Materials
Sprague-Dawley (SD) rats (male, 2 months old, 300–350 g) were purchased from the SLAC Experimental Animal Company (Shanghai, China), then housed under a 12 h light/dark cycle in an air-conditioned room (21 ± 2 ℃). All animals were provided with ad libitum access to tap water and standard chow. All experimental procedures and animal welfare protocols were approved by the Institutional Animal Care and Use Committee at Zhejiang University School of Medicine (Protocol No. ZJU20190068) and strictly adhered to the National Guidelines for Animal Protection.
Stem cell preparation and virus transfection
The Foxg1-positive forebrain NPCs were derived from the human iPSC line DYR0100 based on the updating differentiation method modified from Xu et al.21. Briefly, human iPSCs were cultured on Matrigel (Cat # 354277, Corning) with mTeSR1 medium (Cat # 85850, Stemcell Technologies) for 4 days, then disassociated with TrypLE™ Express Enzyme (Cat # 12605010, Gibco) and cultured in low attachment plates with mTeSR1 medium to form embryonic bodies (EBs). The EBs were cultured for 7 days in suspension and then transferred and attached to Matrigel coated plates. The attached EBs formed Rosette neural aggregates with 14 days culture in neural induction medium containing Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F12, Cat # 11320082, Gibco), 1% N2 supplement (Cat # 17502048, Gibco), 0.1 mM NEAA (Cat # 11140050, Gibco), 1 mM GlutaMAX-I (Cat # 35050061, Gibco), and 2 μg/mL heparin (Cat # H3149, Sigma). The Rosette neural aggregates were then manually isolated to form neurospheres in neurobasal medium (Cat # 21103049, Gibco) containing 2% B27 supplement (Cat # 12587010, Gibco) and 1 mM GlutaMAX-I (Cat # 35050061, Gibco). The following day, the neurospheres were disassociated into single cells using Accutase (Cat # A1110501, Gibco) and placed on poly-D-lysine (Cat # P7886, Sigma) coated plates where NPCs were cultured in differentiation and maturation medium (CN patent 201810298488.9) for further development. The patent medium consisted of neurobasal medium, supplement without B27, 1 × 2-mercaptoenthanol, 0.2 mM ascorbic acid, 20 ng/ml brain-derived neurotrophic factor (BDNF), 20 ng/ml glial cell-derived neurotrophic factor (GDNF), 100 nM SU5402, 200 ng/ml BIBF1120, 10 μM IBMX, and 5 mM glucose. After 4–7 days of culture, the NPCs were disassociated into single cells using Accutase and collected in neurobasal medium at a density of 1.0 × 105 cells/μL, which is suitable for intracranial injections.
For viral transfection, NPCs (1.5 × 105) were added to a 48-well plate, and lentivirus transfection was performed after 3 days of culture at 37 °C. Four lentiviruses carrying different target genes were used, including pLenti-CMV-EGFP-3Flag (ObioTechnology, Shanghai, China) and pLenti-CMV-MCS-EF1a-copGFP (TaiTool Bioscience, Shanghai, China) for labeling grafted progeny cells, pLenti-pCDH-EF1a-Firefly Luciferase-T2A-copGFP (TaiTool Bioscience, Shanghai, China) for bioluminescence imaging tracking, and Lenti-EF1a-Gcamp6s (BrainVTA, Wuhan, China) for Ca2+ imaging. Briefly, 6.98 μL of viral vector was diluted to 400 μL in fresh medium. The vector was transferred evenly to two wells (200 μL/well) in a 48-well plate, followed by incubation for 2 h at 37 ℃, and the addition of 1 mL/well of fresh medium. The plate was returned to the incubator for cultivation at 37 ℃, with the culture medium exchanged with fresh medium after 24 h.
Ischemic brain injury
Cortical photothrombosis was induced in the sensorimotor cortex based on published protocols57. In brief, after anaesthetization with 3% isoflurane, the rats were positioned in a stereotaxic frame. A cranial window (3 × 4 mm) was established over the unilateral sensorimotor cortex (3.5 mm lateral and 0.5 mm anterior to bregma), ensuring that the dura was kept intact. A small piece of opaque tin foil was used to allow light to pass through a 3 × 4-mm hole in the center to irradiate brain tissue. At 2 min before laser irradiation, the rats received an intravenous injection of 1 mL of rose bengal solution (15 mg/mL) via the tail vein. After the cranial window was covered by a masking sheet (3 × 4 mm), focal illumination (3.5-mm diameter focal spot, 25 mW, 532-nm CNI Laser (Changchun, China)) was applied for 15 min to induce occlusion, with medical adhesive then used to cover the skull. The experimental groups included stroke induction + vehicle injection, stroke induction + NPC injection, and sham (rose bengal injection only) groups.
Cell transplantation under MRI guidance
Cell transplantation was performed in accordance with previous research58, with minor adjustments. Before transplantation, all rats underwent MRI scanning to determine the location and extent of infarction (see MRI scanning section). At 7 days after infarction, rats were anesthetized and positioned in a stereotactic device (RWD Instruments, China). In total, 2 × 105 NPCs in a volume of 2.0 μL were injected into the cortex along the anterior-posterior axis (four sites at infarct lesion border) using a microinjection pump (KD Scientific, USA) at a rate of 0.1 μL/min. Before each injection, the stem cell suspension was gently remixed using a 2.5-μL pipette to ensure a constant number of transplanted cells. After each injection, a 33-G Hamilton needle was maintained in place for 5 min, then gradually withdrawn to ensure maximum cell retention at the injection site. The injection coordinates were located at four different points: i.e., 2.5 mm anterior to bregma and 2.8 mm ventral to skull surface; 1.5 mm anterior to bregma and 2.8 mm ventral to skull surface; 0.5 mm posterior to bregma and 2.6 mm ventral to skull surface; and 1.5 mm posterior to bregma and 2.6 mm ventral to skull surface. Body temperature during all operative procedures was monitored (rectal probe) and maintained (heating pad). In this study, all rats in the sham, vehicle and NPC group received an intraperitoneal injection of the immunosuppressant cyclosporine A (MCE, 10 mg/kg) each day for one month and every other day for the remaining months of the experiment5.
Immunostaining
For culture immunostaining, the cells were rinsed using phosphate-buffered saline (PBS), fixed in 4% paraformaldehyde (PFA, 15 min), blocked using blocking buffer (10% donkey serum and 0.2% Triton X-100 in PBS), and incubated overnight (4 ℃) with blocking buffer-diluted primary antibodies. The cells were then washed with blocking buffer (three times), incubated (1 h at room temperature) with corresponding donkey anti-mouse Alexa Fluor 647- or donkey anti-rabbit Alexa Fluor 546-conjugated secondary antibodies (1:500; ThermoFisher), and washed with blocking buffer (three times).
For brain slice immunostaining, the rats were first anesthetized with 1% pentobarbital sodium (0.5 mL/100 g), then rinsed transcardially with 0.9% saline and 4% PFA. The brains were removed and post-fixed overnight in 4% PFA at 4 ℃. Coronal brain slices (40-μm thick) were excised with a sliding microtome (Leica VT1200S, Germany), permeabilized with 0.2% Triton X-100 for 10 min, blocked with PBS solution with 0.5% bovine serum albumin (BSA) and 10% normal goat serum for 1 h at room temperature, and incubated overnight with primary antibodies at 4 °C.
The primary antibodies included goat anti-SOX2 (1:200, SC-17320, Santa), mouse anti-NESTIN (1:1 000, MAB5326, Millipore), rabbit anti-FOXG1 (1:200, ab18259, Abcam), rabbit anti-PAX6 (1:200, 901301, BioLegend), rabbit anti-LHX2 (1:500, AB184337 , Abcam), rabbit anti-NKX2.1 (TTF1, 1:200, ab76013, Abcam), mouse anti-TUJ1 (1:500, AB1637, Millipore), rabbit anti-GFAP (1:1 000, AB5804, Millipore), rabbit anti-MAP2 (1:500, ab183830, Abcam), rabbit anti-BRN2 (1:200, ab137469, Abcam), rabbit anti-TBR1 (1:250, ab183032, Abcam), rabbit anti-VGLUT (1:500, ab77822, Abcam), rat anti-VGAT (1:500, sc-270411, Santo Cruz), rabbit anti-NeuN (1:200, ab177487, Abcam), rabbit anti-synapsin (1:500, ab64581, Abcam), mouse anti-PSD95 (1:500, M1511-4, Huabio), rabbit anti-SATB2 (1:200, ab51502, Abcam), rabbit anti-GABA (1:500, Nbp2-43558, NOVUS), rat anti-somatostatin (SST) (1:200, MAB354, Millipore), and mouse anti-Tau (1:500, Ab80579, Abcam). For identification of transplanted NPCs, the brain slices were co-stained with mouse monoclonal antibodies against human nuclei (HuNu) (1:20, MAB1281, Millipore) or mouse monoclonal antibodies against human cytoplasmic protein (STEM121) (1:500, Y40410, TaKaRa). After rinsing with PBS, the brain slices were incubated with corresponding donkey anti-mouse Alexa Fluor 546- or donkey anti-rabbit Alexa Fluor 647-conjugated secondary antibodies (1:500, A10040 or A32787, Invitrogen) for 1 h at room temperature. After washing (PBS) and counterstaining (4, 6-diamino-2-phenylindole (DAPI) nuclear dye), the slices were fixed on glass slides and fluorescence images were taken using an inverted microscope (Axio Observer 3; Zeiss, Germany) or a Leica sp8 confocal scanning microscope (Germany).
Single-cell nuclear extraction and snRNA-seq library generation and sequencing
At 11 weeks after grafting, the animals were anesthetized with pentobarbital sodium (1%, 0.5 mL/100 g) and perfused with ice-cold PBS. Brains were immediately removed and coronally sectioned with a vibratome (Leica, VT1200S) in ice-cold oxygenated (5% O2 and 95% CO2) artificial cerebral fluid (aCSF) containing (in mM): NaCl 92, KCl 2.5, MgSO4 2, CaCl2·2H2O 2, NaH2PO4·2H2O 1.2, NaHCO3 30, HEPES 20, Na-ascorbate 5, Na-pyruvate 3, thiourea 2, and glucose 25. The GFP-positive regions in brain slices were micro-dissected under a fluorescence stereomicroscope (Leica VT1200S, Germany). To ensure the integrity of nuclear RNA, all solutions were kept on ice and centrifugation was operated at 4℃. The dissected brain tissue was then homogenized in ice-cold RNAase-free homogenization buffer containing: 0.32 M sucrose, 5 mM CaCl2, 3 mM MgAc2, 0.1 mM EDTA, 10 mM Tris-HCl pH 7.6, 0.4 U/μL recombinant RNA inhibitor, 0.1 mM PMSF phenylmethanesulfonyl fluoride, 0.1 mM β-mercaptoethanol, 1% BSA, and 0.01% NP-40 in ultra-pure distilled water. Subsequently, 1.5 mL of 0.32 M sucrose solution containing 0.32 M sucrose, 5 mM CaCl2, 3 mM MgAc2, 0.1 mM EDTA, 10 mM Tris-HCl pH 7.6, 0.4 U/μL recombinant RNA inhibitor, 0.1 mM β-mercaptoethanol, and 1% BSA in ultra-pure distilled water were added to the homogenates, followed by centrifugation at 900 g for 10 min at 4 °C. After withdrawing the supernatant, 1.5 mL of 0.32 M sucrose solution was added to the tube, then resuspended, filtered (35-mm cell-strainer, Cat. No. 352235, Falcon, Corning, USA) and centrifuged at 900 g for 8 min at 4 °C. The precipitate was resuspended in 0.32 M sucrose solution and diluted with an equal volume of 50% OptiPrep density gradient medium (60% OptiPrep density gradient medium, 0.32 M sucrose solution) to give a final concentration of 25% medium solution and centrifuged by 3 000 g for 20 min at 4 °C. After removal of the supernatant, the nuclear pellets were dissolved in 100 mL of PBS. Nuclear density was analyzed using cell counting plates (Cat. No. 177-112C, Watson, Fukae-Kasei, Japan). Single-nucleus capture (target 8 000 nuclei/sample) was performed using a single-cell 3’ Library and Gel Bead Kit V3.1 on the 10X Genomics platform (USA). Single-nucleus libraries from individual samples were pulled and sequenced using the Illumina HiSeq X Ten platform. Nucleus capture and library preparation protocols were performed following the manufacturer’s recommendations (10X Genomics, USA).
scRNA-seq data processing
Reads were aligned to a mixed reference genome (human (hg38) and rat (rn7)) using STAR in CellRanger v3.1.0 with default parameters. Only reads uniquely mapped to the human transcriptome were retained. A unique molecular identifier (UMI) counting matrix was generated and loaded into the R package Seurat (v4.0.0). Cells with less than 500 expressed genes or a minimum of 5% mitochondrial genes were discarded. Genes with less than five detected cells were removed. After filtering, a gene-barcode matrix of 1 988 cells and 15 993 genes was used for downstream analysis. Briefly, the raw UMI counts were normalized to total reads with log-normalization and scaled using a factor of 10 000. The top 2 000 highly variable genes selected by Variance Stabilizing Transformation (VST) were used to perform principal component analysis (PCA). Using 1–14 principal components (PCs), dimension reduction was implemented by t-distributed stochastic neighbor embedding (tSNE). Further cell clusters were identified based on Louvain-clustering. Differentially expressed genes (DEGs) in each cluster were ascertained using Seurat (“FindAllMarkers” function) with default parameters. Cell populations were identified using gene enrichment analysis based on cell-type and layer-specific marker gene sets obtained from published snRNA-seq datasets of the cerebral organoids, developing and adult human cerebral cortices48,59-62.
Pseudo-time analysis across all clusters, except inhibitory neurons, was performed using the R package Monocle2. Briefly, ‘DDRTree’ dimension reduction was performed utilizing 15 PCs based on the top 2 000 highly variable genes. The top 1 000 DEGs across clusters (identified by Seurat) were used to order cells and construct the pseudo-time trajectory. Each ordered cell was assigned a pseudo-time value across pseudo-time trajectory. Pseudo-time genes were identified by Monocle2 with gene expression fit along pseudo-time, with significance at a false discovery rate (FDR) < 0.05. Smooth scatter plots were generated to show dynamic gene expression changes across pseudo-time trajectory. To confirm the accuracy of pseudo-time trajectory, we reperformed pseudo-time analysis using the R package Slingshot. In brief, a lineage dimension reduction implemented by ‘diffusion map’ was generated and loaded into Slingshot. Summarizing cell positions in ‘diffusion map’ embeddings and expression patterns of the top 2 000 highly variable genes, we constructed a pseudo-time trajectory across 16 clusters without supervision.
Electrophysiology
For cell culture electrophysiology, cells were cultured on coverslips, then transferred to a recording chamber perfused with oxygenated (95% O2 and 5% CO2) aCSF containing (in mM): 145 NaCl, 5 KCl, 2 CaCl2, 10 HEPES, 2 CaCl2, 2 MgCl2, and 10 glucose for whole cell patch-clamp recording. In addition, neuronal activity was recorded using the microelectrode array (MEA) system with AxIS software (Axion Biosystems) according to manufacturer’s protocols. Spontaneous activity was recorded for 10 min and the spike-detection threshold was 5.5-fold the standard deviation of noise. The CytoView MEA 48-well plate (Cat # M768-tMEA-48W) for the Maestro MEA system contained 16 electrodes per well. Active electrodes were classified as those electrodes with an average of at least 5 spikes/min.
For brain slice electrophysiology, rats were anesthetized using 1% pentobarbital sodium (0.5 mL/100 g) and perfused with ice-cold oxygenated cutting aCSF containing (in mM): 194 sucrose, 30 NaCl, 4.5 KCl, 0.2 CaCl2, 2 MgCl2, 1.2 NaH2PO4, 26 NaHCO3, and 10 glucose. After quick removal, the brains were sliced (300-μm coronal sections) with a vibratome (Leica VT1200S, Germany) in aCSF, treated for 25–35 min at 34 ℃ in oxygenated routine aCSF containing (in mM): 119 NaCl, 2.5 KCl, 11 glucose, 1.0 NaH2PO4, 26.2 NaHCO3, 2.5 CaCl2, and 1.3 MgSO4, then maintained at room temperature until the experiment. Cell fluorescence was imaged using an Eclipse FN1 microscope (Nikon, Japan), with a 40× water-immersion lens and mercury lamp illumination.
To determine cultured cell and brain slice properties, recording pipettes were loaded with an intracellular solution consisting of (in mM): 125 K-gluconate, 15 KCl, 10 HEPES, 4 MgCl2, 4 Na2ATP, 0.4 Na3GTP, 10 Tris-phosphocreatine, and 0.2 ethylene glycol tetraacetic acid (EGTA). Action potential recordings were obtained with current clamp configuration. Spontaneous EPSCs were obtained under voltage-clamp configuration, with membrane potential held at −70 mV. For miniature iPSC recordings, the extracellular solution contained 1 μM tetrodotoxin (TTX), 5 μM NBQX (AMPAR antagonist) and 50 μM D-AP5 (NMDAR antagonist).
A MultiClamp 700B amplifier and 1440A digitizer (Molecular Devices, USA) were used for whole-cell patch-clamp recordings at room temperature. A series of current steps (50-pA steps, 400 ms) were applied to provoke action potentials. The N-methyl-D-aspartate (NMDA), aminomethylphosphonic acid, and γ-aminobutyric acid A (GABAA) receptors were blocked using D-AP5 (50 μM), NBQX (5 μM), and bicuculine (10 μM), respectively. Responses were filtered (2 kHz), digitized (10 kHz), and evaluated (pClampfit v10.4, Molecular Devices, USA; Mini-Analysis v6.0 Synaptosoft, USA).
Immunoelectron microscopy
Immunoelectron microscopy was conducted based on previously reported protocols63. Rats were first anaesthetized with 1% pentobarbital sodium (0.5 mL/100 g), then perfused transcardially with saline (200 mL), ice-cold 0.05% glutaraldehyde, and 4% PFA (200 mL, pH 7.4) in 0.1 M PBS. Brains were rapidly taken and postfixed by submersion in the same fixative as 4% PFA overnight at 4 °C. Sequential coronal brain sections (50-μm thick), including the cortical region with grafted human cells, were excised with a vibratome (Leica, VT1200S, Germany) and transferred to the same fixative for another 2 h.
Grafted cells were distinguished via immunogold-silver staining. In brief, brain slices were blocked with buffer (1% Triton X-100 in 0.1 M PB and 0.1% BSA-cTM) for 30 min, followed by overnight (4 °C) incubation with primary antibodies (1:200, rabbit anti-GFP, Abcam) and 1-h (room temperature) incubation with secondary antibodies (1:50, Nanogold-labeled goat anti rabbit IgG (H+L), Nanoprobes) diluted with 0.1 M PB and 0.1% BSA-cTM. After rinsing, the sections were postfixed in 2.5% glutaraldehyde in 0.1 M PB for 2 h. Silver enrichment was carried out for 6–8 min in the dark for visualization of GFP immunoreactivity.
Immunolabelled slices were postfixed with 1% buffered osmium tetroxide (OsO4) in 0.1 M PB for 30 min at 4 °C, dehydrated in a graded ethanol series (15 min per grade) and then in 100% acetone, and flat-embedded in Epon 812 amongst plastic sheets. Three to four selected slices from each brain encompassing grafted cell immunoreactivity in the frontal cortex were pruned under a stereomicroscope and fixed on blank resin stubs. Sequential ultra-thin slices were created with an Ultramicrotome diamond knife (Leica EM UC6, Wetzlar, Germany), then placed on formvar-coated mesh grids (6–8 slices per grid) and observed using a 120-kV frozen transmission electron microscope (Tecnai G2 spirit, Czech) equipped with a CCD camera. Synapses were identified based on ≥2–3 synaptic vesicles in a presynaptic terminal, presence of a synaptic cleft, and high postsynaptic density in the postsynaptic structure5.
Adeno-associated virus (AAV)-mediated anterograde transsynaptic tagging
AAV tracing vectors (TaiTool Bioscience, Shanghai, China) were used for anterograde transsynaptic tagging. A volume of 1.5 μL of AAV2/1-hSyn-Cre (titer: 1.23 × 1013 vg/mL) was first injected into the ipsilateral ventral thalamus (AP: -2.85 mm, ML: +3.0 mm, DV: -6.5 mm) 2 months after NPC transplantation, as serum type AAV2/1 exhibits a certain neural anterograde labeling potential42. Then, 2.0 μL of AAV2/8-EF1a-DIO-mCherry-WPRE-HGHpA (titer: 1.27 × 1013 vg/mL) was injected into the four sites (0.5 μL/site) matching the previous NPC transplantation sites. The promoter carried by the AAV2/8-EF1-DIO-mCherry vector is a broad-spectrum EF1α, so all cell types (including host cells and transplanted NPCs) are theoretically tagged when connected to the thalamus. Here, the engrafted progenies could be distinguished from host cells as they carried the GFP marker. We also conducted HuNu staining to distinguish human cells from host cells more reliably, considering the possibility of gene silencing with lentiviral-mediated GFP. Animals were allowed 3–4 weeks to recover following all injections.
Behavioral test
Single-pellet retrieval (SPR) is a widely used task for the precise evaluation of fine sensorimotor function in rat stroke studies57. During the entire SPR experiment, which included pre-training, training, and testing periods, the animals were positioned in a clear Plexiglass box that contained a vertical slit opening (10 × 1 cm) on each side of the front wall. A height-fixed shelf was installed in front of the slit. The rats were trained to pass through the opening to obtain a food pellet (dustless precision pellets/45 mg, banana flavor, Bio-Serv, USA) placed on the shelf. The pre-training period was conducted to identify the dominant forelimb for each rat. After pre-training, rats underwent daily training for 3 weeks before surgery with 50 pellets until a success rate > 60% was achieved over three continuous days. Post-surgical testing was conducted on day 2 after stroke and weeks 1, 2, 4, and 8 after NPC transplantation, with 20 pellets per session per day. The reaching success rate (only considered successful if the pellet was eaten) was used as an indicator for recovery, calculated as [(number of successful reaches / total number of reaches) × 100].
Video-electroencephalography (EEG) monitoring
Video-EEG monitoring was conducted as per earlier research56. EEG screws were attached to stainless steel wires. A stainless-steel screw electrode was implanted in the skull over the transplantation area (M1) adjacent to the infarcted focus as a recording electrode (1 mm anterior and 2 mm lateral from bregma). Two other screws were implanted above the cerebellum as a reference electrode (11 mm posterior and 2 mm lateral to bregma, contralateral to recording electrode) and a ground electrode (11 mm posterior and 2 mm lateral to bregma, ipsilateral to recording electrode). The recording, reference, and ground electrodes were connected to a small homemade plug and stabilized using medical dental cement. All experiments were conducted using freely behaving animals and were performed at 12, 24 and 48 weeks after cell transplantation. All rats were placed in a glass-walled chamber with a multi-channel physiological signal acquisition system (RM-6240B, Chengyi, China) and synchronized video recording (Gz-MG330, Dahua, China). The electrocorticographic signals recorded with the skull screws were defined as EEG recordings. The EEG recordings was visually analyzed using a computer to detect spontaneous seizures, defined as a paroxysmal discharge with rhythmic repetitive waveforms lasting for at least 10 s, with a clear start and end and temporal evolution in amplitude and frequency64. If a seizure was detected, rat behavior was evaluated using video recordings. We analyzed data using custom Matlab software and performed spectral analysis with Matlab (MathWorks) using the wavelet method.
MRI scanning
The MRI data were obtained from a GE 3.0T MR scanner equipped with a rat coil at 3 days before and at 1, 2, and 4 weeks after NPC transplantation. The rats were secured in the prone position and anaesthetized with isoflurane in oxygen (4% for induction, 1.5%–2.5% for maintenance). T2-weighted MRI images was measured using a fast recovery-fast spin echo (FRFSE) sequence with the following parameters: TR = 2 075 ms, TE = 80 ms, FOV = 80 × 80 mm, matrix = 256 × 256, slice thickness = 1.3 mm, spacing = 0.2 mm, voxel size = 0.2 × 0.2 × 1.5 mm.
18F-FDG PET/CT imaging and image analysis
The PET/CT data were obtained via an Argus small-animal PET/CT scanner (Sedecal, Madrid, Spain). Rats were fixed in the prone position while anaesthetized with isoflurane gas anesthesia (4% for induction, 1.5%–2.5% for maintenance). The animals were deprived of food but allowed access to water for 12–20 etc., hours prior to the 18F-FDG injections for PET scans. For PET imaging, ~350 μCi of 18F-FDG (400 μL final volume) was injected intravenously into each rat via the tail vein under 1.5%–2.5% isoflurane gas anesthesia. The PET images were acquired in 3D and reconstructed using OSEM (ordered-subset expectation maximization) algorithm (calculation factor 2.29 MBq/cps), with 16 subsets and 25 iterations. Images corrected for random and scatter events. The CT imaging data were obtained at standard resolution using the scanning parameters: continuous mode, tube voltage 50 kV, tube current 300 µA, number of projections 360, number of shots 9, and axial field-of-view 120 mm. The CT images were restructured with 0.84 Hounsfield for correction attenuation. All scans were recorded without respiratory gating. 18F-FDG accumulation was calculated as the percentage injected dose/gram of tissue using PMOD v.3.902 (PMOD Technologies Ltd., Switzerland).
For metabolic recovery evaluated by PET imaging and analysis, PET data were acquired 6 days before and 1, 3, and 5 weeks after NPC transplantation. Brain PET data were acquired for 10 min using static acquisition mode 40 min after 18F-FDG injection. To assess changes in brain metabolism before and 1, 3, and 5 weeks after NPC transplantation, 3D regions of interest (ROIs) were depicted manually around infarct lesions based on the MRI scans obtained before NPC transplantation. The ROIs in the infarct lesion and pons normal areas were identified from transverse brain section images. The lesion-to-pons (L/P) ratio was used for semiquantitative analysis, calculated as: L/P ratio = mean counts per pixel of lesion ROI / mean counts per pixel of pons region.
For evaluation of EEG and non-EEG seizures, PET data were acquired 12 months after cell transplantation. To assess changes in brain metabolism from EEG and non-EEG seizures post-stroke, the whole-brain cortex was partitioned automatically using an improved rat brain template (Tohoku). ROIs in different cortical regions and pons normal areas were identified from transverse brain section images. For semiquantitative analysis, the M1 brain area was used as the EEG recording region. The region-to-pons (R/P) ratio was calculated as: R/P ratio = mean counts per pixel of M1 brain area / mean counts per pixel of pons region.
For safety evaluation, whole-body PET/CT data were acquired using a whole-body emission protocol for 15 min in two bed positions 1 h after 18F-FDG injection. To assess tumorigenicity, we visually assessed the accumulation of 18F-FDG in various organs, including the brain, heart, lung, stomach, intestine, kidney, liver, and spleen, 12 months after NPC transplantation. Following whole-body 18F-FDG PET scanning, these organs were stained with hematoxylin and eosin and observed under a microscope to rule out any pathological evidence of tumor formation subsequent to forebrain NPC transplantation.
18F-SynVest-1 radiochemistry, PET imaging and image analysis
No-carried added [18F]fluoride was produced by cyclotron bombardment and transferred to an automated radiosynthesizer, where the [18F]fluoride anion in target water (H218O) was trapped on an anionic exchange cartridge Chromafix 45-PS-HCO3, which was preconditioned sequentially with ethanol (5 mL), an aqueous solution of potassium triflate (KOTf, 90 mg/mL, 5 mL), and deionized (DI) water (5 mL). Then the [18F]fluoride was eluted off from the cartridge into a 5 mL reaction V-vial with a mixture of acetonitrile (500 µL), an aqueous solution of potassium carbonate (1 mg/mL, 50 µl), and an aqueous solution of KOTf (10 mg/mL, 450 µL). The eluent was azeotropically dried with anhydrous acetonitrile triplicate under 110°C and a nitrogen bubble. A solution of trimethyltin precursor Me3Sn-SynVest-1 (3 mg) dissolved in anhydrous N,N-dimethylacetamide (DMA, 0.4 mL) was then added to the reaction V-vial, followed by the mixture of pyridine (1 M in DMA, 0.1 mL) and copper(II) triflate (0.2 M in DMA, 67 μL). The reaction mixture was then heated at 110 °C for 20 min, after which the mixture was cooled down and quenched with high-performance liquid chromatography (HPLC) mobile phase (3 mL) and injected into a semi-preparative HPLC for purification (Phenomenex Synergi Hydro-RP column, 10 × 250 mm, 4 µm; mobile phase: 1.25 mL/min acetonitrile and 3.75 mL/min 0.1% aqueous trifluoroacetic acid). The fraction containing purified 18F-SynVesT-1 was collected by monitoring with a radioactivity detector and then diluted with DI water (40 mL). The dilution was passed through a C18 cartridge, preconditioned with ethanol (5 mL) and DI water (10 mL), and the product was trapped on the C18 cartridge. The C18 cartridge was washed with DI water (10 mL) and USP-grade ethanol (0.4 mL). Subsequently, the purified radiotracer was then eluted off with USP-grade ethanol (0.6 mL) into a sterile vial and diluted with USP-grade saline (6 mL). Finally, the mixture was sterilized by a sterile membrane filter (Millex-GV, 0.22 μm) and collected in another sterile vial to afford a formulated solution ready for administration.
The PET/CT data were acquired using an Argus small-animal PET/CT scanner (Sedecal, Madrid, Spain). The rats were immobilized in a prone position following induction with isoflurane gas anesthesia (4% for induction, 1.5%–2.5% for maintenance). Prior to the 18F-SynVest-1 injections for PET scans, the animals were fasted but allowed access to water for a period of 12-20 hours. For PET imaging, each rat was intravenously injected with 300-350 μCi of 18F-SynVest-1 (400 μL final volume) via the tail vein under 1.5%–2.5% isoflurane gas anesthesia. The acquisition of PET images was performed utilizing a 3D and subsequently reconstructed through implementation of the OSEM algorithm (with a calculation factor of 2.29 MBq/cps), utilizing 16 subsets and 25 iterations. The resulting images were corrected for both random and scatter events. The CT imaging data were obtained at a standard resolution, utilizing scanning parameters that consisted of continuous mode, a tube voltage of 50 kV, a tube current of 300 µA, 360 projections, 9 shots, and an axial field-of-view of 120 mm. The CT images underwent reconstruction with a correction attenuation of 0.84 Hounsfield. Notably, all scans were conducted without respiratory gating. To assess alterations in synapse intensity, PET data were collected prior to and at 1, 2, and 4 weeks post NPC transplantation. Brain PET data were obtained through static acquisition mode, 30 minutes after 18F-SynVest-1 injection, for a duration of 30 minutes. 3D regions of interest (ROIs) were manually delineated around the NPC transplantation sites, based on MRI scans acquired prior to transplantation. The ROIs within the NPC transplantation regions were identified from transverse brain section images and quantified as the standard uptake value (SUV) utilizing PMOD v.3.902 (PMOD Technologies Ltd., Switzerland). For semiquantitative analysis, the standardized uptake value ratio (SUVr) of ROI was calculated using the brain stem as a reference region, based on a previous quantification study of SV2A binding in rodents65.
In vitro and in vivo bioluminescence imaging
The IVIS Lumina II Imaging System (PerkinElmer, USA) was used for bioluminescence imaging following previously reported protocols66. For in vitro imaging, the cells were seeded in a 48-well plate, washed with Dulbecco’s Phosphate-Buffered Saline (D-PBS), and incubated at 37 °C with medium containing 150 μg/mL D-luciferin (Gold Biotechnology, USA) for 5 min before analysis, with signals detected using the above imaging system. For in vivo imaging, the rats first received an intraperitoneal injection of D-Luciferin (150 mg/kg) (Synchem, Germany) and were then anesthetized with oxygenated 2% isoflurane. A series of images were acquired at 5–30 min post-injection. The bioluminescence signals in a fixed ROI were detected using IVIS image analysis software (PerkinElmer, USA), with data quantified in units of p/s/cm2/sr.
QUANTIFICATION AND STATISTICAL ANALYSIS
Data are presented as mean ± standard error of the mean (SEM), with P < 0.05 considered statistically significant. Semiquantitative analysis of PET images was conducted using AMIDE v9.2 (Stanford University). GraphPad Prism v6 was used to determine statistical differences, with Student’s t-test and one-way analysis of variance (ANOVA) applied for comparisons between two groups and among multiple groups, respectively.