Infusion of hESC derived Immunity‐and‐matrix regulatory cells improves cognitive ability in early‐stage AD mice

Abstract Objectives In this study, we administered immunity‐and‐matrix regulatory cells (IMRCs) via tail vein (IV) and intracerebroventricular (ICV) injection to 3‐month‐old 5×FAD transgenic mice to assess the effects of IMRC transplantation on the behaviour and pathology of early‐stage Alzheimer's disease (AD). Materials and methods Clinical‐grade human embryonic stem cell (hESC)‐derived IMRCs were produced under good manufacturing practice (GMP) conditions. Three‐month‐old 5×FAD mice were administered IMRCs via IV and ICV injection. After 3 months, the mice were subjected to behavioural tests and electrophysiological analysis to evaluate their cognitive function, memory ability and synaptic plasticity. The effect of IMRCs on amyloid‐beta (Aβ)‐related pathology was detected by thioflavin‐S staining and Western blot. Quantitative real‐time PCR, ELISA and immunostaining were used to confirm that IMRCs inhibit neuroinflammation. RNA‐seq analysis was performed to measure changes in gene expression and perform a pathway analysis in response to IMRC treatment. Results IMRC administration via tail vein injection significantly ameliorated cognitive deficits in early‐stage AD (5×FAD) mice. However, no significant change was observed in the characteristic pathology of AD in the ICV group. Plaque analysis revealed that IMRCs did not influence either plaque deposition or BACE1 expression. In addition, IMRCs inhibited inflammatory responses and reduced microglial activation in vivo. Conclusions We have shown that peripheral administration of IMRCs can ameliorate AD pathology and associated cognitive deficits.

revealed that IMRCs did not influence either plaque deposition or BACE1 expression.
In addition, IMRCs inhibited inflammatory responses and reduced microglial activation in vivo.

| INTRODUC TI ON
Alzheimer's disease (AD) is the most commonly diagnosed agerelated neurodegenerative disease and is characterized by progressive memory decline and cognitive dysfunction. Amyloid deposits, neurofibrillary tangles comprising hyperphosphorylated tau protein and excessive inflammatory response are the main pathologic hallmarks of AD. 1 To date, there is no curative therapy for this disorder.
All the drugs currently used to treat AD only ameliorate the clinical symptoms and have no preventive effect on its pathology. [2][3][4][5][6] These observations underline the need to identify new therapeutic targets for the treatment of AD.
Stem cells have emerged as a potential therapy for a range of neurological insults; however, their application in AD remains limited and the mechanisms underlying the cognitive benefits of stem cells remain to be elucidated. 7,8 Recent studies have highlighted the immune regulatory potential of mesenchymal stem cells (MSCs). [9][10][11] MSCs have emerged as promising agents in combating AD. [12][13][14][15][16] MSCs are adult stem cells that can differentiate into several mesenchymal cell lineages and have the capacity for self-renewal. At present, the main categories of stem cells that can be used in research include bone marrow mesenchymal stem cells (BM-MSCs), 17 adipose-derived mesenchymal stem cells (AD-MSCs), 18 and human umbilical cord mesenchymal stem cells (hUC-MSCs). We have previously identified a hESC-derived MSC-like population with unique abilities in modulating immunity and regulating extracellular matrix production, which we named immunityand-matrix regulatory cells (IMRCs). 19 We showed that the intravenous delivery of IMRCs inhibits both pulmonary inflammation and fibrosis in mouse models of lung injury in a dose-dependent manner.
Additionally, IMRCs were superior to both primary hUC-MSCs and the FDA-approved drug pirfenidone in treating lung injury, and displayed an excellent efficacy and safety profile in both mice and monkeys.
Given the public health crises involving pneumonia, acute lung injury and acute respiratory distress syndrome, our findings suggested that IMRCs are ready for clinical trials to assess their efficacy and safety in the treatment of lung disorders. This study was performed to assess whether IMRCs may also be a suitable therapeutic candidate for AD treatment. We showed that IMRCs administered to AD mice via tail vein injections reduced neuronal loss and improved memory capacity and cognitive deficits by suppressing inflammation.
Only male mice were used because of the gender-specific differences in the progression of AD pathology. Mice were housed at the Laboratory Animal Center of the Institute of Zoology under standard conditions, including a 12:12-hour light/dark cycle, and were allowed free access to food and water. All animal experiments were approved by the Animal Care and Use Committees of the Institute of Zoology, Chinese Academy of Sciences.

| Co-culture assay
BV2 cells were placed on the insert of a Transwell plate (0.4μm polycarbonate filter, Corning, MA, USA), while the IMRCs were placed on the lower chamber. The medium was additionally treated with or without LPS (500 ng/mL) for 12 h during BV2 co-culture with IMRCs.
At the end of the experiment, BV2 cells were washed with PBS and mRNA expression was detected by qPCR.

| Preparation of oligomeric Aβ
Oligomeric Aβ 1-42 (oAβ 1-42 ) was prepared as previously described. 20 Pure Aβ 1-42 peptides were dissolved in hexafluoroisopropanol and volatilized to form a peptide membrane, followed by dissolution in 20 µL of DMSO. Ice-cold phenol-free Ham's F-12 cell culture medium was then added, and the sample was incubated at 4℃ for 24 hours to obtain a 1 mmol/L oAβ 1-42 stock solution.

| Morris water maze test
The water maze was a circular pool (120 cm in diameter, 60 cm in height) with a white inner surface. The escape platform (10 × 10 cm) was fixed in the centre of one quadrant and submerged 1 cm below the water surface. In training sessions, mice were allowed to navigate in the tank to find the hidden platform. If a mouse failed to find the platform within 60 s, it was gently guided to the platform and allowed to stay there for 25 s. Each mouse performed eight training trials per day, starting from different quadrants, for 5 days. Test sessions were performed 24 h after the last training trial. The test session was a single probe test in which the platform was removed and mice were allowed to swim in the tank for 60 s. Behaviours were analysed by video tracking software (EthoVision, Noldus, Netherlands).
Latency to find the platform during the training trials and the time spent in each quadrant during the test session were recorded.

| Y-maze test
The Y-maze apparatus consisted of three radial 30-cm-long arms (named as starting, novel and other arms) originating from the central space to form a 'Y' shape. Mice were placed into the starting arm to explore the maze based on the rodent's innate curiosity to explore novel areas. Briefly, mice were placed into the starting arm to explore and allowed 5 minutes to freely locate the novel arms using spatial clues (training period). After a 2-hour interval, mice were placed into the Y-maze again as part of the training period protocol to evaluate spatial memory. Time, distance, enter times and movement tracks were recorded by an automated video tracking system.

| Novel object recognition
Novel object recognition is widely used in rodents to measure shortterm memory and learning, the preference for novelty and the influence of the hippocampus in the process of recognition. 21 The test was performed in a square-shaped open-field box with objects located opposite the starting point. Briefly, mice were allowed to explore two identical objects (cylinders) in the open field for 10 minutes (learning period). After a 24-hour interval, mice were allowed to explore one familiar object (cylinder) and one novel object (cuboid) as part of the learning period protocol. The time spent exploring familiar and novel objects and the movement tracks of the mice were recorded using a tracking system.

| Open-field activity test
The open-field test measures the exploration of a new environment and anxious behaviour and is based on the idea that mice naturally prefer to be near a protective wall rather than exposed to danger out in an open field. The test was performed in a square-shaped openfield box (as described in the previous section) comprising an inside square (150 mm × 150 mm) as the 'centre area' and an outside square as the 'surrounding area'. Each mouse was gently placed on the floor and allowed to freely explore the area for 10 minutes to investigate their spontaneous locomotor activity. Their overall time spent, distance travelled and movement tracks in the centre and surrounding areas were measured by a tracking system.

| Electrophysiology
Long-term potentiation (LTP). Hippocampal slices were prepared as previously described. 22 Briefly, the brain, including the two hippocampi, was removed and placed into an ice-cold bath con- When the slice to be the hippocampus, according to the experimental needs, slices of hippocampus specific area cut (eg, CA3-CA1 region). The prepared brain sections were transferred to a 32℃ bath

| Immunohistochemistry
Mice were anaesthetized with 2.5% avertin (200 mg/kg body weight) and then perfused with cold PBS followed by 4% paraformaldehyde (PFA). The brains were subsequently removed and postfixed in 4% PFA overnight and then dehydrated with 30% sucrose.
Finally, the brains were coronally sectioned into 40μm-thick slices

| Western blotting
Cultured cells or brain tissues were homogenized in ice-cold RIPA antibodies. After three washes with TBST, the membranes were incubated with horseradish peroxidase-conjugated goat anti-mouse or goat anti-rabbit secondary antibodies at room temperature for 2 hours. The immunoreactive bands were detected using an enhanced chemiluminescence reagent (ECL, Pierce) and quantified using ImageJ software.

| Thioflavin-S staining
For the detection of Aβ plaques, brain sections were first incubated with 0.1% thioflavin-S (Thio-S, Sigma) in the dark for 5 minutes in 50% ethanol, followed by two washes with 50% ethanol and three washes with PBS, and then subjected to antibody staining as described above.

| RNA isolation and RT-qPCR
Total RNA was isolated from mouse brain using the RNAprep Pure

| hESC-derived IMRCs rescued Aβ-induced neural cell damage and LPS-induced inflammation in vitro
Studies have indicated that potential associations exist between MSCs and IMRCs. Several of these have reported that MSCs inhibit LPS-induced pulmonary or inflammatory immune responses in AD. [23][24][25] In this study, hESC-derived IMRCs were prepared as described in our previous research ( Figure 1A-D). 19 The clinical hESC line (Q-CTS-hESC-2) was prepared as previously described. 26 To as-   bodies had adopted a round morphology with shortened, bent, and fractured axons. In contrast, the cells in the CM treatment group displayed and maintained a morphology similar to that of controls ( Figure 2C). In addition, MTT assay results confirmed the increased viability of SH-SY5Y cells after CM treatment ( Figure 2D).
Together, these findings suggested that IMRCs have significant antiinflammatory and neuroprotective properties.

| IMRCs improved spatial learning and memory ability in early-stage AD mice
Next, we investigated whether IMRC injection could alleviate cognitive deficits in the early stages of AD. 5×FAD mice, widely used in preclinical research, exhibit amyloid deposition in the brain and behavioural deficits at 1 and 4 months of age, respectively. 28 Here, we Because 5×FAD mice gradually develop memory deficits that correlate with Aβ deposition at 6 months of age, behavioural tests were performed 90 days after injection when the animals were 6 months old ( Figure 3A). In the open-field test, no significant differences were found for the time spent in the centre square, moving speed or moving distance among the three groups, indicating that IMRCs did not ameliorate motor abilities, exploratory behaviour or anxiety in 5×FAD mice ( Figure 3B). In the Morris water maze test, 31 during the training phase (4 trials per day for 5 successive days), both the IV and ICV groups showed improved latency to the platform compared with that of control mice. There was no significant difference in swimming speed between the 3 groups of mice. In the subsequent probe test phase, control and ICV groups displayed a higher latency to targets and fewer target crossings compared with their littermates in the IV group ( Figure 3C). To assess short-term memory, Y-maze tests were performed. Mice in the IV group also showed improved performance in the Y-maze ( Figure 3D). The novel objective recognition test reflects the learning and memory ability of mice based on their natural tendency to explore novel objects instead of familiar ones when exposed to a novel environment. When administered via tail vein injection, IMRCs ameliorated memory deficits in the AD mice as evidenced by the results of the novel object recognition test ( Figure 3E). Together, these findings suggested that, compared with the NaCl (control) and ICV groups, 5×FAD mice administrated IMRCs by tail vein injection displayed significantly improved spatial learning and memory abilities.

| IMRCs enhanced hippocampal synaptic plasticity in 5×FAD mice
Long-term potentiation (LTP) is characterized by a persistent increase in synaptic strength, a form of synaptic plasticity needed in learning and memory. 32 At the neurophysiological level, AD mice consistently show impaired hippocampal LTP. 33,34 Consequently, we next asked whether IMRCs could also potentially improve synaptic plasticity in AD mice. To this end, we performed brain slice electrophysiology experiments. The NaCl (control) group exhibited a lower slope of evoked field excitatory postsynaptic potential (fEPSP) responses than the IV and ICV groups ( Figure 4A, B). LTP quickly reached baseline levels in the NaCl group, but was maintained above baseline in the IV and ICV groups throughout the recording period.
These results revealed that synaptic plasticity in AD mice was significantly enhanced when IMRCs was administered via tail vein injection. This conclusion was further confirmed by immunofluorescence staining, which showed an increased number of vGLUT1 puncta in hippocampal neurons of mice in the IV group compared with that in the other two groups ( Figure 4C, D).
Combined, the behavioural and electrophysiological studies indicated that the administration of IMRCs via tail vein injection led to a significant improvement in spatial learning and memory compared with that in the control and ICV groups.

| IMRCs did not affect Aβ pathology in 5×FAD mice
Aβ reflects the key neuropathological hall markers of AD pathology. 35 Because BACE1 initiates the formation of Aβ, 36-38 BACE1 inhibition is highly effective in reducing Aβ production. Aβ levels have been reported to be decreased following MSC-based therapy. 16,24,25 To investigate whether IMRC injection can affect BACE1 activity and, consequently, Aβ deposition, we examined BACE1 activity by immunoblot analysis.
No significant reduction in Aβ deposition was seen among the three groups of AD mice at 6 months of age ( Figure 5A, C). To investigate whether the IMRCs affected the amyloid load, we performed thioflavin-S staining and observed no changes in the amyloid load either in the cortex or in the hippocampus ( Figure 5A, B). Western F I G U R E 4 Immunity-and-matrix regulatory cells (IMRCs) enhanced hippocampal synaptic plasticity in 5×FAD mice. a, b, Long-term potentiation (LTP) in the hippocampal CA1 region was induced by high-frequency stimulation (HFS); 5×FAD M-IV , intravenous (IV) IMRC administration; 5×FAD M-ICV , intracerebroventricular (ICV) IMRC administration; 5×FAD NaCl , IV NaCl treatment. A, Averaged slopes of baseline normalized field excitatory postsynaptic potentials (fEPSPs). B, Quantification of mean fEPSP slopes during the last 10 min of the recording after LTP induction (n = 10 slices per group from 3 mice, mean ± SEM, *P < .05; unpaired Student's t test). C, D, Representative confocal images of vGlut1immunostaining in the hippocampus of 5× FAD mice administered IMRCs (5×FAD M-IV ) or NaCl (controls; 5×FAD NaCl ) by tail vein injection (c) and relative quantification (d). Pink, vGlut1; blue, DAPI (n = 5-6 slices per group from 3 mice). Scale bar, 20 µm; data are presented as means ± SEM; ***P < .001, unpaired Student's t test blot results of BACE1 expression also showed no significant differences among the three groups of mice ( Figure 5C, D). These results suggested that IMRC injection is likely to modulate AD pathology through a mechanism other than Aβ deposition. Notably, BACE1 inhibitors may pose a safety risk as, in addition to APP, BACE1 also mediates the cleavage of several other substrates that are important for normal physiology, 39 indicating that IMRCs that do not affect BACE1 may be more suitable for clinical applications.

| IMRCs decrease microglial activation in 5×FAD mice by suppressing inflammation
Substantial evidence supports that Aβ pathology is a key factor in the progression of AD; however, the relationship between Aβ and AD remains contentious. 40 The high failure rate for Aβ-focused drug candidates for the treatment of AD indicates that Aβ may not be the optimal therapeutic target to combat this disease. 41 Dysregulation of the inflammatory system in ageing and AD can also affect brain To further investigate the mechanism in vivo, we sought to identify inflammation-related factors in the hippocampus and cortex by RT-qPCR and ELISA. In the cortex of mice in the IV injection group, the expression of IL-6, encoding a proinflammatory factor, was downregulated, whereas that of CD206, which codes for an anti-inflammatory F I G U R E 5 Immunity-and-matrix regulatory cell (IMRC) treatment does not alter the development of amyloid-beta (Aβ)-pathology in 5×FAD mice. A, B, Representative images of thioflavin-S staining (A) and quantification (B) of the numbers and areas of Aβ plaques in the cortex and hippocampus; 5×FAD M-IV , intravenous (IV) IMRC administration; 5×FAD M-ICV , intracerebroventricular (ICV) IMRC administration; 5×FAD NaCl , IV NaCl treatment. n = 16 to 19 slices from 3 mice per group; data represent means ± SEM; one-way ANOVA with Tukey's multiple comparison test; n.s., nonsignificant. Scale bars: 50 μm (upper); 100 μm (lower). C, D, Representative Western blots (C) and relative quantification (D) of Aβ expression levels in cortical tissues from each group. There was no significant reduction in Aβ deposition in 5×FAD mice (n = 3 mice per group; data represent means ± SEM; one-way ANOVA with Tukey's correction; n.s., nonsignificant) factor, was upregulated ( Figure 6A). In the hippocampus, meanwhile, the expression of the proinflammatory factors IL-6 and TNFα was downregulated, whereas that of the anti-inflammatory factors Arg1, IL-10 and CD206 was increased ( Figure 6B). ELISA for IL-1β, TNFα and IL-10 levels further confirmed these results ( Figure 6C). These data clearly indicate that IMRCs can modulate the inflammatory response. As impaired cognitive function has been linked to central and peripheral inflammation, 49 we also measured cytokine levels in peripheral blood by ELISA and found that the levels of proinflammatory factors were also decreased after IMRC injection ( Figure 6D).
Given that the activation of microglia has been implicated in neuroinflammation during the development of AD, we evaluated microglial activity by immunofluorescence. Microglia in the NaCl (control) group showed a typical activated morphology with hypertrophied cell bodies, whereas microglia in the IV group displayed decreased soma size, suggesting that IMRC injection via the tail vein can inhibit microglial activation ( Figure 6E). RNA-seq analysis revealed that the expression of most of the M1-related (proinflammatory) markers 50 was downregulated in the IV group compared with that of the NaCl group. The expression of M2 (anti-inflammation)-associated genes was upregulated ( Figure 6F).
Taken together, the results demonstrated that IMRC treatment downregulated the expression levels of inflammatory factors in both the brain and the peripheral blood.

| RNA-seq analysis identified biological processes associated with the immune response
To further investigate the role of IMRCs in AD pathology, we performed RNA-seq assays on the brains of mice in the IV (n = 2) and NaCl groups (n = 2). Analysis of the generated heatmap showed that the two groups had distinct expression patterns ( Figure 7A). After IMRC treatment, 134 genes were upregulated and 175 downregulated. Next, the differentially expressed genes (DEGs) were subjected to KEGG pathway and GO enrichment analysis (Figure 7B, The results showed that immune system-associated genes were upregulated. Moreover, we observed an enrichment of pathways associated with neurodegenerative diseases, the immune system and the nervous system ( Figure 7B). GO enrichment analysis of the DEGs revealed an abundance of genes associated with the biological processes of immune response, proinflammatory response, adaptive immunity response and interferon-gamma response ( Figure 7C). This conclusion is identical to the one drawn from our experimental data.

| D ISCUSS I ON
The Alzheimer's Association (AA) 2019 report projected that more than 100 million individuals worldwide will suffer from AD by 2050. F I G U R E 7 RNA-seq analysis of 5×FAD mice after immunity-and-matrix regulatory cell (IMRC) treatment identified the biological process associated with the immune response. A, The expression profile of two distinct clusters of differentially expressed genes (DEGs) between IMRC-treated and control mice. Colouring indicates the log2-transformed fold change. B, KEGG pathway analysis evaluated for the representative profiles of genes involved in Cellular Processes, Environmental Information Processing, Genetic Information Processing, Human Diseases and Organismal Systems. C, GO enrichment analysis in biological process (*q < 0.05; **q < 0.01) Different from adult tissue-derived cells, MSC-like cells differentiated from hPSCs have unique advantages in quality control and large-scale production. We previously showed that hESC-derived

MSC-like cells (IMRCs) exhibit stronger immunomodulatory effects
in the treatment of lung injury and fibrosis compared with that of primary MSC populations. 19 In this study, we report that IMRCs produced under GMP requirements can improve memory ability and enhance cognitive function in AD mice through the suppression of inflammation when administered at an early stage of the disease. Human ESCs provide a limitless supply of IMRCs through cell differentiation.
The efficacy of MSCs has been widely investigated in mouse models of late-stage AD (from 7 to 19 months of age), and good results have been reported. 14,15,24,25,63 In the past few years, substantial progress has been made to understand the pathophysiology and genetic basis of AD. The preclinical stage of AD is considered as the cellular phase. 64 In this stage, alterations in neurons, microglia and astroglia drive the insidious progression of the disease before cognitive impairment is observed. 65 The clinical trial suggests that it might be too late to treat AD and an effective treatment for AD might need an early intervention. Therefore, attempts must focus on increasing early-stage diagnosis and treatment in AD. [66][67][68] Herein, we showed that early intervention (3-month-old mice) can also play a role in disease prevention.
The therapeutic effects of IMRCs were evaluated through IV and ICV routes. All AD mice treated with peripheral MSC therapy exhibited progressive improvement in cognitive function. However, no significant change was observed in the characteristic pathology and symptoms of AD in the ICV group. These results indicated that the ameliorating effect of IMRCs may be related to their potent antiinflammatory effects in the periphery.
Besides therapeutic efficacy, safety is also an important therapy index. 69,70 IMRCs are superior to both primary UCMSCs and the FDA-approved drug pirfenidone in treating lung injury and display an excellent efficacy and safety profile in both mice and monkeys. Regarding AD pathologies, IMRCs did not influence plaque deposition or BACE1 expression, which may be advantageous for AD treatment. Secretase inhibitors have mostly shown disappointing results in clinical trials with an observed worsening of cognitive functions and adverse drug reactions. 4,71,72 In this study, we provided additional evidence for the potential of IMRCs as a novel candidate for use in cell-based therapy.
IMRCs have shown promise in the treatment of AD in vitro and in vivo. This is critical not only for new strategies aimed at AD prevention and early intervention but also for stem cell-based treatment.
Nevertheless, an extensive evaluation of safety, effectiveness and repeatability needs to be carried out before IMRCs can be widely applied in cytotherapy. Project (Z181100001818009 to Q.Z) and National Natural Science

ACK N OWLED G EM ENTS
Foundation of China (31970821 to Y.W).

CO N FLI C T O F I NTE R E S T
The authors declare there is no competing interest, and all authors consented to publish the data.

AUTH O R CO NTR I B UTI O N S
JL, ZH and JW collected and assembled the data, and wrote the manuscript; KL, TG, SY, FM and LW collected and assembled the data; WL provided cell sources; BA and DL contributed to IMRC transplantation and collected the data; YW, JH and BH: conceptualized and designed, analysed and interpreted the data, and gave final approval of the manuscript.

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