified Eagle Medium (DMEM) and centrifuged (1700 g for 5 min). MSC were extracted from the supernatant and cultivated in DMEM Low Glucose and 10% fetal bovine serum. Thereafter, cells were expanded and used between 7th -10th passages, when had reached 70-90% of confluence.

NSC Extraction and Cell Culture
Time-pregnant rats with 13.5 days were euthanized with an overdose of the anesthetic (thiopental 150 mg / kg) after application of lidocaine 10 mg / kg (i.p.). Fetuses were removed from the uterus and selected when expressing GFP with the help of a fluorescence microscope (Nikon) at FITC b-2e/c fluorescence excitation. The telencephalon region was extracted in a stereomicroscope (Nikon). Tissues were incubated in Eppendorf tubes with 0.05% sterile trypsin solution for 5 min at 37 °C and inactivated with FBS (Gibco). The tubes were centrifuged at 1500 g for 7 min and the tissue was mechanically dissociated in DMEM-F12 with DNAse. Following another centrifugation step, the pellet was resuspended in 0.5 mL of DMEM. Cells were grown at a density of approximately 100,000 cells/mL in 25 cm 2 plastic bottles previously treated with poly 2-hydroxyethyl methacrylate (Sigma), a polymer capable of preventing cell adhesion. The standard culture medium used for neurosphere growth was composed of DMEM-F12 50:50 (Invitrogen), supplemented with 1% N-2 (Invitrogen), 1% L-glutamine (200 mM; Invitrogen), penicillin (100 IU / mL), streptomycin (100 ug / mL) and amphotericin B (0.25 ug / mL), Epidermal Growth Factor (EGF) (20 ng / mL; Sigma) and Fibroblast Growth Factor 2(FGF-2) (10 ng / mL; P&D). The cells were kept in a cell incubator at 37ºC and 5% CO 2 for approximately 5 days [21].

Stem Cell Transplantation
Under deep anesthesia (ketamine 115 mg/kg + xylazine 10 mg/kg) APP/PS1 mice were placed in a stereotaxic frame. The skull surface was exposed after scalp incision and cleansed to expose the bregma. Two orifices were opened, and gingival needles were used for stem cell infusion according to the coordinates: AP = -2,06 mm, ML = ± 1,3 mm, DV = -2,7 mm. 400.000 NSC or MSC previously dissociated were diluted in sterile saline and 4μL of this solution was slowly infused in each animal. Then, skull orifices and scalps were closed with bone wax and glue, respectively.
After the procedure, all animals received Ibuprofen (30 mg/kg, orally administered) right after surgery and in 12 h intervals for 3 days and were monitored. Ibuprofen was used to manage pain after surgery as recommended by the Ethical Committee for animal Research (CEUA-UNIFESP). Mice were distributed in experimental groups according to their genotypes (positive or negative for transgenicity) and stem cell transplantation as follows:

Behavioral Tests
The animals were exposed to the Open Field (OF) and Novel Object Recognition (NOR) behavioral tests four weeks after stem cell transplantation. All tests were video-recorded and were examined by two experimenters blind to the experimental groups.

Open Field (OF)
The Open Field (OF) test was performed in a cylindric apparatus of 40 cm-diameter, 50 cm-high and 2 cm-thick walls, as previously described [26]. The arena floor is divided into 19 sections of approximately equal dimensions: 12 peripheric sections (adjacent to the wall), 6 intermediate sections and 1 central section, in three concentric circles of different radiuses (20, 14 and 8 cm, respectively). Animals were placed in the arena for 10 min and allowed to explore it freely. All tests were video-recorded. Locomotion frequency was defined as the number of sections explored by the animal during the test; each locomotion unit corresponds to the animal´s four paws placed inside one of the sectors.

Novel Object Recognition (NOR)
The Novel Object Recognition (NOR) test was carried out in the same arena used for the OF test, 24 h after it. The test consisted of two sessions: training and test. In both sessions, each animal was allowed to explore freely for 5 min. The training session consisted of one trial, where two equal objects (A and A') were fixed on opposite sides of the arena with adhesive tape. The test session consisted of one trial, 1 h after training, where one object was replaced by a new one (B) [27]. Object exploration was defined as smelling and touching objects with snout or front paws. Sitting or walking around the object were not considered object exploration.
The discrimination index (DI) was used to analyze exploration time (T) between new (NT) and familiar (FT) objects. DI is calculated as the difference between familiar object exploration (FT) and new object exploration (NT) divided by the total time spent exploring both objects [DI = (NT-FT) / (NT + FT)]. Positive DI values indicate that the new object was more explored than the familiar one, while negative values indicate the opposite. Zero indicates lack of preference.
Briefly, sections were washed and incubated with the primary antibody overnight. On the following day, sections were incubated for 2 h with a fluorochrome-conjugated appropriate secondary antibody anti-mouse Alexa Fluor 488, anti-rabbit Alexa Fluor 488 anti-rabbit Alexa Fluor 568 (Invitrogen) and washed in PBS. Slides were mounted and sealed with DPX and subsequently analyzed by fluorescence microscopy using the Neurolucida capture software (MBF Bioscience). For M1/M2 phenotype quantification, images were acquired with the TissueFaxs Confocal cytometry (Tis-sueGnostics, Vienna, Austria) and analyzed with the Strata-Quest software (TissueGnostics).

Tissue Cytometry
Hippocampal images were acquired with the TissueFAXS Cytometry (TissueGnostics GmbH, Vienna, Austria) system, coupled to a spinning disk confocal microscope (Zeiss). Data were analyzed with the StrataQuest software (Tissue-Gnostics GmbH) as described elsewhere [28]. iNOS/Iba-1, ARG-1/Iba-1, 6-E10 and Iba-1 immunostaining was examined for the following parameters: number of cells and total area. Six hippocampal sections of 40 µm were captured, resulting in approximately 90 fields of view/section.

Statistical Analysis
Statistical analysis was conducted using GraphPad Prism 8.0 software. OF and NOR tests were analyzed through Kruskal-Wallis test, followed by Dunn's post-test. Immunohistochemical and immunofluorescence quantitative analyses were performed using one-way ANOVA followed by Dunnett's post-test. Significance was assumed at p < 0.05.

MSC and NSC Reduced the Number of Aβ Plaques in the Hippocampus, But Not the Size of the Plaques
After six weeks of NSC and MSC transplantation in the hippocampus of APP/PS1 mice, the number and size (area) of 6E10-stained Αβ plaques were analyzed. We found a significant decrease in the number of Αβ plaques in the APP/PS1 + NSC (2.980 ± 0.27, mean ± SEM) and APP/ PS1 + MSC groups (2.605 ± 0.63) compared to the APP/ PS1 group (6.327 ± 0.67) (Fig. 1A: F 3,20 = 28.99; respectively p = 0.0003 and p < 0.0001). As expected, the APP/ PS1 groups showed a significant increase in the number and size (area) of Aβ plaques in the hippocampus compared to the WT group (Fig. 1B: F 3

APP/PS1 + NSC Animals Showed an Increased Number of Microglial Cells and a Larger Total Area Occupied by Microglial Clusters in the Hippocampus
The number of Iba-1 immunostained cells was increased in APP/PS1 + NSC mice (2628 ± 147) ( Fig. 2A: F 3

NSC Transplantation Reverses the Exacerbated Locomotor Activity of APP/PS1 Mice in the Open Field Test
For locomotor activity analysis, mice were tested in theOF. The total number of entrances into any floor unit revealed that APP/PS1 mice showed a significant increase in locomotor activity compared to WT animals ( Fig. 3A: H = 15.82; p = 0.0012). Peripheral locomotion also increased in APP/ PS1 mice, and NSC grafting was able to restore this behavior to control levels (Figs. 3B: H = 14.36; p = 0.0025).
When mice were tested for the Novel Object Recognition (NOR) task, no statistically significant difference was found in DI between objects regarding the interaction time

Discussion
The results of the present study indicated that transplantation of NSC and MSC reduced the number of Aβ plaques and increased microglia polarization towards the M2 antiinflammatory phenotype in the hippocampus of APP/PS1 mice, although cell transplantation was not sufficient to reduce the size of plaques. In addition, NSC led to an increase in microglial cells in the hippocampus compared to the APP/PS1 group. Also, NSC were able to reverse the hyperlocomotion present in APP/PS1 animals, although neither NSC nor MSC transplantation induced effect on the object recognition memory.
Although both cell types, NSC and MSC, exhibited different properties, they were able to slow down the buildup of new plaques. Considering that MSC were no longer found in the hippocampus six weeks after surgery (data not shown), it is plausible that the cell infusion produced an were quantified using TissueFAXS (iNOS green and Iba-1 orange) in the hippocampus of APP/PS1 (K), APP/PS1 + MSC (L) and APP/ PS1 + NSC mice (M), respectively. Bars represent mean ± SEM; n = 5-6 animals/group. One-way ANOVA followed by Dunnett's posttest; *p < 0.05. Scale bar = 50 µm immunomodulatory effect through a paracrine action that reduced neuroinflammation, which in turn could inhibit the formation of new Aβ plaques [30]. In addition, it has been proposed that MSC act as a CNS immunomodulator, maintaining and repairing brain tissue, not only structurally but also functionally [31]. However, the reduction in the number of plaques caused by NSC transplantation has other underlying mechanisms. NSC transplantation in AD helps to improve cognitive behavior, releasing neurotrophic factors that counteract the the progression of neuroinflammation [32].
Neuroinflammation was examined by quantifying microglial polarization M1/M2 as double-stained of iNOS/Iba-1 and ARG-1/Iba-1. In APP/PS1 mice microglial cells become less responsive to M2 induction signals as they aged and pathology worsened [33]. Our data suggest that APP/PS1 mice with seven months old already present an imbalance in favor of the M1 phenotype of microglia indicating that Aβ induced inflammatory environment and changes microglia phenotypes. The increase in the M1 phenotype of microglia in APP/PS1 mice has also been observed in previous studies, with or without injury induction or treatment. For example, Zhang et al. (2017) observed that acute hypoxia increased M1 and reduced M2 markers in the hippocampus of APP/ PS1 mice [34]. In addition, Wan et al., 2016 observed an increase in the M1 phenotype by iNOS expression and reduction of M2 by ARG-1 expression in APP/PS1 mice, which was prevented by treatment with Ginkgo biloba [35].
Moreover, it is well known that activated microglia with M2 profile are involved in the clearance of Aβ plaques in Alzheimer's disease [36,37] Indeed, together with the reduction in Aβ plaques, the APP/PS1 + NSC group showed a significant increase in microglia frequency in the hippocampus. More interesting, transition from M1 to M2 polarization occurred in the microglia population after NSC and MSC transplantation in APP/PS1 mice. These high concentrations of microglia were found preferentially in the vicinity of Aβ plaques after NSC transplantation, which is in accordance with the suggestion that microglial cells are activated by the presence of Aβ proteins clustering around the plaques to secrete cytokines and neurotoxins and to phagocyte them [38].
In accordance with our findings, other studies also showed that MSC [39] or NSC [40,41] were able to change from M1 to the M2 state. For example, MSC can induce the M2 microglial phenotype and mitigate damage after traumatic brain injury (TBI) [39]. A recent study that investigated the polarization of macrophages in an in vitro coculture system showed that NSC induced M2 polarization and suppressed M1 polarization of macrophages. Gao and co-workers, focusing on neuropathological changes in a mouse TBI model, showed that the transition of microglia/ Fig. 3 Locomotor activity analysis of APP/PS1 and WT mice subjected to the Open Field test. Total (A) and peripheral (B) locomotion was compared between the groups. Bars represent mean ± SEM; n = 13-14 animals/group. Kruskal-Wallis test, followed by Dunn's post-test; *p < 0,05. Discrimination index (DI) between APP/PS1 and WT mice subjected to NOR testis indicated in (C) and (D). Animals were subjected to the NOR test for 5 min and the interaction time (C), and the number of contacts with each object (D) were calculated using the Discrimination Index ( ID = (TN−TF) (TN+TF) ). Bars represent mean ± SEM; n = 13-14 animals/group. Kruskal-Wallis test, followed by Dunn's post-test; *p < 0,05 macrophage cells into the M2 phenotype was supported by NSC [41]. Therefore, our findings are in accordance with other studies where the transplantation of the two cell types, MSC and NSC, favored the M2 polarization of microglia. These findings highlight the relevance of early intervention and reinforce the protective effects of stem cell transplantation and its potential anti-inflammatory effect on AD. Thus, considering that M1 and M2 polarization of microglia plays a significant role in changing from pro-to anti-inflammatory responses, therapeutic strategies that recover the imbalance between microglial polarization states M1 and M2 are promising targets for the treatment of AD.
Since the transplanted cells did not differentiate into microglia, it is reasonable to speculate that the number of microglia increased in NSC-treated mice as a consequence of enhancement of Aβ hippocampal levels, thereby positively affecting the number but not the size of the plaques in the hippocampus. M2 polarization was also increased in MSC transplanted group and, considering that APP/PS1 animals have Aβ plaques as a permanent condition, it is plausible that MSC transplantation induces a continuous expression of M2 phenotype. This in turn results in a lower number of Aβ plaques in this group. Moreover, studies have shown that MSC can reduce the deposition of Aβ peptides, inhibiting the expression of APP-cleaving protein in the β site (BACE1) through immunomodulation and cytokine secretion [31,42].
APP/PS1 mice with seven-months old showed hyperlocomotion in the OF test when compared to WT animals, and only NSC transplantation could reverse this behavior. APP/ PS1 mice at this same age showed an increase in peripheric locomotion in the same test. Clinical observations already describe agitation and an increase in locomotor activity in AD patients [43]. Studies also describe the same behavior in 7 to 8-months-old APP/PS1 mice in both OF and bright-dark tests.
NSC were able to differentiate into neurons in the hippocampus. Literature data show that newly formed neurons can increase connectivity and synaptic density [44][45][46][47]. Although MSC are also capable to release neurotrophic factors and microglial modulators increasing the clearance of Aβ plaques in the hippocampus [48,49], their transdifferentiation may not generate cells functional neurons. These data may explain the ability of NSC to reverse hyperlocomotion in the OF test [38].
APP/PS1 mice show no impairment in recognition memory as revealed by the NOR test, which is consistent with previous reports [50,51]. Previous studies have shown that 9-months-old APP/PS1 animals did not exhibit neuronal loss in the hippocampus, as the amount of Aβ plaques could not yet cause neuronal death [52,53]. Therefore, our study was not able to evaluate whether stem cell transplantation could protect APP/PS1 animals from neuronal loss in later years, as demonstrated by other studies [54], [55] and [56]. Future research needs to examine the same parameters in older animals to determine whether NSC and MSC transplantation could protect recognition memory. The reduction in hyperlocomotion observed in APP/PS1 + NSC animals plays an important role in bringing APP/PS1 mice closer in behavior to their WT counterparts. The fact that MSC were not found in the hippocampus 6 weeks after transplantation may account for the absence of significant differences in the hyperlocomotion between APP/PS1 and APP/PS1 + MSC animals.

Conclusion
Therefore, although Aβ deposition can be managed by MSC, these cells tend to be less effective in repair Aβ-induced damage in APP/PS1 animals than NSC. MSC and NSC transplantation can reduce Aβ plaques and signaling microglia to polarize in M2 phenotype, promoting an anti-inflammatory environment in seven-months old APP/PS1 mice. Despite that, we observed distinct effects on behavior, since only NSC transplantation was efficient in a reversal of peripheral hyperlocomotion activity and increase the frequency of activated microglia in APP/PS1 mice. However, although we observed differences in neuroprotective effect, both NSC and MSC transplantation deserve further investigation regarding the modulatory potential of neuroinflammation in Alzheimer's disease.
As depicted in Fig. 4, APP/PS1 transgenic mice are widely used to study both neurobiology and novel transgenic mice have been widely used to study both the neurobiology and new therapeutic approaches for this AD subtype. AD hallmarks include agitation and increase formation of Aβ plaques in the hippocampus. Accordingly, APP/PS1 transgenic mice showed hyperlocomotion in the OF test along with increased 6E10 immunostaining in the hippocampus (Aβ plaque marker), which was reversed by NSC or MSC transplantation into the hippocampus. These results reinforce the potential beneficial effects of stem cell transplantation for the treatment of AD therapeutic approaches for AD treatment. Hallmarks in AD patients include agitation and increased hippocampal Aβ plaque formation. Accordingly, APP/PS1 transgenic mice showed hyperlocomotion in the OF test along with enhanced hippocampal 6-E10 immunostaining (Aβ plaque marker), which was reversed by NSC or MSC transplantation into the hippocampus. These results reinforce the potential beneficial effects of stem cell transplantation for the treatment of AD.