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

doi:10.21203/rs.3.rs-3220263/v1

Download PDF

Research Article

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

https://doi.org/10.21203/rs.3.rs-3220263/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Alzheimer’s disease (AD) is one of the major causative factors that induces progressive dementia, which is characterized by memory loss and progressive neurocognitive dysfunction. To elucidate the molecular mechanism contributing to AD, we analyzed an RNA-seq cohort of temporal cortex in AD patients using a bioinformatics workflow and demonstrated that transforming growth factor beta receptor 3 is a crucial gene regulating AD. Nevertheless, soluble TGFBR3 (sTGFBR3) rather than membrane-bound TGFBR3 is abnormally elevated in AD patients and animals. We then demonstrated that sTGFBR3 knockdown restored spatial learning and memory deficits in APP/PS1 and STZ-induced tau hyperphosphorylation mice. Mechanistically, sTGFBR3 knockdown promoted microglial polarization to the M2 phenotype from the M1 phenotype, inhibited proinflammatory and chemotactic activity, and enhanced phagocytic activity. In addition, sTGFBR3 knockdown significantly reduced acute LPS-induced neuroinflammation and ameliorated STZ-induced neuronal function impairment. These findings suggest that sTGFBR3 is a potential therapeutic target for AD.

Alzheimer’s disease

sTGFBR3

TGFBR3

Microglial polarization

Bioinformatics

AD is the most common type of dementia, characterized by the deposition of senile plaques outside neurons and the accumulation of neurofibrillary tangles inside neurons in the brain [1]. WHO report shows that more than 55 million people now live with dementia, which brings a substantial economic burden to society and families [2]. However, the molecular mechanisms underlying AD pathology remain unclear, which has hampered advances in drug research. Therefore, it is urgent to discover new molecular mechanisms for the development of effective therapeutic strategies against AD.

The aberrant polarization of microglia in AD has been implicated in neuroinflammation, neuronal damage, phagocytosis, and Aβ propagation [3–5] and its intervention is considered a promising way to restore central function [6, 7]. Aberrant polarization of microglia refers to the excessive transformation of the M2 anti-inflammatory phenotype into the M1 proinflammatory phenotype [8]. This transformation promotes the migration of microglia to pathological locations and contributes to the inflammatory response, aggravating the pathology related to AD [9–11].

TGF-β1, an anti-inflammatory cytokine, can polarize microglia to the M2 phenotype, consequently enhancing the uptake and scavenging of Aβ by microglia [12]. However, several studies have reported that TGF-β1 accumulation is implicated in the polarization of microglia with the M1 phenotype [12, 13]. This accumulation in the AD brain demonstrates neurotoxicity and impaired Aβ uptake rather than a protective effect [12, 14]. In addition, it is accompanied by blockade of Smad-dependent TGF-β signaling [15], which is extremely harmful in aging as well as in AD patients [16, 17]. These contradictory discoveries have forced us to re-examine the TGF-β signaling. The accumulated TGF-β1 induces the expression of inhibitory Smads, thus blocking signal transduction [17]. We also found that NF-κB, a member of non-Smad-dependent TGF-β signaling, can be induced to activation by TGF-β1 [18]. It is known that activation of NF-κB can promote M1-type polarization in microglia [19, 20]. However, NF-κB inhibition can reverse microglia to the M2 phenotype [21]. All of these alterations appear to arise from the aberrant accumulation of TGF-β1. However, the molecular mechanisms underlying this abnormal accumulation are not fully understood.

TGFBR3 is a member of the TGF-β signaling pathway. It can be sheared to produce a soluble form (kDa: 90, 120) by MT1-MMP, which is upregulated in the hippocampus of 5 × FAD mice [22, 23]. The primary function of TGFBR3 in TGF-β signaling is to "present" ligands to the type II TGF-β receptor to activate downstream signaling [24, 25]. Instead, sTGFBR3 inhibits TGF-β signaling through ligand binding competition with TGFBR3 [26]. Therefore, we speculate that sTGFBR3 may be elevated in AD and participate in the abnormal accumulation of TGF-β1.

In this study, we analyzed an RNA-seq cohort of the temporal cortex in AD patients using a bioinformatics workflow and identified TGFBR3 as a crucial regulatory gene for AD. We subsequently found that sTGFBR3, but not TGFBR3, was significantly increased in AD mice and patients, In addition, sTGFBR3 knockdown significantly ameliorated AD-related cognitive impairments in APP/PS1 and STZ-induced tau hyperphosphorylation mice. Mechanistically, sTGFBR3 knockdown increased the polarization of M2 microglia, reducing the inflammatory response, Aβ deposition, and tau hyperphosphorylation, and restoring the functions of neurons and synapses. The results suggest that sTGFBR3 is a potential therapeutic target for AD.

2.1 Materials

Antibodies against β-amyloid (#sc-28365, 1:1000 for western blotting, 1:500 for immunofluorescence and immunohistochemistry), TGFBR3 (#sc-74511, 1:1000 for western blotting, 1:500 for immunofluorescence), Iba1 (#sc-32725, 1:1000 for western blotting, 1:500 for immunofluorescence) and sodium ptassium ATPase/ATP1A1 (#sc-514614, 1:1000 for western blotting) were purchased from Santa Cruz Biotechnology (USA). The antibodies of MT1-MMP (#AF1441, 1:1000 for western blotting), APP (#AF6219, 1:1000 for western blotting) and TREM2 (#AF8229, 1:1000 for western blotting) were purchased from Beyotime (China), The antibodies of TGF-β1 (#WL02193, 1:1000 for western blotting), p-SMAD2(S465 + S467)/p-SMAD3(S423 + S425) (#WL02305, 1:1000 for western blotting), SMAD2/SMAD3 (#WL01520, 1:1000 for western blotting), IL-6 (#WL02841, 1:1000 for western blotting), TNF-α (#WL01581, 1:1000 for western blotting), BACE (#WL02795, 1:500 for western blotting), GSK3β (#WL01456, 1:1000 for western blotting), P-GSK3β (S9) (#WL03518, 1:1000 for western blotting), TAU (#WL03184, 1:1000 for western blotting), IL-10 (#WL03088, 1:1000 for western blotting), NLRP3 (#WL02635, 1:1000 for western blotting), PSD95 (#WL05046, 1:1000 for western blotting) and SYP (#WL03058, 1:1000 for western blotting) were purchased from Wanleibio (China). Phospho-Tau (Thr181) (#D9F4G, 1:1000 for western blotting), phospho-Tau (Ser404) (#D2Z4G, 1:1000 for western blotting), phospho-Tau (Ser202/Thr205) (#D4H7E, 1:1000 for western blotting) and NF-κB p65 (#D14E12, 1:1000 for western blotting) were purchased from Cell Signaling Technology (USA). The antibodies of MAP2 (#GB11128-2, 1:1000 for (immunofluorescence), NeuN (#GB11138, 1:1000 for immunofluorescence) and SYT (#GB111130, 1:1000 for western blotting) were purchased from Servicebio (China). The antibody of GAPDH (#ET1601-4, 1:50000 for western blotting) was purchased from HUABIO (China).

2.2 Acquisition and preprocessing of the RNA-seq cohort from AD patients

An 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 (Cases: 108, Normal: 11. Synapse ID: syn3157743). The cohort was divided into 7 groups according to the CDR, and the CDR score ranged from 0 to 5. 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). We first carried out standardized and usability analyses. Then, a P value < 0.05 was invoked as the cutoff criteria to filter differentially expressed genes using limma packages. General information about all samples including CDR, Braak staging, plaque, age, and sex was recorded in the original data.

2.3 Gene Set Enrichment Analysis

All procedures were carried out as previously described [27]. A 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).

2.4 Weighted gene coexpression network analysis

All procedures were performed as previously described [28]. Briefly, we first preprocessed the RNA-seq cohort of differential genes to remove outliers. Then, the optimal soft threshold for the remaining samples was calculated and seven was chosen to construct a network. According to TOM dissimilarity, the genes were clustered to generate a cluster dendrogram, and the modules with high dissimilarity were merged. Finally, interesting modules were extracted according to the correlation between pathology and modules. Genes with P < 0.01 were collected for further analysis.

2.5 Protein-Protein Interaction Analysis

PPI network was generated using STRING v11.5 [29] and visualized by Cytoscape v3.7.1 [30]. Hub genes were obtained using two strategies: 1, the top 200 genes ranked by MCC (CytoHubba) were collected using the “Check the First-Stage Genes” option. 2, the top 200 genes ranked by degree, calculated by CentiScaPe 2.2, were collected. We collected 180 shared genes by CentiScape and CytoHubba.

2.6 Gene Functional and Pathway Enrichment Analysis

Biological functions and KEGG pathways for differentially expressed genes were analyzed through the online server Metascape[31]. Terms with P < 0.01 were deemed to be statistically significant.

2.7 Convergent functional genomic ranks

CFG ranks as a Bayesian way of cross-validating findings and reducing uncertainty, which integrates several parameters to evaluate the correlation between genes and AD [32]. The correlation between genes and AD was evaluated by several parameters of CFG, including eQTL, GWAS, PPI, early gene expression regulated by AD genetic variants, and Aβ and tau pathology. 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. CFG ranks were generated using Alzdata [33].

2.8 Human brain tissue

Human brain tissues (temporal cortex) from subjects with AD 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: 1:8) containing protease and phosphatase cocktail inhibitors (Sigma). For RNA extraction, an RNA Extraction Kit (Qiagen) was utilized to homogenize frozen tissues in TRIzol. Information on human brain samples of normal controls and patients with AD is contained in Table 1.

Table 1

Details of AD patients’ brains and healthy brain samples from the China Brain Bank Center
Numbering	Age	Sex	Date of death	Groups
20131012	91	F	2013/10/12	AD
20160106	86	M	2016/1/6	AD
20180512	95	F	2018/5/12	AD
20190502	67	M	2019/5/2	Normal
20190220	68	M	2019/2/20	Normal
20201129	69	M	2020/11/29	Normal

2.9 Mice and cell culture

Male APP/PS1 mice (2 months old, 20 ± 1 mg) [34] and age-matched wild-type C57BL/6 male 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. Study results may be influenced by sex, so we only used animals of a single sex for the study [35]. 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 1% antibiotic–antimycotic mixture kept at 37°C with 5% CO2.

2.10 RT-qPCR analysis

For 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′.

2.11 ICV

The mice were anesthetized with isoflurane (3% isoflurane for induction, 2% isoflurane for maintenance). With the bregma as the origin of coordinates, the mouse was positioned in the lateral ventricle, according to the coordinates: bregma: -0.5 mm, interaural line: -1.1 mm, cranium: -3.0 mm, injection speed: 1 µL/min, at 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 penicillin was injected intramuscularly to prevent infection.

2.12 Immunoblotting

The temporal cortex of WT and APP/PS1 mice was dissected and obtained. After extracting total protein, the BCA method was used to evaluate the protein concentration. The protein samples were isolated by 10% SDS polyacrylamide gel electrophoresis and then transferred to PVDF membranes. The PVDF membrane was sealed with a 5% skim milk/5% BSA 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, and TBST was rewashed for 10 min×3 times. Finally, a chemiluminescence imaging system (Bio-Rad, ChemiDoc) was used for blotting exposure, and ImageJ was employed for data analysis.

2.13 Immunofluorescence

Collected human 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 binding to the secondary antibody at room temperature in the dark for 1 hour. The utilized antibodies included anti-amyloid protein (Cell Signaling Technology, D3D2N) and TGFBR3 antibody (Santa Cruz, sc-74511). 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.

2.14 Enzyme-linked immunosorbent assay

Levels of Aβ40 and Aβ42 in soluble mouse brain temporal and hippocampal tissue extracts were measured as described before [36], according to the manufacturer’s instructions (Huabio, Wuhan, China).

2.15 Morris water maze

Learning and memory were measured using the MWM test previously reported but with some modifications.37 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.

2.16 Acute LPS injection

LPS (0.5 mg/kg; Sigma) was administered by intraperitoneal injection according to a previously described procedure [36]. Briefly, the mice were treated with 0.5 mg/ml 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.

2.17 Microglial skeleton analysis

The skeleton analysis of microglia was performed as previously described [37]. 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.

2.18 Acute STZ injection

STZ (50 mg; Sigma) was administered to the lateral ventricle according to a previously described procedure [38]. 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.

3 Statistical analysis

All 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 ± SEM. P values < 0.05 were considered significant.

4.1 A bioinformatics workflow identified that TGFBR3 plays a crucial role in AD pathology

To systematically identify the molecular mechanisms underlying AD, a bioinformatic workflow was established for the discovery of crucial regulatory targets in AD (Fig. 1). After assessing the data quality, we analyzed the hallmarks for the RNA-seq cohort (ID: syn3157743) using GSEA (Supplementary Fig. 1). The results exhibited a significant enrichment of neuroinflammation-related hallmarks, including TNF-α signaling via NF-κB, INF-α response, TGF-β signaling, and IL6-JAK-STAT3 signaling (Fig. 2A).

To demonstrate the contribution of neuroinflammation to AD, we obtained 794 differentially expressed genes from the cohort. Then, several methods, including WGCNA, PPI, and biological process analysis, were combined to identify a group of crucial genes that connect neuroinflammation with typical lesions (CDR, Braak staging, and Aβ plaque). In WGCNA, the MEgray module was positively correlated with all three typical lesions of AD, and 379 genes associated with this module were collected (P < 0.01) using Pearson's correlation

analysis. (Fig. 2B and 2C). In PPI analysis, a total of 180 shared genes were collected from the top 200 differential genes calculated by CytoHubba and CentiScaPe 2.2 (Fig. 2D). In biological process analysis, neuroinflammation-related processes were identified, and 62 hit genes were integrated from these processes (Fig. 2E). Finally, we identified 18 shared genes from the three methods that connected neuroinflammation and typical lesions, which together accelerated the progression of AD (Fig. 2F).

We next ranked the 18 shared genes using the CFG method and identified 6 genes (TGFBR3, VIM, AGT, CD44, PLCE1, FN1) with a CFG score ≥ 3. (Table 2). The expression of 6 genes was then validated using data from multiple platforms, which revealed that only TGFBR3 was elevated across several brain regions (Supplementary Fig. 1). Notably, the KEGG enrichment analysis ranked TGF-β signaling as the top pathway (Fig. 2E), suggesting that TGFBR3 may be implicated in neuroinflammation-related AD lesions by mediating TGF-β signaling. Then, we conducted further investigation and discovered that TGFBR3 exhibited a positive correlation with CDR, plaque, and Braak staging (Fig. 2G-2I). Intriguingly, we also observed that TGFBR3 is highly expressed in microglia using single-cell sequencing analysis, indicating that microglia are implicated in the aforementioned process (Supplementary Fig. 2).

Table 2 The CFG rank of candidate genes in the AlzData online server

GENE	eQTL^a	GWAS	PPI^b	Early_DEG^c	Pathology cor (Aβ)^d	Pathology cor (tau)^e	CFG^f
*TGFBR3*	1	25		yes	0.527, ***	0.399, ns	4
*VIM*	3	0	APP, MAPT	yes	0.918, ***	0.700, **	4
*AGT*	1	0	APP, PSEN1, APOE	yes	-0.359, *	0.002, ns	4
*CD44*	2	0	APP, PSEN1		0.719, ***	0.793, ***	3
*PLCE1*	1	0		yes	0.683, ***	0.562, *	3
*FN1*	1	2	APP, MAPT, APOE				3
*NPPA*	4	0	APP, APOE		-0.006, ns	0.105, ns	2
*HFE*	1	0			0.786, ***	0.852, ***	2
*ARHGEF6*	0	NA	APOE		0.471, **	0.611, *	2
*CD4*	1	0			0.063, ns	-0.606, *	2
*S100B*		0	MAPT		0.406, **	0.313, ns	2
*LTF*	3	0	APP, APOE		-0.148, ns	0.459, ns	2
*WLS*	2	1					2
*TFR2*	1	0					1
*PTPN3*	1	0			-0.064, ns	-0.044, ns	1
*IFIH1*		0					0
*TNFAIP3*	0	0			0.025, ns	0.129, ns	0

^aeQTL: expression of target gene is regulated by AD genetic variants

^bPPI: target gene has significant physical interaction with APP, PSEN1, PSEN2, APOE or MAPT

^cEarly_DEG: target gene is differentially expressed in AD mouse models before AD pathology emergence

^dPathology cor (Aβ): correlation of target gene expression with AD pathology in Aβ line AD mouse models

^ePathology cor (tau): correlation of target gene expression with AD pathology in tau line AD mouse models

^fCFG: total CFG score of the target gene, 1 CFG point is added if any of the above evidence is significant, and the CFG point ranges from 0 to 5.

*P < 0.05, **P < 0.01, ***P< 0.001.

4.2 Increased sTGFBR3 levels are associated with AD

To further investigate the relationship between TGFBR3 expression and AD, we first stained brain tissue from wild-type mice of different ages using a monoclonal anti-mouse TGFBR3 antibody. This antibody was raised against amino acids 511-790 mapping near the C-terminus of the TGFBR3 extracellular domain. Staining with this TGFBR3 antibody was significantly increased with age (Fig. 3A and 3B). However, immunoblotting showed that only sTGFBR3 was increased with age (Fig. 3C and 3D). In addition, MT1-MMP and TGF-β1 were also increased with age, but P-Smad2/3 was decreased, indicating the abnormal accumulation of TGF-β1 and the blockade of TGF-β/Smad signaling (Fig. 3C, 3E-3F, Supplementary Fig. 4A-4B). We also found that sTGFBR3 and MT1-MMP had the same increasing trend in APP/PS1 mice expressing a chimeric mouse/human amyloid precursor protein (Mo/HuAPP695swe) and a mutant human presenilin 1 (PS1-dE9) by immunoblotting analysis (Fig. 3G-3I), indicating that the production of sTGFBR3 was sheared by MT1-MMP. Interestingly, the expression of full-length TGFBR3 (kDa: 180) was also increased in two samples (Fig. 3G). However, the biological function of sTGFBR3 still cannot be underestimated. Furthermore, we performed costaining for sTGFBR3 and Iba1 in 6-month-old APP/PS1 mice, suggesting that the elevated sTGFBR3 was almost entirely derived from microglia (Fig. 3M-3 N).

We then further examined postmortem temporal brain tissue from 3 patients with AD by RT‒qPCR, immunoblotting and immunostaining. First, RT‒qPCR and immunoblotting showed that TGFBR3 was significantly elevated in the temporal lobe compared with normal controls, but mainly sTGFBR3 was upregulated (Fig. 3J-3 L). Likewise, abnormal accumulation of TGF-β1 also existed in the brains of AD patients (Fig. 3J). Next, we performed costaining for human sTGFBR3 and Aβ, and the results suggested that sTGFBR3 was spatially correlated with Aβ plaques (Supplementary Fig. 3C). These data provide direct evidence that increased sTGFBR3 levels are associated with AD.

4.3 MT1-MMP inhibitor restored the impaired TGF-β/Smad signaling in LPS-induced BV2 cells

To verify whether sTGFBR3 was involved in blocking TGF-β/Smad signaling. We treated LPS-induced BV2 cells with marimastat, an inhibitor of MT1-MMP. Immunoblotting analysis showed that sTGFBR3 in the extracellular matrix was significantly increased, and P-Smad2/3 in the cytoplasm was decreased, suggesting the blockade of TGF-β/Smad signaling in LPS-induced BV2 cells (Fig. 4A-4G). However, these abnormalities were reversed in the coculture system of marimastat and LPS. The results showed that sTGFBR3 in the extracellular matrix was significantly decreased, and P-Smad2/3 in the cytoplasm was increased (Fig. 4A-4G). We also observed that the proinflammatory factors IL-6 and TNF-α were increased in LPS-induced BV2 cells, suggesting the formation of M1 proinflammatory microglia, which was also reversed by marimastat (Fig. 4A, 4H-4I), indicating that lowering sTGFBR3 could decrease inflammatory factors and promote microglial M2 polarization.

4.4 sTGFBR3 is associated with cognitive decline in APP/PS1 transgenic mice

Having established that sTGFBR3 levels are strongly associated with AD, we then assessed the relevance of our findings for dementia. To this end, the sTGFBR3 knockdown model was established using a lentivirus to interfere with the expression of sTGFBR3 in the brains of APP/PS1 mice. Notably, sTGFBR3 expression was significantly increased in APP/PS1 mice compared with wild-type controls (Fig. 5A-5B). In contrast, sTGFBR3 expression was decreased in APP/PS1 mice with lentivirus intervention, demonstrating efficient viral transduction and sTGFBR3 knockdown (Fig. 5A-5B). Next, we found a significant association of higher sTGFBR3 expression with increased measures of cognitive decline by the MWM task. APP/PS1 mice spent more time reaching the target platform during the training period and displayed severe loss of spatial learning and memory compared with wild-type controls. In contrast, sTGFBR3 knockdown significantly improved task learning (Fig. 5C). Furthermore, during the test period, sTGFBR3 knockdown resulted in a significantly decreased time to reach the target platform and an increased number of platform crossings but no significant alteration in swimming speed, which suggested alleviation of cognitive impairment (Fig. 5D-5G).

We next determined whether sTGFBR3 knockdown affected Aβ deposition in APP/PS1 mice. As shown in Fig. 5H-5I, the number of Aβ plaques was markedly decreased in lentivirus-treated APP/PS1 mice compared to shRNA ctrl-treated APP/PS1 mice, indicating that sTGFBR3 knockdown could alleviate the Aβ plaque burden in the cortex and hippocampus of APP/PS1 transgenic mice. Consistently, ELISA analysis revealed a significant reduction in both soluble Aβ40 and Aβ42 in the temporal and hippocampal extracts (Fig. 5J-5 M), suggesting that sTGFBR3 plays a crucial role in promoting Aβ deposition in APP/PS1 mice. Furthermore, to clarify whether the decline in Aβ content was due to the decrease in Aβ production or the increase in Aβ clearance, we quantified the protein expression of full-length APP and BACE1. Immunoblotting analysis showed that the critical regulators APP and BACE1 were significantly decreased in lentivirus-treated APP/PS1 mice (Fig. 5N-5P), suggesting that sTGFBR3 knockdown influenced APP metabolism. Here, it is unclear whether sTGFBR3 knockdown affects the clearance of Aβ.

4.5 The effect sTGFBR3 knockdown on microglial function in APP/PS1 mice

Abnormal phagocytosis of microglia mediates the propagation of Aβ within the brain in AD [39]. We therefore examined the effect of sTGFBR3 knockdown on microglial chemotaxis and phagocytosis. As shown in Fig. 6A-6B and S4A-S4B, triple staining for Aβ, Iba1 and TREM2 was used to characterize the chemotaxis and phagocytosis of microglia for Aβ plaques in APP/PS1 mice. The results showed that sTGFBR3 knockdown significantly decreased the number of microglia around plaques. Meanwhile, immunoblotting analysis showed that Iba1, CCR2 and MCP1 were significantly decreased in lentivirus-treated APP/PS1 mice (Fig. S4C, S4E, Supplementary Fig. 4F-4H), indicating that the chemotaxis of microglia to Aβ plaques was weakened.

Moreover, TREM2 mediates abnormal Aβ phagocytosis by microglia [40]. TREM2 staining indicated that phagocytosis from microglia to Aβ plaques was enhanced compared to a lower number of microglia (Fig. 6A-6B). Immunoblotting analysis showed similar results (Fig. S4C-S4D). In addition, we found that sTGFBR3 knockdown significantly decreased the levels of the proinflammatory factors IL-6 and TNF-α but increased the levels of the anti-inflammatory factors IL-4 and IL-10 (Supplementary Fig. 5A-5G). Moreover, sTGFBR3 knockdown significantly decreased NLRP3 and NF-κB p65 (Supplementary Fig. 5H-5J), indicating that microglia polarized into the M2 phenotype from the M1 phenotype. These results suggested that the polarization of the microglia M2 phenotype restores phagocytosis but reduces diffusion of the Aβ protein.

TGF-β/Smad signaling is involved in the pathological process of AD, including Aβ uptake, neuroinflammation and chemotaxis of microglia.[41,42] Immunoblotting analysis showed that P-Smad2/3 was significantly increased in lentivirus-treated APP/PS1 mice (Fig. 6D-6E), suggesting that TGF-β1/Smad3 signaling was enhanced and Aβ uptake in the AD brain was increased. Finally, Griess Reagent analysis showed that NO release was significantly decreased in lentivirus-treated APP/PS1 mice compared with an increased level in APP/PS1 mice (Fig. 6F).

Overall, the knockdown of sTGFBR3 improved TGF-β1/Smad3 signaling and weakened non-Smad-dependent NF-κB signaling. These changes polarized microglia into the M2 phenotype and enhanced the phagocytosis of microglia while limiting their chemotaxis.

4.6 sTGFBR3 knockdown inhibits LPS-induced neuroinflammation in vivo

The above studies have implicated sTGFBR3 in microglial neuroinflammation. To further investigate the role of sTGFBR3 in neuroinflammation, we established an acute LPS injection mouse model as previously described [36]. lentivirus was used to interfere with the expression of TGFBR3 in C57BL/6 mice. Four weeks later, an acute neuroinflammation model was induced by intraperitoneal injection of LPS. We then tested sTGFBR3 protein expression in the brain. sTGFBR3 expression was significantly increased in LPS-induced mice compared with wild-type controls. In contrast, sTGFBR3 expression was decreased in lentivirus-treated LPS-induced mice, demonstrating efficient viral transduction and sTGFBR3 knockdown (Fig. 6G). Next, we found that intraperitoneal injection of LPS significantly increased the levels of certain proinflammatory cytokines, including TNF-α and IL-6. Notably, sTGFBR3 knockdown markedly inhibited the levels of these cytokines (Fig. 6H-6I), suggesting a critical role for sTGFBR3 in mediating neuroinflammation.

Moreover, the changes in microglial morphology were analyzed by Iba1 staining. LPS injections induced significant changes in microglial morphology, characterized by decreased branch numbers, end-point voxels, max branch length, and a significant increase in total branch length. sTGFBR3 knockdown significantly rescued LPS-induced morphological changes (Fig. 6K-6 N).

4.7 sTGFBR3 is associated with cognitive decline in the STZ-induced tau hyperphosphorylation mouse model

To investigate the role of sTGFBR3 in tau hyperphosphorylation, an STZ-induced tau hyperphosphorylation mouse model was established as described earlier [38]. The expression of sTGFBR3 in the C57BL/6 mouse brain was interfered with by the above method. Four weeks later, a tau hyperphosphorylation mouse model was induced by intraperitoneal injection of STZ. Except for the effective sTGFBR3 knockdown in lentivirus-treated STZ-induced mice, we also observed that TGF-β1/Smad3 signaling was significantly enhanced compared with the decreased density of P-Smad2/3 in STZ-induced mice (Fig. 7A-7B). We then checked whether sTGFBR3 knockdown affected learning and memory impairment using the MWM task. STZ-induced mice spent more time reaching the platform in the training period, suggesting the presence of a cognitive deficit. Notably, sTGFBR3 knockdown significantly improved task learning (Fig. 7C). During the test period, sTGFBR3 knockdown resulted in a significantly decreased time to reach the platform, increased preference for the target quadrant, and increased number of platform crossings. However, no significant alteration in swimming speed suggested alleviation of cognitive impairment (Fig. 7D-7F).

Mechanistically, immunoblotting analysis showed that P-GSK3β (Ser9) was significantly increased and P-tau (Thr181, Ser202+Thr205, and Ser404) was significantly decreased in lentivirus-treated STZ-induced mice compared with the reverse changes in STZ-induced mice. This result suggests that tau hyperphosphorylation was reduced (Fig. 7G-7I). The structural integrity of neurons is the basis of learning and memory, and the destruction of neuronal structures often presupposes damage to cognitive function [43]. We found that the neurons in the cortex and hippocampus suffered from impairment in STZ-induced mice by Nissl staining, characterized by atrophy, deep staining, and a decreased number of neurons. However, sTGFBR3 knockdown rescued the impairment to neurons and restored the number of neurons (Supplementary Fig. 6A-6B). Immunofluorescence analysis revealed that the function of neurons was obviously improved by staining NeuN and MAP2 in lentivirus-treated STZ-induced mouse brains (Fig. 7J-7 L). Meanwhile, immunoblotting analysis showed that the density of synapse-related proteins, including PSD95, SYP, and SYT, was significantly increased in lentivirus-treated STZ-induced mouse brains (Supplementary Fig. 6C-6F), suggesting that sTGFBR3 knockdown restored synaptic plasticity.

Overall, sTGFBR3 knockdown reduced tau hyperphosphorylation, restored neuronal function, and maintained synaptic plasticity by restoring TGF-β1/Smad3 signaling in STZ-induced tau hyperphosphorylation.

Microglia mediate crucial functions to support the central nervous system and play critical roles in the pathogenesis of AD. Initial microglial investigations found that changes in abnormal microglial overactivation are largely detrimental, positioning microglia at the center of neuroinflammation, neurodegeneration, and synaptic loss [3, 4]. The inhibition of microglial activity has become a powerful strategy for treating AD. However, activated microglia exhibited M1 and M2 phenotypes separately [7]. The M1 phenotype is harmful, not the M2 phenotype. Thus, it is not advisable to broadly inhibit microglial activity. The M1 phenotype enhances the neuroinflammatory response by secreting proinflammatory cytokines; in contrast, the M2 phenotype exerts neuroprotective effects by inhibiting neuroinflammation [44, 45]. The two phenotypes can be transformed and balanced under various conditions. However, the balance in AD tends to favor the M1 phenotype. This phenotype promotes microglial migration to pathological locations and participation in the inflammatory response, worsening AD-related pathology [9–11]. Therefore, inhibiting microglial polarization to the M1 phenotype or promoting microglial polarization to the M2 phenotype has become a potential therapeutic strategy for AD.

In this study, we identified a gene, TGFBR3, that serves as an intersection to connect neuroinflammation with Aβ and tau pathology in a bioinformatics workflow. We then demonstrated that sTGFBR3, but not TGFBR3, was elevated in the temporal cortex of an AD mouse model and human AD patients. This elevation can cause abnormal accumulation of TGF-β1 and block TGF-β1/Smad3 signaling to produce neurotoxicity. In addition, non-Smad-dependent NF-κB signaling is activated to produce neuroinflammation. These results polarize microglia to the M1 phenotype, producing abnormal phagocytosis and propagation of Aβ and tau. However, these changes were reversed after treatment with sTGFBR3 knockdown lentivirus. Additionally, sTGFBR3 knockdown restored neuronal function and synaptic plasticity, which enhanced cognitive function. Interestingly, we also observed that microglia experienced the same changes in the brains of aging mice. These results indicated that a key control point for the transition from aging to AD may be altered microglial polarization, which is regulated by sTGFBR3-mediated blockade of TGF-β1/Smad3 signaling and enhancement of non-Smad-dependent NF-κB signaling. Therefore, it is crucial to intervene pharmacologically on sTGFBR3.

sTGFBR3 was composed of core protein and covalently attached glycosaminoglycan (GAG) chain modifications that included heparan sulfate (HS) and chondroitin sulfate (CS). Several studies have demonstrated the implication of HS in the pathology of AD. HS-Aβ interactions are associated with almost all stages of AD pathogenesis, including Aβ production, clearance, accumulation, aggregation, and toxicity [46–48]. In addition, several studies have revealed that HS-tau interactions enhance tau uptake by microglia [49, 50], which can promote the propagation of tau in the brain. HS has also been found to aggravate neuroinflammation [51, 52]. However, sTGFBR3 acts as a part-time proteoglycan that may not carry sporadic HS in a different context than other full-time proteoglycans. The function of HS could be negligible in relation to the function of the core protein. Therefore, we did not consider the interference potentially caused by HS in this study. The core protein is involved in the presentation of three TGF-β ligands (TGF-β1, 2, 3) for type II TGF-β receptors [53]. Increased sTGFBR3 contributes to the accumulation of inefficient excessive levels of extracellular TGF-βs, leading to brain toxicity [54–56]. The TGF-β family was found to be significantly associated with AD through analysis of the RNA-seq cohort (Supplementary Fig. 7). Notably, TGF-β2 exhibits a higher correlation with AD, which can be attributed to the higher affinity of TGF-β2 for sTGFBR3 [57]. Additionally, a correlation between TGFBR1 and AD was found to be inconsistent with previous studies. Elevated TGF-βs induce Smad7 to recruit the E3 ubiquitin protein ligases Smurf1 and Smurf to ubiquitinate TGFBR1 for degradation [58, 59]. Similarly, this inconsistency also occurred with TGFBR2. All of these differences may be attributed to differences in transcriptome and protein levels.

There are also some puzzling aspects of this study. The decrease in Aβ can be attributed to either increased clearance or decreased production. In this study, we focused on Aβ clearance and therefore did not further investigate the reasons for the decrease in Aβ production. In AD, increased processing of APP by β and γ secretases leads to the production of high levels of Aβ, which can aggregate into extracellular plaques [60]. We found that APP expression was significantly reduced in the APP/PS1 mouse brain after lentivirus treatment. However, it was shown that HS can block the shearing of APP by β-secretase by binding directly to β-secretase, APP [61]. Thus, sTGFBR3-attached HS did not exert a role in blocking Aβ production, which is consistent with our findings discussed above. We hypothesize that lentivirus treatment blocks inflammatory signaling and that nuclear translocation of NF-κB may be reduced, which may interfere with BACE1 transcription and result in reduced BACE1 expression [11]. Moreover, MT1-MMP is highly expressed in AD and can produce Aβ by shearing APP [62]. However, the type of Aβ produced remains unclear.

Overall, in this study, we demonstrate that at the intersection of AD pathology, sTGFBR3, as a key regulatory receptor for microglial polarization dysregulation, is a potential therapeutic target for AD. Antagonist design targeting the sTGFBR3 endogenous ligand binding site is a promising therapeutic approach: (1) ligands competitively bound by sTGFBR3 can be released to enhance TGF-β signaling; (2) the reduction of aberrant accumulation of ligands inhibits non-Smad-dependent NF-κB signaling.

Aβ = Amyloid-β; APP = Amyloid Beta Precursor Protein; BACE1 = β-Secretase; CCR2 = Chemokine Receptor 2; CDR = Clinical Dementia Rating; CFG = Convergent Functional Genomic; ELISA = Enzyme-Linked Immunosorbent Assay; eQTL = Expressional Quantitative Trait Loci; GSEA = Gene Set Enrichment Analysis; GWAS = Genome-Wide Association Study; Iba1 = Ionized Calcium Binding Adaptor Molecule 1; ICV = Intra-Cerebroventricular Injection; KEGG = Kyoto Encyclopedia Of Genes And Genomes; LPS = Lipopolysaccharides; MAP2 = Microtubule-Associated Protein; MCP1 = Monocyte Chemotactic Protein 1; MT1-MMP = Membrane-Type 1 Matrix Metalloproteinase; MWM = Morris Water Maze; NF-κB = Nuclear Factor Kappa-B; NLRP3 = NOD-Like Receptor Thermal Protein Domain Associated Protein 3; PPI = Protein-Protein Interaction; PSD95 = Postsynaptic Density-95; STZ = Streptozotocin; SYP = Synaptophysin; SYT = Synaptotagmin; Transforming growth factor beta receptor 3 = TGFBR3; TREM2 = Triggering Receptor Expressed On Myeloid Cells-2; WGCNA = Weighted Gene Co-Expression Network Analysis; WHO = World Health Organization

Acknowledgments

We are grateful to Prof. Jiapei Dai (Chinese Brain Bank Center, Central South University for Nationalities, Wuhan, China) for his help in the experiment and analysis of human subject samples. This work was supported by the National Natural Science Foundation of China.

Funding

This work was supported by the National Natural Science Foundation of China (Grant numbers [81973209], [82173716] and [82104154]).

Author Contributions

Prof. Q.C.Z. and Dr. X.W.J. designed this study. Z.L.J. performed this research and wrote the paper. N.W., W.Z.F., and X.L. assisted in isolating animal tissue. Z.H.X., Q.W., and J.X.C. collected data. X.M.J., X.Y.N., and Z.T.Q. analyzed data. All authors read and approved the final manuscript.

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of General Hospital of Northern Theater Command (Date: 2022.08.01/No: Y (2022) 102).

Consent for publication

Not applicable.

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Alzheimer's disease facts and Figures. Alzheimers Dement. Mar 14 2022;doi:10.1002/alz.12638
World health statistics, World failing to address dementia challenge. 2021. 2021.
Sierra A, Paolicelli RC, Kettenmann H. Cien Años de Microglía: Milestones in a Century of Microglial Research. Trends Neurosci. Nov 2019;42(11):778-792. doi:10.1016/j.tins.2019.09.004
Tay TL, Béchade C, D'Andrea I, et al. Microglia Gone Rogue: Impacts on Psychiatric Disorders across the Lifespan. Front Mol Neurosci. 2017;10:421. doi:10.3389/fnmol.2017.00421
d'Errico P, Ziegler-Waldkirch S, Aires V, et al. Microglia contribute to the propagation of Aβ into unaffected brain tissue. Nat Neurosci. Jan 2022;25(1):20-25. doi:10.1038/s41593-021-00951-0
Hu X, Leak RK, Shi Y, et al. Microglial and macrophage polarization—new prospects for brain repair. Nat Rev Neurol. Jan 2015;11(1):56-64. doi:10.1038/nrneurol.2014.207
Wang Q, Yao H, Liu W, et al. Microglia Polarization in Alzheimer's Disease: Mechanisms and a Potential Therapeutic Target. Front Aging Neurosci. 2021;13:772717. doi:10.3389/fnagi.2021.772717
Nimmerjahn A, Kirchhoff F, Helmchen F. Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo. Science. May 27 2005;308(5726):1314-8. doi:10.1126/science.1110647
Glass CK, Saijo K, Winner B, Marchetto MC, Gage FH. Mechanisms underlying inflammation in neurodegeneration. Cell. Mar 19 2010;140(6):918-34. doi:10.1016/j.cell.2010.02.016
Ising C, Venegas C, Zhang S, et al. NLRP3 inflammasome activation drives tau pathology. Nature. Nov 2019;575(7784):669-673. doi:10.1038/s41586-019-1769-z
Leng F, Edison P. Neuroinflammation and microglial activation in Alzheimer disease: where do we go from here? Nat Rev Neurol. Mar 2021;17(3):157-172. doi:10.1038/s41582-020-00435-y
Tichauer JE, von Bernhardi R. Transforming growth factor-β stimulates β amyloid uptake by microglia through Smad3-dependent mechanisms. J Neurosci Res. Oct 2012;90(10):1970-80. doi:10.1002/jnr.23082
Kiyota T, Okuyama S, Swan RJ, Jacobsen MT, Gendelman HE, Ikezu T. CNS expression of anti-inflammatory cytokine interleukin-4 attenuates Alzheimer's disease-like pathogenesis in APP+PS1 bigenic mice. Faseb j. Aug 2010;24(8):3093-102. doi:10.1096/fj.10-155317
Uribe-San Martín R, Herrera-Molina R, Olavarría L, Ramírez G, von Bernhardi R. Reduction of beta-amyloid-induced neurotoxicity on hippocampal cell cultures by moderate acidosis is mediated by transforming growth factor beta. Neuroscience. Feb 18 2009;158(4):1338-47. doi:10.1016/j.neuroscience.2008.11.002
López-Rovira T, Chalaux E, Rosa JL, Bartrons R, Ventura F. Interaction and functional cooperation of NF-kappa B with Smads. Transcriptional regulation of the junB promoter. J Biol Chem. Sep 15 2000;275(37):28937-46. doi:10.1074/jbc.M909923199
Tichauer JE, Flores B, Soler B, Eugenín-von Bernhardi L, Ramírez G, von Bernhardi R. Age-dependent changes on TGFβ1 Smad3 pathway modify the pattern of microglial cell activation. Brain Behav Immun. Mar 2014;37:187-96. doi:10.1016/j.bbi.2013.12.018
Tominaga K, Suzuki HI. TGF-β Signaling in Cellular Senescence and Aging-Related Pathology. Int J Mol Sci. Oct 10 2019;20(20)doi:10.3390/ijms20205002
Yeh YY, Chiao CC, Kuo WY, et al. TGF-beta1 increases motility and alphavbeta3 integrin up-regulation via PI3K, Akt and NF-kappaB-dependent pathway in human chondrosarcoma cells. Biochem Pharmacol. Mar 15 2008;75(6):1292-301. doi:10.1016/j.bcp.2007.11.017
Ślusarczyk J, Trojan E, Głombik K, et al. Targeting the NLRP3 Inflammasome-Related Pathways via Tianeptine Treatment-Suppressed Microglia Polarization to the M1 Phenotype in Lipopolysaccharide-Stimulated Cultures. Int J Mol Sci. Jul 5 2018;19(7)doi:10.3390/ijms19071965
Wang H, Liu C, Han M, Cheng C, Zhang D. TRAM1 Promotes Microglia M1 Polarization. J Mol Neurosci. Feb 2016;58(2):287-96. doi:10.1007/s12031-015-0678-3
Zhang H, Li Y, Yu J, et al. Rho kinase inhibitor fasudil regulates microglia polarization and function. Neuroimmunomodulation. 2013;20(6):313-22. doi:10.1159/000351221
Py NA, Bonnet AE, Bernard A, et al. Differential spatio-temporal regulation of MMPs in the 5xFAD mouse model of Alzheimer's disease: evidence for a pro-amyloidogenic role of MT1-MMP. Front Aging Neurosci. 2014;6:247. doi:10.3389/fnagi.2014.00247
Velasco-Loyden G, Arribas J, López-Casillas F. The shedding of betaglycan is regulated by pervanadate and mediated by membrane type matrix metalloprotease-1. J Biol Chem. Feb 27 2004;279(9):7721-33. doi:10.1074/jbc.M306499200
Villarreal MM, Kim SK, Barron L, et al. Correction to Binding Properties of the Transforming Growth Factor-β Coreceptor Betaglycan: Proposed Mechanism for Potentiation of Receptor Complex Assembly and Signaling. Biochemistry. Jul 18 2017;56(28):3689. doi:10.1021/acs.biochem.7b00566
Knelson EH, Nee JC, Blobe GC. Heparan sulfate signaling in cancer. Trends Biochem Sci. Jun 2014;39(6):277-88. doi:10.1016/j.tibs.2014.03.001
Zhang M, Zola H, Read L, Penttila I. Identification of soluble transforming growth factor-beta receptor III (sTbetaIII) in rat milk. Immunol Cell Biol. Jun 2001;79(3):291-7. doi:10.1046/j.1440-1711.2001.01013.x
Mootha VK, Lindgren CM, Eriksson KF, et al. PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. Jul 2003;34(3):267-73. doi:10.1038/ng1180
Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. Dec 29 2008;9:559. doi:10.1186/1471-2105-9-559
Szklarczyk D, Gable AL, Nastou KC, et al. The STRING database in 2021: customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res. Jan 8 2021;49(D1):D605-d612. doi:10.1093/nar/gkaa1074
Shannon P, Markiel A, Ozier O, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. Nov 2003;13(11):2498-504. doi:10.1101/gr.1239303
Zhou Y, Zhou B, Pache L, et al. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat Commun. Apr 3 2019;10(1):1523. doi:10.1038/s41467-019-09234-6
Bertsch B, Ogden CA, Sidhu K, Le-Niculescu H, Kuczenski R, Niculescu AB. Convergent functional genomics: a Bayesian candidate gene identification approach for complex disorders. Methods. Nov 2005;37(3):274-9. doi:10.1016/j.ymeth.2005.03.012
Xu M, Zhang DF, Luo R, et al. A systematic integrated analysis of brain expression profiles reveals YAP1 and other prioritized hub genes as important upstream regulators in Alzheimer's disease. Alzheimers Dement. Feb 2018;14(2):215-229. doi:10.1016/j.jalz.2017.08.012
Radde R, Bolmont T, Kaeser SA, et al. Abeta42-driven cerebral amyloidosis in transgenic mice reveals early and robust pathology. EMBO Rep. Sep 2006;7(9):940-6. doi:10.1038/sj.embor.7400784
Xiong J, Kang SS, Wang Z, et al. FSH blockade improves cognition in mice with Alzheimer’s disease. Nature. 2022/03/01 2022;603(7901):470-476. doi:10.1038/s41586-022-04463-0
Wendeln AC, Degenhardt K, Kaurani L, et al. Innate immune memory in the brain shapes neurological disease hallmarks. Nature. Apr 2018;556(7701):332-338. doi:10.1038/s41586-018-0023-4
Li S, Liao Y, Dong Y, et al. Microglial deletion and inhibition alleviate behavior of post-traumatic stress disorder in mice. J Neuroinflammation. Jan 5 2021;18(1):7. doi:10.1186/s12974-020-02069-9
Chen Y, Liang Z, Blanchard J, et al. A non-transgenic mouse model (icv-STZ mouse) of Alzheimer's disease: similarities to and differences from the transgenic model (3xTg-AD mouse). Mol Neurobiol. Apr 2013;47(2):711-25. doi:10.1007/s12035-012-8375-5
McQuade A, Kang YJ, Hasselmann J, et al. Gene expression and functional deficits underlie TREM2-knockout microglia responses in human models of Alzheimer's disease. Nat Commun. Oct 23 2020;11(1):5370. doi:10.1038/s41467-020-19227-5
Chen Y, Colonna M. Two-faced behavior of microglia in Alzheimer's disease. Nat Neurosci. Jan 2022;25(1):3-4. doi:10.1038/s41593-021-00963-w
Estrada LD, Oliveira-Cruz L, Cabrera D. Transforming Growth Factor Beta Type I Role in Neurodegeneration: Implications for Alzheimer´s Disease. Curr Protein Pept Sci. 2018;19(12):1180-1188. doi:10.2174/1389203719666171129094937
Huang WC, Yen FC, Shie FS, et al. TGF-beta1 blockade of microglial chemotaxis toward Abeta aggregates involves SMAD signaling and down-regulation of CCL5. Journal of neuroinflammation. Apr 29 2010;7:28. doi:10.1186/1742-2094-7-28
Zhao C, Deng Y, He Y, Huang X, Wang C, Li W. Decreased Level of Exosomal miR-5121 Released from Microglia Suppresses Neurite Outgrowth and Synapse Recovery of Neurons Following Traumatic Brain Injury. Neurotherapeutics : the journal of the American Society for Experimental NeuroTherapeutics. 2021;18(2):1273-1294. doi:10.1007/s13311-020-00999-z
Orihuela R, McPherson CA, Harry GJ. Microglial M1/M2 polarization and metabolic states. Br J Pharmacol. Feb 2016;173(4):649-65. doi:10.1111/bph.13139
Mantovani A, Sica A, Sozzani S, Allavena P, Vecchi A, Locati M. The chemokine system in diverse forms of macrophage activation and polarization. Trends Immunol. Dec 2004;25(12):677-86. doi:10.1016/j.it.2004.09.015
van Horssen J, Kleinnijenhuis J, Maass CN, et al. Accumulation of heparan sulfate proteoglycans in cerebellar senile plaques. Neurobiol Aging. Jul-Aug 2002;23(4):537-45. doi:10.1016/s0197-4580(02)00010-6
Cui H, Freeman C, Jacobson GA, Small DH. Proteoglycans in the central nervous system: role in development, neural repair, and Alzheimer's disease. IUBMB Life. Feb 2013;65(2):108-20. doi:10.1002/iub.1118
Zhang GL, Zhang X, Wang XM, Li JP. Towards understanding the roles of heparan sulfate proteoglycans in Alzheimer's disease. Biomed Res Int. 2014;2014:516028. doi:10.1155/2014/516028
Perry G, Siedlak SL, Richey P, et al. Association of heparan sulfate proteoglycan with the neurofibrillary tangles of Alzheimer's disease. J Neurosci. Nov 1991;11(11):3679-83. doi:10.1523/jneurosci.11-11-03679.1991
Zhao J, Zhu Y, Song X, et al. 3-O-Sulfation of Heparan Sulfate Enhances Tau Interaction and Cellular Uptake. Angew Chem Int Ed Engl. Jan 27 2020;59(5):1818-1827. doi:10.1002/anie.201913029
Zhang X, Wang B, Li JP. Implications of heparan sulfate and heparanase in neuroinflammation. Matrix Biol. Apr 2014;35:174-81. doi:10.1016/j.matbio.2013.12.009
Collins LE, Troeberg L. Heparan sulfate as a regulator of inflammation and immunity. J Leukoc Biol. Jan 2019;105(1):81-92. doi:10.1002/jlb.3ru0618-246r
Jenkins LM, Horst B, Lancaster CL, Mythreye K. Dually modified transmembrane proteoglycans in development and disease. Cytokine Growth Factor Rev. Feb 2018;39:124-136. doi:10.1016/j.cytogfr.2017.12.003
Hashimoto Y, Nawa M, Chiba T, Aiso S, Nishimoto I, Matsuoka M. Transforming growth factor beta2 autocrinally mediates neuronal cell death induced by amyloid-beta. J Neurosci Res. May 1 2006;83(6):1039-47. doi:10.1002/jnr.20804
Noguchi A, Nawa M, Aiso S, Okamoto K, Matsuoka M. Transforming growth factor beta2 level is elevated in neurons of Alzheimer's disease brains. Int J Neurosci. Mar 2010;120(3):168-75. doi:10.3109/00207450903139689
Chong JR, Chai YL, Lee JH, et al. Increased Transforming Growth Factor β2 in the Neocortex of Alzheimer's Disease and Dementia with Lewy Bodies is Correlated with Disease Severity and Soluble Aβ42 Load. J Alzheimers Dis. 2017;56(1):157-166. doi:10.3233/jad-160781
López-Casillas F, Wrana JL, Massagué J. Betaglycan presents ligand to the TGF beta signaling receptor. Cell. Jul 2 1993;73(7):1435-44. doi:10.1016/0092-8674(93)90368-z
Ebisawa T, Fukuchi M, Murakami G, et al. Smurf1 Interacts with Transforming Growth Factor-β Type I Receptor through Smad7 and Induces Receptor Degradation*. Journal of Biological Chemistry. 2001/04/20/ 2001;276(16):12477-12480. doi:https://doi.org/10.1074/jbc.C100008200
Kavsak P, Rasmussen RK, Causing CG, et al. Smad7 Binds to Smurf2 to Form an E3 Ubiquitin Ligase that Targets the TGFβ Receptor for Degradation. Molecular Cell. 2000/12/01/ 2000;6(6):1365-1375. doi:https://doi.org/10.1016/S1097-2765(00)00134-9
David S. Knopman HA, Ronald C. Petersen, Gäel Chételat, David M. Holtzman, Bradley T. Hyman, Ralph A. Nixon & David T. Jones Alzheimer disease. Nat Rev Dis Primers. May 13 2021;7(1):34. doi:10.1038/s41572-021-00275-0
Scholefield Z, Yates EA, Wayne G, Amour A, McDowell W, Turnbull JE. Heparan sulfate regulates amyloid precursor protein processing by BACE1, the Alzheimer's beta-secretase. J Cell Biol. Oct 13 2003;163(1):97-107. doi:10.1083/jcb.200303059
Paumier JM, Py NA, García-González L, et al. Proamyloidogenic effects of membrane type 1 matrix metalloproteinase involve MMP-2 and BACE-1 activities, and the modulation of APP trafficking. Faseb j. Feb 2019;33(2):2910-2927. doi:10.1096/fj.201801076R

No competing interests reported.

Supplementarymaterial.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

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

Status:

Version 1

Abstract

Figures

Introduction

Materials and methods

2.1 Materials

2.2 Acquisition and preprocessing of the RNA-seq cohort from AD patients

2.3 Gene Set Enrichment Analysis

2.4 Weighted gene coexpression network analysis

2.5 Protein-Protein Interaction Analysis

2.6 Gene Functional and Pathway Enrichment Analysis

2.7 Convergent functional genomic ranks

2.8 Human brain tissue

2.9 Mice and cell culture

2.10 RT-qPCR analysis

2.11 ICV

2.12 Immunoblotting

2.13 Immunofluorescence

2.14 Enzyme-linked immunosorbent assay

2.15 Morris water maze

2.16 Acute LPS injection

2.17 Microglial skeleton analysis

2.18 Acute STZ injection

3 Statistical analysis

Results

Discussion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1