If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
1 Chunxia Chen and Xihe Tang contributed equally to this work.
Chunxia Chen
Footnotes
1 Chunxia Chen and Xihe Tang contributed equally to this work.
Affiliations
School of Pharmaceutical Sciences, Shenzhen Campus of Sun Yat-sen University, Shenzhen, Guangdong P. R. ChinaDepartment of pharmacy, The People's Hospital of Guangxi Zhuang Autonomous Region & Guangxi Academy of Medical Sciences, Nanning, Guangxi P. R. China
1 Chunxia Chen and Xihe Tang contributed equally to this work.
Xihe Tang
Footnotes
1 Chunxia Chen and Xihe Tang contributed equally to this work.
Affiliations
Department of neurosurgery, The People's Hospital of Guangxi Zhuang Autonomous Region & Guangxi Academy of Medical Sciences, Nanning, Guangxi P. R. ChinaDepartment of neurosurgery, Aviation General Hospital, Beijing, P. R. China
1 Chunxia Chen and Xihe Tang contributed equally to this work.
Zhaohui Lan
Footnotes
1 Chunxia Chen and Xihe Tang contributed equally to this work.
Affiliations
Key Laboratory for the Genetics of Development and Neuropsychiatric Disorders (Ministry of Education), Shanghai Key Laboratory of Psychotic Disorders, and Brain Science and Technology Research Center, Bio-X Institutes, Institute of Psychology and Behavioral Sciences, Shanghai Jiao Tong University, Shanghai, P. R. China
Department of Emergency, The People's Hospital of Guangxi Zhuang Autonomous Region & Guangxi Academy of Medical Sciences, Nanning, Guangxi, P. R. China
Key Laboratory for the Genetics of Development and Neuropsychiatric Disorders (Ministry of Education), Shanghai Key Laboratory of Psychotic Disorders, and Brain Science and Technology Research Center, Bio-X Institutes, Institute of Psychology and Behavioral Sciences, Shanghai Jiao Tong University, Shanghai, P. R. China
Department of Neurology, The People's Hospital of Guangxi Zhuang Autonomous Region & Guangxi Academy of Medical Sciences, Nanning, Guangxi, P. R. China
The CLU rs11136000C mutation (CLUC) is the third most common risk factor for Alzheimer's disease (AD). However, the mechanism by which CLUC leads to abnormal GABAergic signaling in AD is unclear. To address this question, this study establishes the first chimeric mouse model of CLUC AD. Examination of grafted CLUC medial ganglionic eminence progenitors (CLUC hiMGEs) revealed increased GAD65/67 and a high frequency of spontaneous releasing events. CLUC hiMGEs also impaired cognition in chimeric mice and caused AD-related pathologies. The expression of GABA A receptor, subunit alpha 2 (Gabrα2) was higher in chimeric mice. Interestingly, cognitive impairment in chimeric mice was reversed by treatment with pentylenetetrazole, which is a GABA A receptor inhibitor. Taken together, these findings shed light on the pathogenesis of CLUC AD using a novel humanized animal model and suggest sphingolipid signaling over-activation as a potential mechanism of GABAergic signaling disorder.
CLUC has emerged as the third most common risk factor for AD. However, the underlying mechanism is unclear. This study establishes the first chimeric mouse model of CLUC AD. Increased of GAD65/67 and a high frequency of spontaneous releasing events occurred in this model. Moreover, cognition was impaired and Gabrα2 was higher in CLUC hiMGEs chimeric mice. Interestingly, cognitive impairment in chimeric mice was reversed by pentylenetetrazole treatment. The over-activation of sphingolipids signaling mainly contributing to disordered GABAergic signaling caused by CLUC. Our findings shed light on the pathogenesis of CLUC AD by establishing a humanized animal model.
INTRODUCTION
Alzheimer's disease (AD), a neurodegenerative disease characterized by progressive brain atrophy,
Safety and efficacy of losartan for the reduction of brain atrophy in clinically diagnosed Alzheimer's disease (the RADAR trial): a double-blind, randomised, placebo-controlled, phase 2 trial.
due to its role in excitatory and inhibitory signaling (E/I) imbalance. As GABAergic neurons interact with astrocytes to form tripartite synapses, they are necessary for GABAergic homeostasis.
In the AD brain, hyperactive astrocytes impair neural synapses via elevated intracellular Ca2+ transients and release excessive GABA inhibitory neurotransmitters.
This vicious cycle of pathogenic interactions among aggregations of amyloid-β (Aβ), neurofibrillary tangles of hyperphosphorylated tau, and GABAergic dysfunction exacerbates AD pathogenesis.
With recent genome-wide association studies (GWAS) involving larger populations, the CLU rs11136000C mutation (CLUC) has been identified as the third strongest genetic risk factor for AD.
A study of 85 healthy participants genotyped for CLU rs11136000 and assessed by magnetic resonance imaging (MRI) during a working memory task (WM) found that the neural functional network associated with memory was significantly impaired in people with CLUC.
However, until now, there has been no reliable animal model of this specific mutation.
The ability to transform somatic cells into induced pluripotent stem cells (iPSC) by specific transcription factors is a major breakthrough in cell reprogramming.
One advantage of this breakthrough is the ability to reprogram the somatic cells of patients with specific genetic defects to establish disease-specific or patient-specific iPSCs, enabling research on the pathogenesis of the disease and providing a promising tool for AD modeling.
Additionally, since iPSCs are derived from autologous somatic cells, this approach avoids immunologic rejection and ethical issues associated with the use of embryonic stem cells (ESCs).
Genotype-dependent, disease-associated phenotypes have also been displayed under CRISPR/Cas9 genome-modified technologies to generate a panel of isogenic knock-in human iPSC lines carrying APP and/or PSEN1,
However, sporadic AD cell and animal models with CLUC have not yet been established using the iPSC platform and, as a result, the underlying mechanisms remain unclear.
In neurodegenerative disease research, human-induced neural precursor cells (hiNPCs) better reflect neural development for reproducing the occurrence and development of diseases.
In a previous study, we successfully reprogrammed healthy human peripheral blood mononuclear cells (PBMNCs) directly into hiNPCs and then confirmed that the cells survived, differentiated, and formed synapses well in host mice.
In the present study, PBMNCs from AD patients with CLUC were reprogrammed to hiNPCs using a similar reprogramming method with the Sendai virus to generate a CLUC cell line, which was further differentiated into a group of GABAergic precursor cells during neuronal maturation termed medial ganglionic eminence progenitors (MGEs). A humanized mouse model was then constructed to study the pathogenesis of AD mutation at this locus.
RESULTS
CLUC hiNPCs and GABAergic neurons derived from CLUC hiNPCs secrete high levels of GABA
To establish an in vitro cellular model of CLUC, we generated hiNPC lines (n = 3) from human CLUC mutation PBMNCs using Sendai viral vectors encoding SOX2, KLF4, OCT3/4, and c-MYC (Fig 1A). Age-matched hiNPCs derived from healthy people served as a control (n = 3). PBMNCs were expanded for 14 days and then transfected with the Sendai virus. Eleven days after transfection, hiNPC clones appeared and were selected after 10 days. The PBMNCs were reprogrammed to hiNPCs at a high efficiency, as indicated by the expression of the NPC markers SOX2, NESTIN, and PAX6 as well as the proliferation marker Ki67 (Fig 1, B and C, and Supplementary Fig 1, A). The primer sequences used are listed in Supplementary Table I. All 6 reprogrammed hiNPC lines showed a normal karyotype (Supplementary Fig 1, B). Polymerase chain reaction (PCR) sequencing revealed a point mutation (rs11136000) in the CLU locus in 3 of the AD patient-derived hiNPCs (Fig 1, D).
Fig. 1Establishment of the CLUC hiNPCs cell line and differentiation into hiMGEs. A, Schematic representation of the hiNPCs induction procedure. B, Representative immunostaining of NPC markers. C, RT-PCR analysis of markers expressed by expanded hiNPCs. D, Genotypic analysis of a representative CLUC hiNPC cell line by PCR amplification and sequencing. E, Schematic procedure for directed differentiation of hiNPCs to GABAergic interneurons. Representative immunostaining (F and G) and Western blot analysis (H and I) of hiMGEs. Representative immunostaining for GABA interneuron (J and K) markers in vitro. L, Quantification of GABA neurotransmitter concentration corresponding to the peak area of each group (n = 3). *P < 0.05. Scale bar = 100 μm. Data are presented as the mean ± SD. Cont, Control group, CLUC, CLU rs11136000C mutation group (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Next, hiNPCs were differentiated toward GABAergic neurons (Fig 1, E). On day 37, nearly all cells expressed the MGE markers FOX1 (86%) and NKX2.1 (93%) (Fig 1, F and G). High expression of FOX1 and NKX2.1 in hiNPCs-derived MGEs (hiMGEs) was also demonstrated by Western blot analysis of the 3 CLU mutation cell lines (Fig 1, H and I). On day 48, an average of 94% ± 4% and 95% ± 7% cells expressed the GABAergic neuronal system marker NKX2.1 and the neurotransmitter GABA, respectively (Fig 1, J and K). Supernatant from hiMGEs neuronal culture was collected to quantify the GABA level by high performance liquid chromatography (HPLC). Compared to controls, the GABA level in CLUC hiMGEs was significantly higher (Fig 1, L).
CLUC hiMGEs neuronal chimeric mice showed high GAD65/67 levels and dysfunction of grafted GABAergic neurons
To observe neurogenesis in CLUC AD, lentiviral-CMV-EGFP-puro labeled CLUC hiNPCs were transplanted into the dentate gyrus (DG) of the right hippocampus. Six months after transplantation, about 75% of CLUC hiNPCs expressed the mature neurons marker NeuN, which was not significantly different than the control (Cont) hiNPCs; however, 22% of grafted cells expressed the astrocyte marker GFAP, which was higher than in transplanted Cont hiNPCs (Supplementary Fig 2, A and B). Imaging analysis revealed that the morphology of CLUC hiNPCs-derived astrocytes differed from Cont hiNPCs-derived astrocytes. Next, GFP and GFAP positive cells were traced using the Fiji plugin software simple neurite tracer and Sholl analysis (Supplementary Fig 2, C). We found that the numbers of primary and secondary branches were both greater and the ramification index was higher in CLUC hiNPCs-derived astrocytes (all P < 0.05) (Supplementary Fig 2, D–F), suggesting increased astrogenesis in CLUC hiNPCs.
To examine the differentiation of transplanted hiNPCs towards GABAergic neurons, we co-stained the brain sections with GFP and GAD65/67. As shown in Supplementary Fig 2, G, grafted cells (GFP positive) expressing GAD65/67 denoted that the hiNPCs differentiated to GABAergic neurons. A much higher GAD65/67 fluorescence intensity was observed in CLUC hiNPCs-derived cells than in Cont hiNPCs (P < 0.05) (Supplementary Fig 2, H). These data indicate excessive differentiation toward GABAergic neurons or dysfunction of GABAergic neurons during the CLUC hiNPCs differentiation procedure. Notably, none of the grafted cells expressed NESTIN or Ki67 (Supplementary Fig 2, I–K) 6 months after transplantation.
We then focused on CLUC GABAergic neurons. We differentiated lentiviral-CMV-EGFP-puro labeled hiNPCs to hiMGEs and then stereotaxically injected them into the DG of the right hippocampus (Fig 2, A). Six months after transplantation, at least 62% of cells in the mouse brain were GFP positive (Fig 2, B). Both immunofluorescence (Fig 2, C and D) and Western blot (Fig 2, E and F) analyses revealed higher GAD65/67 protein expression in CLUC hiMGEs chimeric mice. These data suggest the prominent role of GABA transmitter synthesis abnormalities in the subsequent pathogenesis of CLUC AD.
Fig. 2Delayed neuronal development and abnormal neuronal function 6 months after CLUC hiMGEs transplantation. A, Schematic diagram of GFP labeled hiMGEs. B, Flow cytometry analysis showing the percentage of GPF positive cells in mice transplanted with CLUC hiMGEs for 6 months (n = 3). C, Representative image of double immunofluorescent staining for GFP and GAD65/67 6 months after transplantation of hiMGEs. Scale bar = 40 μm. D, Quantification of GAD65 expression in vivo (n = 3). E and F, Representative scan strip and quantities relative to GAD65/67 protein expression. G, Recording of a GFP positive neuron in a hippocampal slice. H, Action potential induced by a depolarizing current ramp from 0 to +200 pA in transplanted CLUC hiMGEs derived GABAergic neurons. I, Spontaneous postsynaptic currents (sPSCs) and spontaneous postsynaptic potentials (sPSPs) recorded under the voltage clamp (V-clamp) and current clamp (I-clamp) modes, respectively. J, Representative traces of mIPSCs from transplanted CLUC hiMGEs derived GABAergic neurons and Cont hiMGEs derived GABAergic neurons (above) voltage clamped at -70 mV. K, The frequency of mIPSCs was increased 2-fold (n = 5, P < 0.05), whereas the amplitude (L) exhibited an upward trend (n = 5, P > 0.05) in engrafted CLUC hiMGEs derived GABAergic neurons. ns, no significant difference. *P < 0.05, **P < 0.01. Data are presented as the mean ± SD (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Next, to evaluate the function of transplanted neurons, patch-clamp recordings of brain slices were performed using a fluorescence microscope to locate GFP-positive neurons followed by a bright field to capture the target cells (Fig 2, G). Engrafted cells were elicited to generate action potentials under ramp current injections from 0 to +200 pA (Fig 2, H), which were characterized by a series of spikes with a trend of decreased amplitudes. To determine if transplanted cells produce functional synapses with host cells, spontaneous postsynaptic currents (sPSCs) and spontaneous postsynaptic potentials (sPSPs) were recorded (Fig 2, I). As expected, sPSCs and sPSPs were detected in GFP positive neurons. To test whether the CLUC hiMGEs derived GABAergic neurons-mediated inhibitory effect was due to changes in pre– and postsynaptic activity, miniature inhibitory postsynaptic currents (mIPSCs) were recorded in the presence of tetrodotoxin (1 μM), which blocks the opening of voltage-gated sodium channels and thus prevents the generation of action potentials. With the voltage-clamp set at −70 mV, the frequency and amplitude of mIPSCs of CLUC hiMGEs derived GABAergic neurons were altered relative to Cont group, as shown in Fig 2, J. The results showed a significant increase in the frequency (Cont, 1.1 ± 0.6; CLUC, 2.3 ± 0.4; P < 0.05; Fig 2, K) and a trend of increased amplitude (Cont, 17.0 ± 2.0; CLUC, 20.5 ± 4.0; P > 0.05; Fig 2, L) of mIPSCs in CLUC hiMGEs derived GABAergic neurons, suggesting greater synaptic inhibition in grafted CLUC hiMGEs.
Cognitive impairment and AD-related pathology are induced by transplantation of CLUC hiMGEs after 6 months
To evaluate the effects of CLUC hiMGEs on cognitive function in the humanized mice model, APP/PS1 mice and matched WT mice as a control underwent behavioral testing. To exclude the effect of cell reprogramming, Cont hiNPCs and Cont hiMGEs formed other control pairs. Starting 3 months after transplantation of CLUC hiMGEs, signs of aging were gradually observed in the mice, including reduced action and white hair. The transplanted mice also showed abnormal cognitive function, as demonstrated by the Morris maze results. In the navigation trial (during the first 5 days), normal control mice (WT group) could quickly find the platform (latency to platform) starting from day 3. However, a smaller number of APP/PS1 group mice (as a positive control) managed to reach the platform from day 4 onwards. Similar to the APP/PS1 group, only a few animals reached the platform during the day 5 trials (Supplementary Fig 3, A and B). As shown in Supplementary Fig 3, C, compared with the WT group, the escape latency was increased and the number of target platform crossings and quadrant occupancy were decreased in all 3 groups in the probe test, including the APP/PS1, CLUC hiNPCs, and CLUC hiMGEs groups. These results confirm successful establishment of the chimeric mouse model by cell transplantation for 3 months.
Relative to performance 3 months after transplantation, WT mice could find the platform faster 6 months after cell transplantation. However, most mice in the CLUC hiNPCs and CLUC hiMGEs groups struggled to reach the platform within 60 seconds (Fig 3, A and B). When the underwater platform was removed, mice in the CLUC hiNPCs and CLUC hiMGEs groups took longer to reach the previous platform site, crossed the target platform fewer times, and spent less time in the target quadrant compared to the WT group (Fig 3, C). WT mice that received transplanted CLUC hiMGEs showed a trend of more severe cognitive impairment, as evidenced by a longer escape latency, fewer target platform crossings, and less time spent in the target quadrant compared with the APP/PS1 and CLUC hiNPCs groups.
Fig. 3Impaired cognitive function and AD-related pathology in mice transplanted with CLUC hiMGEs for 6 months. A, Representative navigation training trial (day 5) of the Morris maze test (n = 10). B, Latencies to find the submerged platform during the 5 days of navigation trials. The probe test was performed on day 6 and the submerged platform was removed. C, The time taken for the mouse to find the removed platform, the frequency of crossing the target platform, and the time spent in the target quadrant were recorded. For the Y maze test (n = 10), the total number of arm entries was counted (D) and spontaneous alternations were calculated (E). F, Representative immunohistochemical staining images of p-Tau and Aβ1-42 (n = 5). G, Average optical density value of p-Tau and Aβ1-42 in the 5 groups. H, Representatives Western blot bands of t-Tau and Aβ1-42 expression (n = 3). I, Quantification of relative protein expression. J, The cells were cultured and the culture medium was collected for ELISA detection (n = 3). K, p-Tau and Aβ levels in different cells during reprogramming. 1, WT group, 2, Cont hiNPCs group, 3, Cont hiMGEs group, 4, CLUC hiNPCs group, 5, CLUC hiMGEs group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the mean ± SD (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
To further confirm the effect of CLUC hiMGEs on WT mice, the Y maze was employed. Both the total number of arm entries (Fig 3, D) and spontaneous alternations (Fig 3, E) were significantly decreased in the APP/PS1 and CLUC hiMGEs groups. Notably, WT mice treated with Cont hiNPCs and Cont hiMGEs exhibited similar behavior as the WT group, which excludes the effect of reprogrammed cells on the mouse brain but supports that transplantation of CLUC hiMGEs causes cognitive decline in normal WT mice.
Since transplantation of CLUC hiMGEs induced cognitive decline, we then explored whether AD-related pathological changes also occur in mice. Two typical pathology indexes, namely p-Tau/t-Tau and Aβ1-42, were used to evaluate the effect of transplantation of CLUC hiMGEs on WT mice. Microscopy of hippocampal sections confirmed higher p-Tau and Aβ1-42 average optical density (OD) (Fig 3, F and G) in CLUC hiNPCs and CLUC hiMGEs chimeric mice; in contrast, the Cont hiNPCs and Cont hiMGEs groups were similar to the WT group. The Western blot results supported these pathological changes, as shown by similar patterns (Fig 3, H and I). The protein concentration at the cellular level was also determined (Fig 3, J). The levels of p-Tau and Aβ1-42 in the supernatant of the 3 cell types during the programming and differentiation processes including PBMNCs, hiNPCs, and hiMGEs were higher in the above-mentioned 3 CLUC cell lines (Fig 3, K). Based on these results, CLUC cells with high expression of p-Tau and Aβ1-42 cause behavioral disorders and pathological changes in the brains of humanized mice.
Transcriptome analyses reveal altered gene expression in WT mice transplanted with CLUC hiMGEs
To explore the underlying mechanism of CLUC hiMGEs-induced dementia, RNA sequencing was performed using WT mice brains transplanted with CLUC hiMGEs for 6 months and WT mice engrafted with Cont hiMGEs and WT mice as controls (n = 3, respectively). A total of 22,274 genes were detected, of which 18,260 had known sequences (81.98%). Comparing gene expression between the CLUC hiMGEs and WT groups identified only 8 differentially expressed gene (DEGs), including Gabra2 (Supplementary Fig 4, A), with a false discovery rate (FDR) <0.05 and | log2 fold change | >1. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed that the 8 DEGs were mainly enriched in GABAergic synapses, neuroactive ligand-receptor interactions, and ether lipid metabolism (Supplementary Fig 4, B). First, the gene transcripts encoding GABAergic signaling pathways were examined. As shown in the heatmap (Supplementary Fig 4, C), the CLUC hiMGEs group had higher gene expression levels of Gabrα2, Gabrβ2 (GABA A receptor subunit beta 2), and Gabrβ1 (GABA A receptor subunit beta 1). Reactome analysis indicated that these genes were enriched in pathways of interest, including GABA transport Cl− from extracellular regions to cytosol (which support the electrophysiological findings of brain slices) and GABA receptor activation (Supplementary Fig 4, D).
To explore additional genes in the pathways of interest, the DEG screening standards were adjusted to P < 0.05 and fold change >1.5 There were 129 DEGs between the CLUC hiMGEs and WT groups (C57BL/6J mice) (Fig 4, A). The target genes related to Gabrα2, lipid metabolism, and ion channels (Fig 4, B), and the most significant enrichment pathway according to KEGG is lipids and atherosclerosis (Fig 4, C). These results suggest that Gabra2 upregulation, Cl− hyperactivity, and lipid imbalance occur in chimeric mice engrafted with CLUC hiMGEs.
Fig. 4Transcriptional analysis of CLUC hiMGEs chimeric mice and verification in clinical samples. A, Venn diagram of differentially expressed gene (DEGs) between the WT, CLUC hiMGEs, and Cont hiMGEs groups with conditions of a 1.5 fold change and P < 0.05 (n = 3). B, Heatmap of genes related to Gabrα2, lipid metabolism, and ion channels in all samples. 1–3, WT group, 4–6, Cont hiMGEs group, 7-9 CLUC hiMGEs group. C, KEGG pathway analysis of target genes. Representative images of immunohistochemical analysis (D) (n = 5), Western blots (E) (n = 3), and quantification of Gabrα2 protein expression (F) in the retrosplenial cortex. HPLC analysis of GABA neurotransmitter levels in the serum of grafted mice (G) (n = 4) and samples from clinical patients (H) (n = 9). *P < 0.05, **P < 0.01, ***P < 0.001. Data are presented as the mean ± SD (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Gene set enrichment analysis (GSEA) was then used to compare the CLUC hiMGEs group with the WT group. The former played a more positive role in GABAergic synapse (Supplementary Fig 4, E). DEGs related to other neurological diseases were also mapped by GSEA and showed links to AD-related pathways with the transplantation of CLUC hiMGEs (Supplementary Fig 4, F). Interestingly, WT mice transplanted with CLUC hiMGEs showed a larger portion of T-C base transition (Supplementary Fig 4, G).
Next, GABAergic pathways were successively verified in the mice and clinical samples. PCR was initially performed using mice brain homogenate, and the results showed that Gabrα2 expression was significantly higher in WT mice transplanted with CLUC hiMGEs (Supplementary Fig 4, H). The expression of Gabrα2 in the retrosplenial cortex of mice that received CLUC hiMGEs injection was 5.4-fold and 1.5-fold higher than that of WT mice, as shown by immunohistochemical staining (Fig 4, D) and Western blot (Fig 4, E and F), respectively. HPLC was used to detect the GABA level in mouse serum. As shown in Fig 4, G, the level was higher in mice treated with CLUC hiMGEs than WT mice. GABA levels were also higher in the serum of AD patients with CLU mutation (Fig 4, H). Taken together, these results demonstrate dysfunction of the GABAergic pathway in CLUC AD.
GABAergic neuron imbalance and AD-related pathology in hiMGEs with low CLU expression
To further explore the relation between CLUC and AD, a CLU knock-down hiNPCs (shCLU hiNPCs) cell line was established using an interference carrier (pLV-hU6- CLU shRNA03-hef1a-mScarlet-P2APuro) with the target sequence (CCAGGAAGAACCCTAAATTTA) (Fig 5, A). shCLU hiNPCs were successfully transfected (Fig 5, B–D) and then differentiated to hiMGEs (shCLU hiMGEs). To accelerate the progression of AD lesions, 4 groups were created: Cont hiMGEs (Cont), shCLU hiMGEs (shCLU), and Aβ1-42, and shCLU hiMGEs that received Aβ1-42 treatment (AD-acce). In the AD-acce group, the protein levels of GAD65/67 and Gabrα2 increased by 62% and 1.3-fold (Fig 5, E–G), respectively. In addition, t-Tau protein expression was 1.8-fold higher in the AD-acce group compared to the Cont group, while Aβ1-42 protein expression did not show a significant difference (Fig 5, E, H, and I). The increased expression of GA65/67, Gabrα2, and t-Tau protein mimicked the process of CLUC AD, supporting that GABAergic pathway dysfunction and AD-related pathology are induced by CLU gene deficit.
Fig. 5AD-related pathology in CLU knock-down (shCLU) hiMGEs. A, Construction of a plasmid vector carrying the target sequence. Representative band (B), quantity of clusterin assessed by Western blot (C) (n = 3), and quantity of CLU gene expression assessed by RT-PCR (D) (n = 3) in shCLU hiNPCs. Representative band (E) and quantities (F–I) of characteristic protein (GAD65/67, Gabrα2, t-Tau, and Aβ1-42) expression of hiMGEs (n = 3). *P < 0.05. Data are presented as the mean ± SD (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
The CLU gene encodes apolipoprotein J, which participates in the lipid metabolism process. As shown by Fig 4, B and C, lipid metabolism is an important pathway in CLUC hiMGEs chimeric mice. Lipomics is thus a suitable method to investigate the relationship between CLUC and the GABAergic pathway. To do so, the culture hiMGEs media from the 3 groups (Cont, shCLU, and CLUC) were analyzed. A total of 32 lipid classes and 1016 lipid species were identified. 49 differential lipid (DL) species were screened (Fig 6, A) between the CLUC hiMGEs and Cont groups, which were mainly clustered in the Sphingolipids including ceramide (Cer) and sphingomyelin (SM) (Fig 6, B) subtypes. Lipid molecular content in SMs with a carbon chain length of 34 or 44 was higher in the CLUC hiMGEs group than in the Cont group (Fig 6, C). The fatty acid chain saturation of SM was somewhat higher in the CLUC hiMGEs group than the Cont group (Fig 6, D). Next, clinical serum samples were used to verify Sphingolipids molecules. The results confirmed that the contents of Cer and SM were all higher in CLUC patients compared to healthy controls (Fig 6, E and F). Taken together, these data suggest that changes in sphingolipids content, mainly due to an increase in Cer, may contribute to the GABAergic signaling anomalies caused by CLUC.
Fig 6Lipidomics findings related to CLUC. A, Volcano plot of DL species between the CLUC hiMGEs and Cont groups (n = 3). B, Clustering heatmap of DL. Carbon chain length statistics of SM (C). Fatty acid chain saturation of SM (D). Levels of Cer (E) and SM (F) in clinical serum samples (n = 7). *P < 0.05, **P < 0.01. Data are presented as the mean ± SD (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Pentylenetetrazole (PTZ) improves recognition memory in AD chimeric mice
RNA sequencing and verification data suggest that upregulation of Gabrα2 plays an important role in CLUC AD progression. Thus, we investigated whether inhibition of Gabrα2 improves memory. Seven months after transplantation, mice performed behavioral tests including the Morris maze and Y maze. For the Morris maze, the time to reach the platform in CLUC hiMGEs chimeric mice that received PTZ treatment decreased on the day 5 trial, although it did not reach statistical significance (Fig 7, A). In the probe test, chimeric mice treated with PTZ had a shorter escape latency (Fig 7, B) and greater number of target platform crossings (Fig 7, C), and spent more time in the target quadrant (Fig 7, D). In the Y maze test, chimeric mice treated with PTZ showed more arm entries (Fig 7, E) and had a higher spontaneous alternation rate (Fig 7, F). AD chimeric mice without PTZ treatment still showed significant cognitive impairment compared with WT mice. The RT-PCR results revealed a relative decrease in the mRNA expression of 9 subunits belonging to GABA A receptors in dementia mice treated with PTZ (Fig 7, G), as well as a significant decrease in Gabrα2. Analysis of serum samples confirmed that the higher levels of Cer and SM in CLUC hiMGEs chimeric mice were significantly decreased by PTZ treatment (Fig 7, H and I), indicating that PTZ may improve recognition memory in AD chimeric mice through regulation of the sphingolipid signaling pathway.
Fig 7Pentylenetetrazole (PTZ) treatment reversed cognitive dysfunction in CLUC hiMGEs chimeric dementia mice. A, Latency to find the platform during the 5 days of Morris maze navigation trials (n = 10). Escape latency (B), number of target platform crossings (C), and quadrant occupancy (D) were recorded during the probe test. The number of arm entries and spontaneous alternations (%, F) in the Y maze test (n = 10). G, Relative expression level of GABA A receptor (including 9 subunits) in CLUC hiMGEs chimeric mice with or without PTZ treatment (n = 3). Serum levels of Cer (H) and SM (I) in CLUC hiMGEs chimeric mice (n = 6) *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the mean ± SD (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
In this study, PBMNCs of AD patients carrying the CLU gene mutation were reprogrammed into hiNPCs. The primary observation was that GABA synthase GAD65/67 was significantly increased in these cells. Subsequently, we differentiated CLUC hiNPCs to hiMGEs progenitors, injected them into the hippocampus of immunosuppressed mice, and successfully constructed a CLUC humanized mouse model for the first time. Dysfunction of GABAergic neurons and abnormally high GABA neurotransmitter secretion were confirmed in this model. Chimeric mice exhibited behavioral dysfunction 3 months after cell transplantation (compared to age-matched WT control mice), and these effects persisted 6 months later. RNA sequencing revealed that Gabrα2 expression was significantly upregulated in chimeric mice. Up-regulated Gabrα2 was further verified in the retrosplenial cortex of an established dementia mouse model. Next, we showed that memory loss in dementia mice was reversed by PTZ treatment, which is a GABA-A receptor inhibitor. Similarly, AD-related pathology was induced in a shCLU cell model. Alteration of lipid homeostasis may thus being the biochemical basis of GABAergic signaling abnormalities.
Somatic reprogramming of PBMNCs into neural fate cells is an efficient and time-saving research method, as previously described,
in which epithelial-like cells from urine are reprogrammed into NPCs. The main advantages of this method are that blood samples are easy to obtain and PBMNCs are easy to separate and purify.
In the present study, hiNPCs from healthy people mainly differentiated to NeN positive neurons, as well as a few GFAP positive astrocytes, after 1 month of differentiation, which is consistent with previous reports.
Notably, higher GAD65/67 fluorescence intensity was observed in CLUC-hiNPCs-derived neurons both in vivo and in vitro. MAO-B inhibitors have been shown to restore memory impairments in APP/PS1 mice
possibly by inhibition of GABA synthesis. Therefore, we speculated that GABAergic signaling plays an important role in AD. Next, CLUC hiNPCs were differentiated into hiMGEs in vitro. We found that the secretion of GABA transmitters in CLUC hiMGEs increased in a time-dependent manner. Notably, previous studies have reported aberrant GABA production in reactive astrocytes.
Postmortem brain tissues from AD patients show reduced GABA levels in multiple regions, including the frontal cortex, hippocampus, and occipital cortex.
However, this study did not perform classification analysis of risk genes in AD patients. In the present study, increased GABA levels were detected in serum samples from CLUC hiMGEs of chimeric mice and AD patients carrying CLUC. Moreover, we found that the GABAergic neurons derived from CLUC hiMGEs in cheremic mice made functional connections with host neurons, as observed in cerebral cortical slices. This result was further evidenced by patch-clamp recordings of APs, sPSCs, and sPSPs. The frequency of mIPSCs was greatly increased in GFP positive CLUC hiMGEs compared with Cont hiMGEs. Combined with the change in the frequency of spontaneous releasing events reflecting presynaptic effects, our data indicate presynaptic dysfunction in CLUC hiMGEs.
Simple verbal memory tests are reported to show moderate-to-strong associations with visuospatial working memory in humans.
In AD mouse models such as APP/PS1 and 5 × FAD, spatial-based working memory tests, such as the Morris water maze and alternation tasks (Y-maze), are widely used to assess behavior.
In our study, mice with CLUC hiNPCs and mice with hiMGEs grafts derived from model mice both showed cognitive impairment at 3 months. Moreover, mice transplanted with these 2 cell lines showed memory impairments by 6 months. Cognitive dysfunction was most severe in mice transplanted with CLUC hiMGEs, compared to the APP/PS1 and CLUC hiNPCs groups, as evidenced by the Morris water maze and Y-maze test results. These results support the important role of the CLU gene in cognition, consistent with other studies.
Interestingly, dementia induced by transplantation of CLUC hiMGEs for 7 months was significantly reversed by treatment with PTZ, which is a GABA A receptor inhibitor. To our knowledge, this is the first sporadic AD animal model with CLU loci mutation established using somatic reprogramming technology. A previous study of female ApoE4 knock-in mice reported spatial learning deficits at 16 months using the Morris water maze.
Another report described a chimeric AD model established by grafting hiPSC-derived neurons with the ApoE4 genotype, which is similar to the method used in our study; however, the prior study did not describe the onset time or duration of memory impairment.
The presence of higher levels of Tau oligomers, an intermediate form of tau, correlates with memory loss, as demonstrated in studies injecting WT mice with human brain tau oligomers
Based on the association of higher expression of clusterin in neurons without tau aggregation in nondementia cases, and lower expression of clusterin in neurons with tau aggregation in autopsy samples of AD patients, clusterin may play a protective role in neuronal degeneration.
We used p-Tau (S404), the phosphorylated form of tau, to quantify tau protein in mice. In mice transplanted with CLUC hiMGEs, higher levels of p-Tau protein clusters were found in cell membranes and in brain tissue next to grafted cells. Aβ deposition was also observed in grafted cells and near grafted cells, consistent with previous studies.
These findings suggest that neurons derived from hiNPCs of AD patients with CLUC result in alterations in Aβ deposition and tau aggregates, which could be associated with memory impairment. Some studies have reported that clusterin assists in clearing cerebral Aβ load,
which is consistent with our study. Given that the expression intensity and range of p-Tau protein in the dementia model were larger than those of Aβ, we speculate that the pathological features of AD in this mutation type are mainly due to tau overexpression. Thus, tau is considered a correlate of cognitive dysfunction in this mutation type.
As a crucial component of the GABAergic signaling system, there is growing evidence of GABA receptor dysfunction in AD
. There are 2 types of GABA receptors: GABA A and GABA B. GABA A receptors comprise up to 16 subunits (including α1-6, β1-3, γ1-3, δ, ε, θ, π) arranged around an integral Cl− channel.
International Union of Pharmacology. LXX. Subtypes of gamma-aminobutyric acid (A) receptors: classification on the basis of subunit composition, pharmacology, and function.
TPA-023 attenuates subchronic phencyclidine-induced declarative and reversal learning deficits via GABA (A) receptor agonist mechanism: possible therapeutic target for cognitive deficit in schizophrenia.
RNA sequencing showed that Gabrα2 and Cl− channels are upregulated in WT mice transplanted with CLUC hiMGEs. Consistent with a previous report that the retrosplenial cortex is an important site for GABAergic receptors,
Gabrα2 was highly expressed in our novel AD model. An increased Gabrα2 level was also confirmed in serum samples from CLUC AD patients. These results suggest that GABAergic neuron dysfunction and upregulation of Gabrα2 in the retro splenial cortex may be associated with the pathogenic phenotypes observed in the CLUC AD animal model.
Lipomics was employed to further explore the mechanism linking CLUC and GABAergic impairment using cultured hiMGEs. The results revealed that CLUC resulted in sphingolipid imbalance, as manifested by increased Cer and SM levels. Cer can be converted to SM, thereby contributing to cell senescence.
The following evidence supports over-activation of the sphingolipid pathway leading to accelerated neuronal senescence as a potential mechanism of CLUC-induced dementia. Firstly, abnormalities in sphingolipid metabolism—including accumulation of Cer and SM and decreased sphingosine-1-phosphate (S1P) levels—accelerate aging, as seen in the AD brain.
The humanized model developed in this study showed sharp upregulation of Gabrα2 and Cer, supporting this notion. Secondly, as a second messenger of the IL-1β-dependent pathway, the permeable analogue of Cer, 2-Cer, has been shown to increase spontaneous GABA events by increasing the frequency of mIPSC.
Consistent with this report, we found that CLUC hiMGEs exhibited a higher frequency of mIPSC. Thirdly, Cer and Aβ synergized with each other to accelerate aging and neuronal cell death. Cer stabilizes the APP cleavage enzyme 1 (BACE1) and promotes Aβ biogenesis.
The present study detected Cer and SM upregulation in CLUC hiMGEs culture medium, which is consistent with the development of AD-related pathological changes in chimeric mice. To the best of our knowledge, our study provides the first evidence that PTZ inhibits high Gabrα2 expression in the mouse brain induced by transplanted CLUC hiMGEs and attenuates cognitive impairment. Moreover, PTZ decreased Cer and SM levels in AD cheremic mice, consistent with a previous study.
Based on these findings, we speculate that the underlying mechanism through which PTZ treats CLUC AD is related to regulation of the sphingolipid signaling pathway.
In summary, this study established a novel model to study AD pathogenesis using CLUC humanized mice. Model animals showed increased enzyme synthesis (GAD65/67) of the GABA neurotransmitter in AD GABAergic interneurons and active synaptic release function. Notably, memory impairment, AD-related pathology, and Gabrα2 upregulation were all observed in mice receiving CLUC hiMGEs injections. Our results support that over-activation of the sphingolipid pathway may be the underlying mechanism of these effects.
METHODS
hiNPCs generation and characterization
Control and CLUC hiNPC lines were generated from human PBMNCs using the Sendai virus (Invitrogen) infection method, as previously described.
Blood samples were obtained from the People's Hospital of Guangxi Zhuang Autonomous Region, China. This study was approved by the Ethics Committee of the People's Hospital of Guangxi Zhuang Autonomous Region (No.KY-LW-2020-17). Written informed consent was given by the parents in each case. Three AD hiNPC lines carrying CLUC were generated from 3 patients. The same number of control hiNPC lines was generated from 3 healthy individuals. Quality control of hiNPC lines was performed by karyotyping. The identity and pluripotency of hiNPCs were established by clonal expansion with various markers including SOX2, NESTIN, PAX6, OCT4, and Ki67. The differentiation properties of hiNPCs in vitro were observed over 4 weeks of spontaneous differentiation procedures. hiNPCs were seeded on PDL/laminin coated coverslips with DMEM/F12 (Gibco) supplemented with 1 × N2 (Stemcell), 1 × B27 (Gibco), 1 × NEAA (Gibco), and 1 × GlutaMAX (Gibco). GABA interneurons were differentiated from hiNPCs through 2 steps based on a previously reported method
with some modifications. For more details, see the Supplemental Experimental Procedures.
Animals and cell transplantation
Male 8-week-old C57BL/6J mice were obtained from Changsha Tianqin Biotechnology Co., Ltd. (Hunan, China). Male 16-week-old APP/PS1 Tg (APPswe, PSEN1dE9)85Dbo/Mmjax mice were purchased from Cavens Co., Ltd. All experimental procedures followed the guide for the Care and Use of Laboratory Animals (NIH publication N. 85-23, revised in 1996) and were approved by the Animal Ethics Committee of Guangxi Medical University (No. 202008002).
GFP lentivirus (Guangzhou Getein, China) labeled hiNPCs and their derived MGEs (at day 15 of hiNPCs differentiation) were both intracranially injected into the dentate gyrus (DG) of the right hippocampus. Approximately 5 × 106 cells were injected in a suspension of 5 μL PBS. The injection site was located 2 mm behind the bregma, 0.75 mm to the right of the midline, and 2 mm below the skull, as determined using a brain stereotactic instrument. To avoid host rejection, cyclosporine A (7 mg/kg, Novartis Pharma Stein AG, Swiss) was injected subcutaneously 1 day before transplantation. At 6.5 months after cell transplantation, some mice underwent PTZ (1 mg/kg) treatment for 2 weeks.
Behavioral tests
Morris maze
The Morris maze test was used to evaluate cognitive function at 3, 6, and 7 months after cell transplantation. The test apparatus consisted of a circular pool with a diameter of 120 cm and an automatic tracking camera system (HKvision, Huangzhou, China).
Y maze
The Y maze test was performed 6 and 7 months after cell transplantation. The apparatus consisted of 3 arms (length × width × height = 40 cm × 10 cm × 20 cm). The mice were placed at the intersection of the 3 arms and allowed to move freely for 5 minutes. The total number of times the mice entered the 3 arms and the number of entries into the 3 arms in order (alternating times) were recorded. Spontaneous alternations were calculated.
Electrophysiology
CLUC hiMGEs chimeric mice were anesthetized with 1% sodium pentobarbital and perfused with cold (0 °C) artificial cerebrospinal fluid (ACSF) containing (in mM) 125 NaCl, 25 NaHCO3, 3.5 KCl, 3 MgCl2, 1.25 NaH2PO4, 0.1 CaCl2, and 10 glucose (295 ± 5 mOsm; pH 7.3; oxygenated with 95% O2 and 5% CO2). The brains were removed and 300 µm thick coronal hippocampal slices were cut using a vibratome (Leica VT1200S) submerged in ACSF (0 °C) at a speed of 0.08 mm/s. After incubation in 32 °C ACSF for 30 minutes, slices were allowed to recover in room temperature ACSF for 1 hour. The slices were then transferred to a recording chamber and immersed in a continuous flow of ACSF driven by a peristaltic pump (LongerPump BT100-2J) at a speed of 4 rpm. A patch pipette that can produce 6 MΩ resistance was made by a pipette puller (Model P-97, Sutter instrument) using a glass pipette (A-M system) and filled with solution containing (in mM) 115 K gluconates, 20 KCl, 10 KOH, 10 HEPES, 2 MgCl2, 4 Na2ATP, 1 Na3GTP, and 0.4 EGTA (pH 7.2, 290–310 mOsm). GFP positive neurons were randomly selected for electrophysiological recording. Action potentials were fired in a current clamp and spontaneous postsynaptic currents (sPSCs) were recorded under −70 mV using a voltage clamp. All data were collected using a patch clamp amplifier system (Multiclamp 700B, Axon instruments) in the digital mode, transformed by a low-noise data acquisition system (Digidata 1550B, Axon instruments), and analyzed using Clampex software (Axon instruments).
RNA isolation and RT-PCR
Total RNA extraction was performed using the RNAeasy extraction kit (TRIzol reagent, Invitrogen) according to the manufacturer's instructions. The Fast All-in-One RT kit (ES Science, China) was used for reverse transcription of RNA to DNA. PCR reactions were carried out using a 7500 Real Time PCR System (Applied Biosystems). The reactions consisted of 10 μL 2x Super SYBR Green qPCR Master Mix (ES Science, China), 15 ng cDNA, 0.4 μL forward primer, and 0.4 μL reverse primer. All primers are listed in Supplementary Table I. GAPDH was used as the reference. mRNA expression in the samples was calculated using the 2−△△Ct method.
Immunostaining
Immunofluorescence analysis was performed as previously described.
All primary and secondary antibodies are listed in Table S2. Briefly, cells were fixed with 4% paraformaldehyde and then blocked with fetal bovine serum (BSA).
Brains were removed and cut into 15 μm slices using a freezing microtome (Leica CM1860 UV). Next, the sections were fixed with acetone for 10 minutes and dehydrated with 30% sucrose for 15 minutes. The slice staining protocol was the same as the cell staining protocol. Images were captured using a Leica confocal laser-scanning microscope (TCS SP8, Germany). Fluorescence intensity was analyzed using Image J software (NIH image) with the control group as a reference. To assess the morphological complexity of AD GABAergic interneurons, GFP stained positive neurons and their axonal branches were traced and analyzed using the Simple Neurite Tracer.
Immunohistochemistry of mice brain sections was performed using routine methods. Images were obtained using a microscope (Olympus BX51, Germany) and analyzed using Image J software (NIH image).
Flow cytometry
GFP labeled cells were digested by accutase, centrifuged at 650 × g for 3 minutes at room temperature, and then collected. After washing twice with PBS, cells were suspended in PBS and analyzed using FACSCanto Ⅱ (BD Biosciences) and FACSdiva software (BD Biosciences). The signals were stimulated by a blue laser 488 nm and received using the FITC channel.
High performance liquid chromatography (HPLC) analysis of GABA levels
GABAergic neurons were differentiated on PLO/laminin-coated 6-well plates. The cell culture supernatant was collected on day 27 after NPCs differentiated into GABAergic.
Chromatography was performed using an Ultimate 3000 system (Thermo Scientific) equipped with a fluorescence detector and autosampler. The target substance was separated using a hypersil gold column (250 × 4.6 mm, 25,005–254,630, Thermo Scientific) with 5 μL injection volume and 30 °C column temperature. The mobile phase consisted of methanol and H2O (40:60, v/v) at a flow rate of 1.0 mL/min. The excitation wavelength and emission wavelength were set at 338 nm and 425 nm, respectively. The GABA content of the sample was calculated by converting the peak value corresponding to the concentration in the standard solution (at the same retention time).
Western blot
Proteins were extracted using the BCA method, separated by SDS-PAGE, and then transferred to methanol-acidified PVDF membranes. PVDF membranes containing protein information were blocked, incubated with the primary antibody (1:1000) at 4 °C overnight, and then incubated with the secondary antibody at room temperature for 1 hour. Lastly, the target protein was developed using a 2 color infrared laser imaging system (Odyssey Fc).
RNA sequencing
RNA sequencing was performed for WT mice transplanted with CLUC hiMGEs for 6 months (n = 3) and control WT mice (n = 3). Total RNA was extracted using the TRIzol reagent kit (Invitrogen). Libraries were constructed by isolation of mRNA, synthesis and purification of cDNA double strands, end repairing, addition of the A base, screening of 100 bp cDNA (AMPure XP beads), and amplification. Tests, including agarose gel electrophoresis, revealed that the brightness of the 28s rRNA band was twice that of the 18s rRNA, and the RNA integrity number (RIN) was close to 10, which is the main index of library quality control. The products were sequenced using an Illumina Novaseq6000 system (Illumina) and the raw reads were filtered by fastp (version 0.18.0). After alignment with the reference genome (Ensembl_release104) using HISAT 2.2.4 with “-rna-strandness RF,” the mapped reads of each sample were assembled using StringTie v1.3.1. Next, the fragments per kilobase of transcript per million mapped reads (FPKM) value was calculated using RSEM software to quantify expression abundance.
DEGs were determined using DESeq2 software with the criteria of fold change >1.5 and P value < 0.05. KEGG pathway enrichment analysis identified significant transduction pathways in DEGs. A heatmap was drawn using the “pheatmap” package in R software. GSEA was performed to identify differences in the set of genes in the KEGG pathway between the 2 groups. GATK (version 3.4-46) was used for single-nucleotide polymorphism (SNP) analysis.
Lipidomics
Absolute quantitative lipidomic analysis was performed as previously described.
Briefly, the supernatant was vortex mixed with the moderate internal lipid standard mixture, namely methyl tert-butyl ether (MTBE). After undergoing cryogenic ultrasound for 20 minutes, standing at room temperature for 30 minutes, and centrifugation at 14,000 × g for 15 minutes, the upper organic phase was acquired. Samples were separated by Nexera LC-30A ultra high performance liquid chromatography (Shimadzu) using a CSH C18 column (1.7 µm, 2.1 mm × 100 mm, Waters) and then detected using a Q Exactive mass spectrometer (MS, Thermo Scientific) with electrospray ionization (ESI) in positive and negative ion modes. Lipid molecules and internal standards were analyzed using LipidSearch software (Thermo Scientific), including identification of lipid quantity, lipid composition analysis, and differential lipid (DL) analysis.
Statistical analysis
All data are reported as the mean ± standard deviation (SD). Navigation trials in the Morris maze were analyzed using two-way analysis of variance (ANOVA). Other data were analyzed using one-way ANOVA followed by the least significant difference (LSD) post hoc test. Differences with P < 0.05 were considered statistically significant. The analyses were done in SPSS 20.0 and GraphPad Prism 8.0.
DATA AVAILABILITY
Original data for the figures in this study are available from the lead contact, W.B. Deng.
ACKNOWLEDGMENTS
Conflicts of Interest: All authors have read the journal's policy on disclosure of potential conflicts of interest and have none to declare.
This work was supported by the National Natural Science Foundation of China (82260258, 81960246, 81701089, 81772449, and 81971081), the Innovation and Technology Fund of Shenzhen (JCYJ20180307154653332), Science, Technology & Innovation Commission of Shenzhen Municipality (JCYJ20180307154606793), the Innovation and Technology Fund of Guangzhou (201803010090), Shenzhen Science and Technology Program (KQTD20190929173853397), and the Guangxi Natural Science Foundation (2020GXNSFAA238003 and 2017GXNSFBA198010). All authors have read the journal's authorship agreement and that the manuscript has been reviewed by and approved by all named authors.
Author contributions are as follows: C.X.C., X.H.T., Z.H.L., G.Y., M.C.Y., and W.B.D. designed the research study and had full access to all data. X.H.T. and C.X.C. performed cell reprogramming. Z.H.L. and W.D.L. performed electrophysiological recording. C.X.C., W.C., Y.G., and H.S. carried out most experiments, Y.X.L. and X.Z. collected clinical samples. W.C collected the data. C.X.C., X.H.T., and Z.H.L. drafted the manuscript and summarized all data. G.Y., M.C.Y., and W.B.D. revised the manuscript.
Safety and efficacy of losartan for the reduction of brain atrophy in clinically diagnosed Alzheimer's disease (the RADAR trial): a double-blind, randomised, placebo-controlled, phase 2 trial.
International Union of Pharmacology. LXX. Subtypes of gamma-aminobutyric acid (A) receptors: classification on the basis of subunit composition, pharmacology, and function.
TPA-023 attenuates subchronic phencyclidine-induced declarative and reversal learning deficits via GABA (A) receptor agonist mechanism: possible therapeutic target for cognitive deficit in schizophrenia.
00087-7&id=gr1.jpg){kind=link}
00087-7&id=gr2.jpg){kind=link}
00087-7&id=gr3.jpg){kind=link}
00087-7&id=gr4.jpg){kind=link}
00087-7&id=gr5.jpg){kind=link}
00087-7&id=gr6.jpg){kind=link}
00087-7&id=gr7.jpg){kind=link}