Neural Stem Cell Transplantation Improves Spatial Learning and Memory via Neuronal Regeneration in Amyloid-β Precursor Protein/presenilin 1/tau Triple Transgenic Mice

Abstract

Neural stem cell (NSC) transplantation has recently become a main research target for Alzheimer’s disease (AD) treatment. In the present study, we transplanted NSCs from C57BL/6 mice into the hippocampus in the 12-month-old triple transgenic model of AD (3 × Tg) and determined whether NSC transplantation can alleviate impairments in spatial learning and memory via neuronal regeneration in AD mice. Two months after transplantation, Morris water maze tests suggested that spatial learning and memory in the 3 × Tg mice receiving NSCs was significantly improved compared to 3 × Tg mice not receiving NSCs. Furthermore, quantification of Nissl staining revealed that the number of neurons in the hippocampus of 3 × Tg mice receiving NSCs was significantly greater than that in 3 × Tg mice not receiving NSCs, indicating that new neurons were generated. These results may demonstrate that NSC transplantation can improve spatial learning and memory via neuronal regeneration in amyloid-β precursor protein/presenilin 1/tau 3 × Tg mice.

Introduction

Alzheimer’s disease (AD) is the most common form of dementia in the elderly patients, with characteristic neuronal deficits in the cerebral cortical and hippocampal areas associated with cognitive decline. ^{1 –3} Drug treatment is always the main strategy for AD, especially in its early stages. ⁴ Unfortunately, this conventional method can alleviate only some clinical symptoms but not reverse the pathological course by replacing lost neurons. To date, there is no effective treatment against AD that is capable of reversing neuronal loss.

A stem cell is a potential donor cell source that may be used for cell replacement. Cell-replacement strategy has been recently developed with ever-growing advances in stem cell transplantation. ⁵ Neural stem cells (NSCs) are capable of self-renewal and can differentiate into various brain cell types such as neurons, astrocytes, and oligodendrocytes. ^6,7 This ability, combined with other biological features of NSCs, makes them a potential target for replacement of impaired neurons in the brain. Indeed, previous studies in several animal models of AD have shown that NSC transplantation can be used to prevent the loss of learning and memory function in AD. ^3,8,9 However, among other shortcomings, these mouse models cannot perfectly mimic the behavioral impairment and neuropathology observed in patients with AD. A triple transgenic model of AD (3 × Tg-AD) has been generated to express both senile plaques (SP) and neurofibrillary tangles (NFT) that are characteristic of patients with AD and has become the mainstream AD animal model. ^10,11 In the present study, we transplanted NSCs from C57BL/6 mice into the hippocampus of 3 × Tg mice expressing mutated amyloid-β precursor protein (APP), presenilin 1 (PS1), and tau (APP/PS1/tau 3 × Tg). Additionally, we tested spatial learning and memory using Morris water maze (MWM) and quantified neuronal number after NSC transplantation via Nissl staining. The aim of this research was to determine whether NSCs can alleviate, or even reverse, spatial learning and memory impairment via neuronal regeneration in APP/PS1/tau 3 × Tg mice.

Materials and Methods

Experimental Animals

Male homozygous APP/PS1/tau 3 × Tg mice expressing 3 mutant genes, ¹⁰ APP (APPswe), PS1 (PS1M146V), and tau (P301L) were purchased from the animal model center of Nanjing University (Nanjing, China). The derived descendants were genotyped and raised in separate cages in a12-hour light–dark cycle at constant temperature with free access to food and water. In the present study, we used 12-month-old APP/PS1/tau 3 × Tg mice (n = 20) and age-matched wild-type (wt) littermate males (n = 10). The study was approved by the Nanjin Medical University Ethics Committee, and all experiments were performed in accordance with guidelines from the Chinese Animal Welfare Agency.

Neural Stem Cell Proliferation

Original generation NSCs were derived from the hippocampus and subependymal zone of the fetal brain of GFP-expressing C57BL/6 mice at age E12.5 (Cyagen Biosciences; No. MUBNF-01001). The NSCs were thawed, transferred to centrifuge tubes containing the medium, and centrifuged at 1500 rpm for 5 minutes. After removing the supernatant, cells were suspended with 2 mL medium, dispersed gently, and mixed well. The cell suspension was cultured in the proliferating media (37°C, 5% CO₂, and 80% relative humidity) for 5 days. The NSCs were aggregated into neurospheres then subsequently trypsinized and triturated to a single-cell suspension. Passaged culture was performed at a 3-day interval, depending on the growth rate of the cells and color change of the medium. By passage 3, Nestin immunofluorescence staining was performed for identification.

Neural Stem Cell Transplantation

At 12 months of age, twenty 3 × Tg mice were randomly assigned into 2 groups (n = 10 per group), the first group receiving NSCs (NSC group) and the second group receiving phosphate-buffered saline (PBS group). An additional 10 wt mice of the same age not receiving any treatment were used as a blank control (WtC group).

The 3 × Tg mice were anesthetized with 1% pentobarbital (40 mg/kg) injected intraperitoneally (ip) and received a bilateral hippocampal CA1 graft of either NSCs or PBS. Briefly, mice were fixed on a stereotaxic frame (Blue Star wi290044, Xin-min biological technology co. LTD, Shanghai, China), and the skin was shaved and sterilized with iodine. A spindle-shaped median incision was made on the scalp with the periosteum dissected. The plate was burned with 1% hydrogen peroxide to expose the sagittal suture and coronal suture. Bregma was used as a reference point with the following stereotaxic coordinates, medial/lateral ±1.8 mm, anterior/posterior +2.0 mm, and dorsal/ventral −5.0 mm. A volume of 2 µL NSC solution (cell number: 1 × 10⁶/µL) was injected gently into the brain within 10 minutes by fixing a microsyringe on the stereotaxic apparatus and inserting the needle 2.5 mm into the dura. The microsyringe was left in place for 5 minutes after injection withdrawn gently. The holes were sealed after completion of NSC transplantation and repeatedly washed with normal saline. After a small amount of penicillin powder was spread over the incision to prevent infection, the scalp was finally sewn up.

Morris Water Maze

Before transplantation and 8 weeks after transplantation, all groups (n = 10 per group) were tested in the MWM (Shanghai Mobile Datum Information Technology Co, Ltd, Shanghai, China) to assess spatial learning and memory. Trial were conducted in a large pool (180 cm in diameter, 55 cm in height, and 42 cm in water depth) filled with water (22°C ± 2°C). The water pool was divided into 4 equal quadrants. A transparent platform (8 cm in diameter and 40 cm in height) was place in the center of the third quadrant (35 cm from the pool wall and 2 cm below the water surface). A video system was installed above the maze. The procedure consisted of 6 days of hidden platform tests and a probe trial 24 hours after the last hidden platform test.

The hidden platform test was used to the spatial message acquisition capabilities of the mice in the water maze. During the 6-day experiment, each day was divided into 2 time segments, morning and afternoon, with 2 trials per segment during which any of the 4 quadrants was selected randomly as the plunge point. The mouse was gently plunged into the water pool facing the wall. If the mouse climbed up the platform after a period of swimming in water and stayed on the platform for more than 5 seconds, it was considered a success. The software system automatically recorded the duration of swimming, which was regarded as the escape latency of looking for the platform. If the animal failed to find the platform within 60 seconds, it was led onto the platform, allowed to stand on it for 10 seconds, and given an escape latency of 60 seconds. The minimum interval between 2 exercises was 15 minutes. The escape latency and path length that the animal used to find the platform were observed and recorded for each trial. Twenty-four hours after the last hidden platform trial, all mice performed a probe trial to evaluate their retention of the spatial learning and memory for the platform.

The spatial probe trial was used to test memory retention. Twenty-four hours after the completion of acquisition trials, the platform was removed from the maze, and the animal was placed into water from any of the 4 quadrants facing the wall and allowed to swim freely for 60 seconds. The time spent in the target quadrant and the number of platform crossings were measured, while searching strategy was analyzed.

Nissl Staining

After MWM examination, mice (2-3 animals/group) were anesthetized with 1% pentobarbital (60 mg/kg) ip and laid supine to incise the chest and expose the heart and aorta. A catheter was inserted from the left ventricle to the aorta, through which 100 mL of 4°C normal saline was infused to wash blood quickly, followed by perfusion with 100 mL of 4% paraformaldehyde. The posterior brain was removed and fixed in 4% paraformaldehyde. After a 30-minute wash with PBS, the brain tissue was immersed in PBS containing sucrose solution overnight, fully frozen at −20°C, and serially sectioned (4μm coronal sections) with the injection site as the center using a freezing microtome. Frozen sections were stained with 1% toluidine blue at 56°C for 30 minutes, washed with water, color separated with 95% ethanol, dehydrated, hyalinized, and mounted for microscopic examination. To analyze the levels of neurons, quantification was performed using a 400× objective and a 1 mm × 1 mm grid on 15 fields from 8 to 10 consecutive stained sections that contained the hippocampal formation (Nikon E 800 microscope image analysis system).

Statistical Analysis

Statistical descriptions were expressed as the mean ± SD (x ± s). Data from all groups were analyzed with SPSS for Windows (version 17.0; SPSS Inc, Chicago, USA) software. The water maze acquisition data (escape latency and path length) were analyzed by 2-way repeated-measures analysis of variance (ANOVA), and post hoc comparisons were made using Bonferroni corrections. One-way ANOVA was used to analyze the possible differences in the probe trials of the MWM. Comparisons of histology indexes among 3 group mice were also conducted by a 1-way ANOVA. A P value less than .05 was considered statistically significant.

Results

Neural Stem Cell Proliferation and Nestin Staining

Five days after the inoculation, NSCs from C57BL/6 mice remained undifferentiated and were amassed in aggregates as neurospheres (Figure 1A). Nestin immunofluorescence of neurospheres was strongly positive (Figure 1B), indicating that neurospheres were composed of undifferentiated NSCs.

Figure 1.

A, Cultured neural stem cells (NSCs) were amassed in aggregates as neurospheres (×40 magnification). B, Nestin staining of the neurosphere show strong positive labeling (×40 magnification).

Morris Water Maze

Figure 2 shows the experimental results for the different groups in the hidden platform test. Before transplantation (Figure 2A and B), the escape latency gradually decreased over days for mice in the NSC, PBS, and WtC groups (F _5,165 = 114.057, P < .001), indicating that mice were able to learn the task and significantly improve their performance during the hidden platform training. However, there was a significant overall effect of genotype on escape latency (F _2,33 = 20.735, P < .001), and post hoc analysis showed that 3 × Tg mice in the NSCs and PBS group needed more time to find the platform compared to Wtc mice (P < .05). No significant differences were present between the NSC and PBS groups (P = .085). Similarly, swimming distance to locate the platform during training also decreased daily (F _5,165 = 61.578, P < .001), and significant effects of genotype were observed in the path length measured (F _2,33 = 12.492, P < .001). All results indicated 12-month-old Tg mice had a spatial learning and memory deficit. After transplantation (Figure 2C and D), the escape latency and path length in the hidden platform test were also shortened gradually in all mice after 6 days of training (F _5,165 = 132.009 and 71.037, respectively, P < .001), indicating that learning occurred in all of the tested groups. For the escape latency, an effect of genotype was shown (F _2,33 = 23.786, P < .001), and post hoc analysis showed the escape latency in the 3 × Tg mice receiving NSCs was significantly lower than in the 3 × Tg mice not receiving NSCs (P < .05). However, there was no significant difference in escape latency between the 3 × Tg mice receiving NSCs and WtC mice (P = .063), indicating that 3 × Tg mice in the NSC group exhibited a significant improvement in learning ability compared to mice in the PBS group. Furthermore, swimming distance in the NSC group also showed a significant decrease compared to the PBS group (P < .05), but no significant difference compared to the WtC group (P = .052), indicating that the 3 × Tg mice receiving NSCs were able to swim significantly shorter distances to find the platform than 3 × Tg mice without NSCs.

Figure 2.

Results of the hidden platform. A and B, The escape latency and path length for each group before transplantation showed a significant difference between the NSCs and PBS groups compared to the WtC group (P < .05). C and D, The escape latency and path length after transplantation showed a significant difference between the NSC group and PBS group (P < .05) but not between the NSC group and the WtC group (P >.05). The results are mean ± SEM. NSC indicates neural stem cell; PBS, phosphate-buffered saline; SEM, standard error of the mean.

The experimental results for different groups obtained from the probe trial are shown in Figure 3A and B. Before transplantation, there was a significant effect of genotype on time spent in the target quadrant and the number of platform crossings (F _2,33 = 20.839 and 11.34, respectively, P < .001). Post hoc analysis showed that 3 × Tg mice of the NSC and PBS group tended to spend less time in the platform quadrant (third quadrant) than WtC mice (P < .05). The number of platform crossings in the NSC and PBS groups was also significantly decreased compared to WtC mice (P < .05), indicating impaired memory retention of the Tg mice. After transplantation, the 3 × Tg mice receiving NSCs showed an exclusive preference for the target quadrant, and spent more time in the third quadrant compared to 3 × Tg mice not receiving NSCs (P < .05). However, post hoc comparisons indicated that there was no significant difference between the NSC and WtC groups for time spent in the target platform (P = .089). For the number of platform crossings, similar results were also shown (P = .112).

Figure 3.

Results of the probe trial (A and B). Before transplantation, the NSC and PBS group tended to spend less time in the target quadrant than WtC group (P < .05); the number of platform crossings was also significantly decreased compared to WtC group (P < .05). After transplantation, the time spent in the target quadrant and the number of crossings of platform in the NSC group were significantly increased compared with the PBS group (P < .05), and there was no significant difference between the NSC and WtC groups (P >.05). The results are mean ± SEM. NSC indicates neural stem cell; PBS, phosphate-buffered saline; SEM, standard error of the mean.

In addition, 3 main strategies (tendency strategy, marginal strategy, and random strategy) for locating the target platform were found in the probe trial. Before transplantation, 3 × Tg mice of the NSC and PBS groups often displayed a random (Figure 4A) or marginal (Figure 4B) strategy type. However, the 3 × Tg mice receiving NSCs primarily showed a tendency strategy compared to 3 × Tg mice not receiving NSCs after transplantation (Figure 4C). These data also indicated that memory retention of 3 × Tg mice improved significantly after transplantation.

Figure 4.

Three representative searching patterns during the probe trial. A, Random strategy, (B) marginal strategy, and (C) tendency strategy.

Nissl Staining

Nissl staining showed that neurons were observed in good and dense arrangements in the hippocampus of WtC mice, with rich Nissl bodies in the cytoplasm at higher magnification (Figure 5A). For the 3 × Tg mice not receiving NSCs, Nissl bodies in the cytoplasm decreased or disappeared, which marked neuronal damage and reduction were observed (Figure 5B). In contrast, in the NSC group, the hippocampus pyramidal cell layer of the 3 × Tg mice was densely arranged (Figure 5C), and the number of neurons was significantly greater than that in the PBS group (P < .05; Figure 5D). However, there was no significant difference in the number of neurons between the NSC and WtC groups (P >.05), indicating that the impaired neurons had been repaired or new neurons had been generated.

Figure 5.

Representative image of Nissl staining. A, Neurons were observed in dense arrangements in the hippocampus of the WtC group, with the rich Nissl bodies in cytoplasm at higher magnification; (B) the pyramidal cell layer in the hippocampus of the PBS group was broken, and neurons decreased or absent. C, The hippocampus pyramidal cell layer in the NSCs group was densely arranged compared with the PBS group, and Nissl bodies were rich in cytoplasm at higher magnification; (D) quantitative analysis of Nissl staining shows the number of neurons in the NSC group was significantly greater than that in the PBS group (P < .05). Scale bar = 50 μm. NSC indicates neural stem cell; PBS, buffered saline.

Discussion

As one of the most common neurodegenerative disorders, AD displays a number of pathological characteristics. The massive neuronal loss is one of the most obvious pathological changes, ^12,13 and cognitive impairment is a classical clinic symptom. ^14,15 So far, drug therapy that can delay cognitive deficits to some extent has still been the main treatment method for AD, such as acetylcholinesterase inhibitors, NMDA antagonists, and antioxidants. ^{16 –18} However, these drugs do not compensate for lost neurons in the hippocampus and cortex, so the curative effect is still limited.

Since Anderson discovered NSCs in 1989, ¹⁹ there has been a new understanding for cerebral neuron regeneration and the treatment of neurodegenerative diseases. The NSCs, which can be derived from embryonic central nervous tissue or adult mammalian cortex, hippocampus, lateral ventricles, and spinal cord, are self-renewing cells that have the remarkable capability to replace degenerating neurons. ²⁰ In vitro experiments have shown that NSCs can differentiate into the major cell types of the brain, including neurons, astrocytes, and oligodendrocytes. ^6,7,21 Due to their high-survival rate and ability to differentiate into neural function cells following transplantation into damaged tissue, NSCs have been proposed as alternative sources of cells for transplantation into the brain in neurodegenerative disorders. Previous studies demonstrate benefits of NSCs as cell source for transplant therapy for AD transgenic mice. Also, these cells are safe and effective in an animal model of AD. In the present study, NSCs were obtained from the hippocampus and subependymal zone of the fetal brain of C57BL/6 mice, and previous research shows that the original generation of NSCs has good breeding potential and multidirectional differentiation capacity. ²² By microscopy, we found that NSCs derived from C57BL/6 mice continually divided and gradually gathered together to form neurospheres. Nestin immunofluorescence also demonstrated that neurospheres were composed of purified NSCs after subculturing.

There have been various AD animal models that seek to understand the causal factors and pathological changes in the disease, but they cannot mimic AD in a comprehensive and accurate way. Comparatively, cerebral pathologic changes in triple Tg mice are similar to those of humans and can be used to simulate part of the neuropathologic changes such as SP, NFT, and decrease of synapses. However, the APP/PS1/tau 3 × Tg strain used in the present study was the only one to harbor 3 major genes associated with AD and has been generated to express both β-amyloid (Aβ) plaques and NFT pathology. ¹¹ The 3 × Tg mice are not only characterized by the early appearance of Aβ plaques, neuronal degeneration, and synaptic loss in the brain but also exhibit the corresponding behavioral disorders similar to clinical presentations of AD. ²³ Thus, APP/PS1/tau 3 × Tg mice are a relatively valuable model for behavioral and pathologic studies of AD and may provide useful experimental clues for the clinical study of AD.

The SP, NFT, and neuronal loss are the most obvious pathological characteristics of AD. However, according to Feng and Gao, ²⁴ cognitive dysfunction correlates best not with SP or NFT but rather with hippocampal synaptic density. It is clear that decreased synaptic density resulting from massive neuronal loss eventually leads to devastating cognitive deficits. Several studies ^{25 –28} have revealed that progressive neuronal degeneration eventually leads to spatial learning and memory impairments in AD. Hagan et al ² also found that the degree of deficit is positively correlated to the amount of neuronal loss. However, exogenous NSCs may prevent or delay the deterioration of the cognitive function. Therefore, with gradual in-depth study of stem cells in recent years, using NSCs to replace degenerating or lost cerebral neurons (namely neuronal regeneration) has become a new strategy for the treatment of AD. ^{3,29 –31}

In the present study, our aim was to determine the effects of transplanted NSCs in 3 × Tg mice by observing behavioral and histopathological changes. Although the hidden platform test showed that the escape latencies decreased along with the repeated MWM training, we found that significant spatial learning and memory impairment was observed in 12-month-old APP/PS1/tau 3 × Tg mice compared to WtC mice in MWM testing. Two months after NSC transplantation, significant chances in the escape latency, path length, times spent in the target quadrant, and number of platform crossings were observed in the NSC group in the behavioral tests, but there was no significant change in Tg mice not receiving NSCs in spite of the escape latency gradually decreased over days, indicating that grafted NSCs improved spatial learning and memory function. Furthermore, quantification of Nissl staining showed that the number of neurons in the NSC group was significantly greater than the PBS group after transplantation, indicating that the impaired neurons regenerated or had been repaired in addition to the improved cognitive function observed in behavioral tests.

Although our study demonstrated that NSCs can facilitate significant functional improvement, the exact mechanism remains unclear. It may be related to (1) neuronal replacement, (2) brain microenvironment, or (3) neuronal circuits reconstruction. Research ^32,33 shows that NSCs that were transplanted into AD mice can migrate to a damaged site and differentiate into neurons and glial cells that express nerve mark protein, replenishing lost of neurons. In addition to replacing neurons, NSCs may secrete or release neurotrophic factors (such as NTF, BDNF, and FGF) into the local environment and consequently protect or repair damaged neurons. ³⁴ Finally, research has shown that NSCs transplanted into the brain can regenerate synapses to establish functional neuronal circuits, release neurotransmitters, and recover cell function. ³⁵

In the present study, we transplanted NSCs into the hippocampus of APP/PS1/tau 3 × Tg mice and conducted quantitative analyses of behavioral and histopathological changes. The results of transplantation suggest that the impaired neurons had been repaired or regenerated with a corresponding improvement in cognitive function. In summary, we believe that NSC transplantation may be a potential strategy for AD treatment.