Soluble TGF-β decoy receptor TGFBR3 exacerbates AD lesions by modifying microglial function

Acquisition and preprocessing of the RNA-seq cohort from AD patientsAn independent RNA-seq cohort for the temporal cortex was collected from the Mount Sinai School of Medicine cohort in the Accelerating Medicines Partnership AD reprocessed RNA-seq (Synapse ID: syn3157743). CDR 0 (n = 11), CDR 0.5 (n = 17), CDR 1 (n = 16), CDR 2 (n = 23), CDR 3 (n = 17), CDR 4 (n = 18), CDR 5 (n = 17). General information about all samples, including Clinical dementia rating (CDR), Braak staging, plaque, age, and sex, was recorded in the original data.Gene Set Enrichment Analysis (GSEA)All procedures were carried out as previously described. CDR score = 0 of the RNA-seq cohort was classified as a normal group, and the other samples were classified as AD patient groups. The results were generated by GSEA v4.1.0 and visualized by the R package (grid).Weighted Gene Co-Expression Network Analysis (WGCNA)All procedures were performed as previously described. Briefly, the optimal soft threshold =7 for the samples was calculated, and was chosen to construct a network. According to TOM dissimilarity, the genes were clustered to generate a cluster dendrogram. Finally, interesting modules with P < 0.01 were extracted according to the correlation between lesions and modules.Protein‒Protein Interaction (PPI) AnalysisThe PPI network was generated using STRING v11.5 and visualized by Cytoscape v3.7.1. 180 shared genes were obtained using MCC (CytoHubba) and CentiScaPe 2.2 methods.Gene Functional and Pathway Enrichment AnalysisBiological functions and pathways for differentially expressed genes were analyzed using online server Metascape. Terms with P < 0.01 were deemed to be statistically significant.Convergent functional genomic (CFG) ranks CFG ranks were used to assess the correlation between genes and AD using eQTL, GWAS, PPI, early gene expression regulated by AD genetic variants, and Aβ and tau pathology in Alzdata database. The CFG score ranges from 0 to 5. The higher the CFG score of the gene, the stronger the correlation with the molecular mechanism of AD. Human brain tissueHuman brain tissues (temporal cortex) from AD subjects and normal subjects were obtained in the form of frozen blocks by the China Brain Bank Center at the South-Central University for Nationalities. For protein extraction, each sample was homogenized in RIPA buffer (wt/vol: 1:7) containing protease and phosphatase cocktail inhibitors (Sigma). Information on human brain samples of normal controls and patients with AD is contained in Supplementary Table 1.Total RNA extraction and real-time PCR (RT-PCR) determinationFor AD patient brain samples, total RNA was extracted using TRIzol reagent (Qiagen), and cDNA was synthesized from total RNA (2 μg) using the first-strand synthesis system (Vazyme). The cDNA was diluted to 2 ng/µL, and then 4 µL to 10 µL of 2×FastStart Universal SYBR Green PCR Master Mix (Vazyme) was added. Each sample was tested in triplicate using the StepOnePlus qPCR system (Applied Biosystems 7500). The Ct value was normalized to the housekeeping gene GAPDH, which was amplified in parallel. The following qPCR primers were used: TGFBR3-F, 5′-TGGTGTCTGAGGGTTCTGTGG-3′ and TGFBR3-R, 5′- TTCTTGCTATCTTGAGTTCGGTG-3′; GAPDH-F, 5′-TGGTGTCTGAGGGTTCTGTGG-3′ and GAPDH-R, 5′- TGATGACCCTTTTGGCTCCC-3′.Mice and cell cultureAPP/PS1 mice (6 months old, 20 ± 1 mg) and age-matched wild-type (WT) C57BL/6 mice were purchased from Beijing HFK Bioscience Co. Ltd. and fed at a room temperature of 26 ± 1°C under 12 hours of light/dark cycles under standard conditions with unlimited access to standard rodent chow and clean water. The mouse cell line BV2 was acquired from the Pricella Cell Line Bank (Procell, China). BV2 cells were cultured in minimum essential medium (Procell, China) containing 10% fetal bovine serum and a 1% antibiotic–antimycotic mixture kept at 37°C with 5% CO2.Intracerebroventricular (ICV) injectionMice were anesthetized with isoflurane (3% isoflurane for induction, 2% isoflurane for maintenance), and were immobilised on a brain stereotaxic apparatus. The injection coordinates were: bregma: -0.5 mm, interaural line: -1.1 mm, cranium: -3.0 mm, injection rate: 1 μL/min, with an interval of 1 min, and the needle was stopped for 3 min after the last injection. The incision was then sutured and disinfected, and 0.1 mL of penicillin was injected intramuscularly to prevent infection. lentiviral-expressing TGFBR3 shRNA was purchased from Wanleibio (viral titer: 1×10^8 TU/mL). Experimental mice (n= 7-10) received 3 μL of lateral ventricle injection each, and WT mice (n= 7-10) received the same dose of control lentivirus.Acute LPS/STZ injectionLipopolysaccharide (LPS, 0.5 mg/kg; Sigma) was administered by intraperitoneal injection according to a previously described procedure [38]. Briefly, the mice were treated with 0.5 mg/kg LPS twice, each injection was separated by 24 hours, and saline was administered to the control group. The whole brain was collected 3 hours after the second injection for immunofluorescence experiments. STZ (50 mg; Sigma) was administered to the lateral ventricle according to a previously described procedure [39]. Briefly, each mouse received a single ICV injection of 3.0 mg/kg STZ in 3.0 μL 0.9% saline into the left ventricle of the brain, and relevant detection was conducted 21 days later.Morris water mazeLearning and memory were measured using the MWM test previously reported [40]. For the spatial exploration test, mice were trained in two daily trials for 6 days (22 ±2°C). Each mouse was trained for 60 s, and the interval between two pieces of training was more than 10 min. The mouse found the platform within 60 s, stayed for 5 s, and finished the training task. Otherwise, training results will be suspended and automatically recorded. On day 7, the platform was removed for the directional navigation experiment. The time spent crossing the platform for the first time, cross-platform number, quarter preference, and velocity were recorded.ImmunoblottingThe temporal cortex and hippocampus of WT, model mice and sTGFBR3 deficiency mice were dissected and obtained. the BCA method then was used to evaluate the protein concentration after extracting total protein. The protein samples were separated by 10% SDS polyacrylamide gel electrophoresis and transferred to PVDF membranes. The PVDF membrane was sealed with a 5% skim milk shaking table at room temperature for 2 h and then incubated with primary antibody at 4°C overnight. A secondary antibody was added for incubation for 1 h. Finally, detection was performed using a chemiluminescent imaging system (Bio-Rad, ChemiDoc) and data were analysed using ImageJ.ImmunofluorescenceHuman brain specimens were embedded in OCT reagent and then sliced on a sliding microtome in 10-micron increments (Thermo, Cryotome E). The sample was incubated with the primary antibody overnight at 4 ℃ and then incubated with a suitable fluorescent probe bound to the secondary antibody at room temperature in the dark for 1 hour. The sections were subsequently washed in TBST and incubated with secondary fluorescent antibodies for 2 hours at room temperature. After washing in TBST, a buffer containing DAPI was added, and the slides were washed with TBST and mounted.Enzyme-linked immunosorbent assayLevels of Aβ40 and Aβ42 in soluble mouse brain temporal and hippocampal tissue extracts were measured as described previously, according to the manufacturer’s instructions (Huabio, China).Microglial skeleton analysisSkeletal analysis of microglia was performed as previously described. Briefly, pannoramic MIDI:3Dhistech was employed for scanning stained brain slices. The Skeletonize plugin of ImageJ software (National Institutes of Health, USA) was used for skeleton analysis, and the AnalyzeSkeleton plugin was used for length and process number analysis.Statistical analysisAll data were analyzed using GraphPad Prism 8. Pairwise comparisons were analyzed using a two-tailed Student’s t test, and one-way analysis of variance (ANOVA) was used for multigroup comparisons. Human transcription samples were analyzed using Spearman’s correlation coefficient. All values are expressed as the means ± SEMs. P values < 0.05 were considered significant.