text
stringlengths 4
179k
| predicted_class
stringclasses 5
values | confidence
float64 0
100
|
|---|---|---|
ROS is a key regulator of NLRP3 inflammasome1125. We then investigated whether IgA ICs induced NLRP3 inflammasome activation through ROS-mediated pathways. We showed that IgA ICs induced ROS generation, but this effect was inhibited by N-acetyl-L-cysteine (NAC), an ROS scavenger (Fig. 3a). NAC also significantly reduced protein expression levels of NLRP3 and pro-IL-1β (Fig. 3b), caspase-1 activation (Fig. 3c) and IL-1β secretion (Fig. 3d) in IgA ICs-activated macrophages. | other | 30.52 |
Mitochondrial ROS and oxidized mitochondrial DNA have been shown to play a crucial role in the process of NLRP3 inflammasome activation1226. We showed that IgA ICs induced mitochondrial ROS generation and mitochondrial DNA release into the cytosol (Fig. 3e) of J774A.1 macrophages. In experiments using mito-TEMPO, an inhibitor of mitochondrial ROS, we showed that protein expression levels of NLRP3 and pro-IL-1β (Fig. 3f), caspase-1 activation (Fig. 3g), and IL-1β secretion (Fig. 3h) in IgA ICs-activated macrophages were significantly inhibited by mito-TEMPO. | other | 34.38 |
We tested whether IgA ICs can induce IL-1β secretion directly in renal intrinsic cells. The data show that IgA ICs induced caspase-1 activation (Fig. 4a) and IL-1β secretion (Fig. 4b) in primary MCs from WT mice, but these effects were significantly reduced in primary MCs from NLRP3 KO mice. IgA ICs also induced caspase-1 activation (Fig. 4c) and IL-1β secretion (Fig. 4d) in renal TECs transfection with shSC, but these effects were inhibited in TECs transfection with shNLRP3. These data suggest that NLRP3 inflammasome is involved in the IgA ICs-mediated inflammatory reaction in both the renal intrinsic cells. | other | 33.2 |
NLRP3 inflammasome controls caspase-1 activity and IL-1β release in various inflammatory diseases13141516, including IgAN2021. We chose an IgAN model passively induced by repeated injections of IgA and pneumococcal C-polysaccharide (PnC) antigen to make it feasible inducing the experimental IgAN in NLRP3 KO mice (NLRP3 KO + IgAN mice), with which we examined the impact of lack of NLRP3 inflammasome activation on the development of the IgAN model. As shown in Fig. 5, compared to WT mice treated with saline (WT + saline mice), renal protein levels of NLRP3 (Fig. 5a,b), IL-1β (Fig. 5a,c) and IL-18 (Fig. 5a,d) were significantly increased in the IgAN model in WT mice (WT + IgAN mice) on days 14 and 36, but these effects were significantly inhibited in NLRP3 KO + IgAN mice. In addition, NLRP3 KO + IgAN mice showed much lower caspase-1 were observed in NLRP3 KO + IgAN mice, compared to NLRP3 KO mice treated with saline (NLRP3 KO + saline mice). There were no increased expression levels of renal NLRP3, IL-1β and IL-18 in NLRP3 KO + IgAN mice compared to NLRP3 KO + saline mice on days 14 and 36. | other | 32.9 |
Besides, WT + IgAN mice presented increased albuminuria beginning on day 7 and persisting at high levels until day 36, as demonstrated by the urine albumin/creatinine (Cr) ratio, compared to WT + saline mice, whereas this effect was significantly decreased in NLRP3 KO + IgAN mice on days 28 and 36 (Fig. 6a). NLRP3 KO + IgAN mice still showed increased albuminuria compared to NLRP3 KO + saline mice. | other | 30 |
In addition, compared to WT + saline mice, WT + IgAN mice showed significantly increased serum levels of blood urea nitrogen (BUN) (Fig. 6b) and Cr (Fig. 6c) on days 14 and 36, while, in NLRP3 KO + IgAN mice, BUN levels were significantly lower than in WT + IgAN mice on days 14 and 36 and Cr levels were significantly decreased on day 36. NLRP3 KO + IgAN mice showed increased serum levels of BUN and of Cr on days 14 and 36, compared to NLRP3 KO + saline mice. | other | 33.6 |
Moreover, compared to WT + saline mice, WT + IgAN mice showed significantly increased levels of glomerular proliferation, glomerular sclerosis and periglomerular mononuclear leukocyte infiltration by light microscopy on days 14 and 36, and the severity of these renal lesions was greatly reduced in NLRP3 KO + IgAN mice (Fig. 6d,e). Increased levels of these renal lesions were observed in NLRP3 KO + IgAN mice on days 14 and 36 compared to NLRP3 KO + saline mice. | other | 31.12 |
The NLRP3 inflammasome plays a critical role in innate immunity and is involved in various inflammatory responses91012 and Th1 and Th2 differentiation can be altered by activation of the NLRP3 inflammasome11142728. We therefore examined the effect of NLRP3 deficiency on systemic levels of proinflammatory cytokines in the mice. As shown in Table 1, on days 14 and 36, serum levels of IL-1β were significantly increased in WT + IgAN mice compared to WT + saline mice, but this effect was not seen in NLRP3 KO + IgAN mice. On days 14 and 36, serum levels of TNF-α were significantly increased in WT + IgAN mice compared to WT + saline mice and this effect was markedly lower in NLRP3 KO + IgAN mice on day 36. Serum levels of IL-17A showed no significant differences between the 4 groups on day 14, but a significantly increase was seen in WT + IgAN mice compared to WT + saline mice on day 36, which was not seen in the NLRP3 KO + IgAN mice. Serum levels of IL-4 showed no difference between the 4 groups on day 14, but a significant increase in the NLRP3 KO + IgAN mice compared to WT + IgAN mice on day 36. Increased serum levels of IFN-γ were seen in WT + IgAN mice on day 36, and this effect was lower in NLRP3 KO + IgAN mice on day 36, although there was no statistical significance between the mice and WT + IgAN mice. Besides, NLRP3 KO + IgAN mice showed increased serum levels of TNF-α (days 14 and 36) and IL-4 (day 36) compared to NLRP3 KO + saline mice. There was no significant difference in serum levels of IL-1β (days 14 and 36), IL-17A (days 14 and 36) and IL-4 (day 14) between NLRP3 KO + IgAN mice and NLRP3 KO + saline mice. | other | 34.12 |
The NLRP3 inflammasome plays a crucial role in the innate immune response and links it to the adaptive immune response810112728. Notably, mononuclear leukocytes are often present around the inflamed glomeruli (periglomerular infiltration) in human IgAN3. In our study, on days 14 and 36, WT + IgAN mice showed markedly increased numbers of F4/80+ macrophages (both glomerular and periglomerular) (Fig. 7a,d), and CD11c+DCs (periglomerular) (Fig. 7b,e) and CD3+ T cells (periglomerular) (Fig. 7c,f) compared to WT + saline mice and these effects were significantly reduced in NLRP3 KO + IgAN mice. Increased numbers of periglomerular F4/80+ macrophages, CD11c+DCs and CD3+ T cells were observed in NLRP3 KO + IgAN mice on days 14 and 36, compared to NLRP3 KO + saline mice. | other | 31.89 |
Deficiency of the NLRP3 inflammasome can attenuate systemic T cell immune responses in nephrotoxic serum nephritis14. We showed IgA ICs induced IL-1β secretion through NLRP3 inflammasome in macrophages and DCs, and subsequent T cells activation (Figs 1 and 2). These data suggest a role of the NLRP3 inflammasome in the interaction between these antigen presenting cells and T cells in IgAN. In the spleen, as shown in Fig. 8, WT + IgAN mice had significantly increased numbers of CD4+CD44hiCD62lo-hi (Fig. 8a) and CD8+CD44hiCD62lo-hi (Fig. 8b) effector/memory T cells compared to WT + saline mice on both days 14 and 36, and this effect was absent in NLRP3 KO + IgAN mice. In addition, the NLRP3 inflammasome mediates the production of IL-1β, which can further promote the differentiation of Th17 cells2729. In our study, WT + IgAN mice had a significantly increased percentage of CD4+ IL-17A+ T cells in the spleen compared to WT + saline mice on both day 14 and day 36, and this effect was absent on day 14 and significantly reduced on day 36 in NLRP3 KO + IgAN mice (Fig. 8c). In addition, NLRP3 KO + IgAN mice had significantly increased percentage of CD4+CD25+ FoxP3+ Treg cells compared to WT + IgAN mice on days 14 and 36. There was no significant difference between NLRP3 KO + saline and WT + saline mice (Fig. 8d). | other | 32.84 |
To further test whether blockade of NLRP3 has therapeutic potential for IgAN and thereby also confirmed the pathogenic role of NLRP3 inflammasome in the disease model, we used kidney-targeting, ultrasound-mediated microbubble gene delivery of shNLRP3 into the mice. First, we tested how long injected shNLRP3-luciferase survived in vivo and, as shown in Fig. 9, luciferase activity in the mice, as detected by an in vivo imaging system, peaked at day 2, although elevated expression was also seen at days 3 and 7. We therefore decided to inject the mice with shNLRP3 before and after induction of IgAN. | other | 32.94 |
shSC (shSC + IgAN) or shNLRP3 (shNLRP3 + IgAN) was injected every 3 days starting one day before induction of IgAN and tested the mice at various days. shNLRP3 + IgAN mice showed greatly reduced severity of disease compared to shSC + IgAN mice, including significantly lower albuminuria levels on days 14 to 36 (Fig. 10a), better renal function on day 36, as demonstrated by serum levels of BUN and Cr (Fig. 10b), milder renal histopathology, such as lower glomerular proliferation, glomerular sclerosis, and periglomerular mononuclear leukocyte infiltration in the renal interstitium on day 36 (Fig. 10c,d), and renal infiltration of macrophages (F4/80+) and T cells (CD3+) on day 36 (Fig. 10c,d). | other | 34.4 |
In addition, although renal protein levels of NLRP3, IL-1β and IL-18 (Fig. 10e,f) and renal caspase-1 activity (Fig. 10g) on day 36 were increased in the shSC + IgAN mice compared to the controls, this effect was significantly inhibited in the shNLRP3 + IgAN mice. | other | 29.66 |
To mimic clinical status for treatment of IgAN, administration of shNLRP3 was started one week after the start of induction of IgAN in mice (IgAN + shNLRP3), which were sacrificed on day 36. Again, the results showed that IgAN + shNLRP3 mice demonstrated significantly improved renal conditions compared to IgAN + shSC mice, including albuminuria levels (Fig. 11a); renal function as demonstrated by serum levels of BUN, but not serum levels of Cr (Fig. 11b); renal histopathology, such as glomerular proliferation, glomerular sclerosis, and periglomerular mononuclear leukocyte infiltration in the renal interstitium (Fig. 11c,d); and renal infiltration of macrophages (F4/80+) and T cells (CD3+) (Fig. 11c,d). | other | 34.8 |
The results from gene delivery with shRNA NLRP3 were consistent with the findings in the experiments involving the use of NLRP3 KO mice, suggesting that NLRP3 inflammasome activation plays a pathogenic role in IgAN and blockade of NLRP3 may have therapeutic potential for the disease. | other | 38.1 |
The potential pathogenic role of NLRP3 inflammasome in IgAN remains unknown. In this study, we examined the causal-relationship between the NLRP3 inflammasome and the development of IgAN and identified the molecular mechanism underlying the pathogenic role of NLRP3 inflammasome in IgAN. Our data showed that IgA ICs-induced NLRP3 inflammasome activation in macrophages involving disruption of mitochondrial integrity and induction of mitochondrial ROS, and renal intrinsic cells (e.g., MCs and renal TECs); IgA ICs-induced DCs activation and resultant CD4+ T cells activation and differentiation/polarization; improved renal function and histopathology in the IgAN model in NLRP3 KO mice or kidney-targeting gene delivery of shNLRP3. These findings suggest that the NLRP3 inflammasome plays a pathogenic role in the development of IgAN in part by activation of T cells and mitochondrial damage and ROS production. To our knowledge, this is the first report on the NLRP3 inflammasome specifically involved in both antigen-presenting cells and glomerular MCs and renal TECs in the development of an IgAN model in mice, although, in other glomerular and tubulointerstitial conditions, the inflammasome has been shown to be upregulated in immune cells, glomerular MCs, podocytes, or renal TECs9101930. | study | 31.52 |
ROS released from mitochondria promotes mitochondrial permeability and facilitates the release of mitochondrial DNA into the cytosol, and the latter can activate the NLRP3 inflammasome1227. We showed that IgA ICs induced mitochondrial dysfunction in macrophages, as evidenced by an increased mitochondrial ROS production and mitochondrial DNA copy numbers (Fig. 3e). The IgA ICs-induced increase in levels of NLRP3 and pro-IL-1β in (Fig. 3b,f) and IL-1β secretion (Fig. 3d,h) by macrophages was inhibited by two specific inhibitors of ROS, including NAC and mito-TEMPO. Together, the results suggest a central role of mitochondrial ROS in the IgA ICs-mediated activation of the inflammasome. These results suggest the importance of mitochondrial integrity in IgA ICs-induced NLRP3 inflammasome activation in macrophages. In addition to mitochondria, the roles of other cellular sources of ROS in NLRP3 inflammasome activation have been investigated. Recently, NADPH oxidase 4 (NOX4), a source of cellular superoxide anions, promotes NLRP3 inflammasome activation by increasing the expression of carnitine palmitoyltransferase 1A, a key mitochondrial enzyme in the fatty acid oxidation pathway31. Ives et al. showed that ROS derived by xanthine oxidase regulates NLRP3 activation in macrophages, leading to excessive IL-1β and IL-18 secretion32. In our previous studies, we demonstrated that ROS derived by cyclooxygenase-2 increases NLRP3 inflammasome activation by increasing LPS-induced NLRP3 expression and caspase-1 activation, and these effects were associated with increased NF-κB activation and mitochondrial damage, respectively33. Furthermore, ROS derived by NADPH oxidase regulates both NLRP3 expression and caspase-1 activation in macrophages34. | study | 33.12 |
In addition, NLRP3 deficiency has been shown to inhibit systemic T cell immune responses in a mouse model of nephrotoxic serum nephritis through the activation of CD11c+ DCs14. In the present study, we showed that IgA ICs activated NLRP3 inflammasome in macrophages and DCs, and thereby activating T cells and in consistence with this, inhibited T cells activation in the spleen in NLRP3 KO + IgAN mice. Recently, we have shown that IL-17 is implicated in the pathogenesis of IgAN in the mouse model20, and it has been demonstrated that early Th17 cell differentiation involves IL-1 signaling in T cells2729. In the present study, we showed that NLRP3 deficiency resulted in decreased numbers of CD4+ IL-17A+ T cells (Fig. 8c) and increased percentage of Treg cells (Fig. 8d) in the mouse model of IgAN, and reduced production of IL-17A and IFN-γ although enhanced production of IL-4, by activated T cells (Fig. 2c). An exaggerated pro-inflammatory T cell response in WT + IgAN mice, demonstrated by increased effector/memory T cells (Fig. 8a,b). This might be contributed an imbalance caused by a reduction in Treg cells (Fig. 8d), at least in part, by a restoration in Treg cells. In addition, it has been shown that reduced levels of IL-17A can inhibit Th1 responses3536 and promote Th2 responses38 involving the activation of Treg cells. In consistence with these findings, in serum samples from NLRP3 KO + IgAN, levels of IL-17A and IFN-γ were decreased compared to those of WT + IgAN mice, although increased IL-4 levels were observed (Table 1). | other | 33 |
In contrast, IL-4, which favors Th2 responses has a protective role in renal inflammatory responses3738. Increased production of IL-4 in the CD4+ T cells activated by IgA ICs-primed BMDCs from NLRP3 KO mice, consistent with increased serum levels of the cytokine in the IgAN model of NLRP3 KO mice (the NLRP3 KO + IgAN mice), may further support the protective effect of NLRP3 deficiency in the IgAN mice. All together, our data support that the NLRP3 inflammasome plays a role in the T cell activation and differentiation/polarization underlying the observed renal inflammation in the mouse IgAN model. | other | 35.06 |
On the other hand, our data also showed that a caspase-11 dependent, non-canonical pathway of NLRP3 inflammasome activation is involved in the IgA ICs-mediated activation of macrophages (Fig. 1e,f). In the caspase-11-mediated activation of noncanonical NLRP3 inflammasome, gram-negative bacterial LPS are sensed by TLR4, and LPS-mediated endocytosis of TLR4 induces IFN-α/β expression via activating IRF3-IRF7 complex. Secreted IFN-α/β binds to the IFNAR1/IFNAR2 receptor increasing the caspase-11 gene expression through JAK/STAT pathway39. Although the PnC antigen contained in the IgA ICs was prepared from Gram-positive Streptococcus pneumoniae, we found that IgA ICs increased caspase-11 mRNA levels in macrophages, this result suggested that IgA ICs might induce IFN-α/β expression. There are two possible mechanisms of caspase-11 activation. One is that the induction of procaspase-11 expression is necessary and sufficient for its auto-activation40. The other might be a mechanism mediated by unidentified scaffold/receptor that is induced by intracellular bacteria40. In the present study, we show that knockdown of caspase-11 inhibited IL-1β secretion induced by IgA ICs (Fig. 1f). This finding suggests that caspase-11 induced by IgA ICs might lead to auto-activation of caspase-11 and subsequence of caspase-1 activation and IL-1β secretion. However, we can’t exclude the possibility of uptake of IgA ICs by macrophages, thereby activating the caspase-11 as a similar finding observed in bacteria-mediated caspase-11 activation. Consistently, Andersen et al.14 reported that a non-canonical mechanism of NLRP3 inflammasome activation is involved in renal injury in nephrotoxic serum nephritis in mice. In the present study, however, whether a non-canonical pathway of NLRP3 inflammasome was operated in renal intrinsic cells (e.g., epithelial cells, MCs, and endothelial cells) in IgAN remains further investigation. | other | 34.84 |
Very recently, Chun et al.18 reported that decreased expression of NLRP3 mRNA and protein in kidney biopsies of IgAN patients. The authors also showed that NLRP3 has been implicated in the pathogenesis of chronic kidney disease16. In view of their important findings and based on our data, we infer that the pathogenic role NLRP3 plays very likely in the initial or accelerating/progressing stages of the renal disorder, although further investigations are required to validate the hypothesis. | other | 35.8 |
It should be noted that the mouse model of IgAN was significantly attenuated by injection of a kidney-targeting shNLRP3 before or after induction of the mouse IgAN model (Figs 10 and 11). Recently, two anti-NLRP3 compounds, MCC95015 and β-hydroxybutyrate41, were shown to reduce the severity of NLRP3 inflammasome-mediated experimental autoimmune encephalomyelitis, Muckle-Wells syndrome, and autoinflammatory syndrome, etc. These findings further support that administration of a kidney-targeting shNLRP3 may be a potential therapeutic for IgAN. | other | 33.28 |
In this study, we decided to use the IgAN model passively induced by repeated injections of IgA and PnC antigen so that it was more feasible in performing the induction of the experimental IgAN in NLRP3 KO mice, and also because it was indeed a very stable IgAN model to reproduce the characteristic granular immunofluorescence pattern of IgA and C3 mesangial deposits, with renal inflammation and fibrosis720214243 and has been used in various experiments involving the role of different IgA subtypes in IgA ICs glomerular deposition, complement activation and ICs deposition, the clearance kinetics of circulating IgA ICs and the role of hepatic Kupffer cells in their elimination, the impact of the nature of antigen in IgA ICs in resultant renal injury, and the synergy between extra-renal cytokines and IgA mesangial deposits in the development of renal injury and dysfunction and in the evolution of renal histopathologic changes. | other | 36.75 |
However, in the present study, the exact role of the inflammasome in antigen-presenting cells and glomerular cells is not clearly established, this would require cell-specific gene-inactivation or at least bone marrow transplantation44. In addition, we could not exclude the possibility of other inflammasomes45 or their related pathways46 also involved in the mouse IgAN model at least because the treatment with shNLRP3 could not entirely prevent or improve the IgAN mice. This proposition should be warranted for further investigation. | other | 34.8 |
For IgAN induction, 8-week-old, female, NLRP3 KO mice (Center for Molecular and Clinical Immunology, Chang-Gung University, Tao-Yuan, Taiwan) and C57BL/6 mice (WT) (National Laboratory Animal Center, Taipei, Taiwan) were used. Briefly, IgAN was induced by 36 daily injections of purified IgA anti-phosphorylcholine antibodies and PnC into NLRP3 KO mice and WT mice as described previously720214243. For OVA-specific T cell proliferation assay, 8-week-old, female OT-II mice (provided by Dr. C. Lowell, University of California, San Francisco, CA) were used as described previously47. All animal experiments were performed after approval by the Institutional Animal Care and Use Committee of the National Defense Medical Center, Taiwan and were consistent with the NIH Guide for the Care and Use of Laboratory Animals. | other | 38.72 |
Murine J774A.1 macrophages (TIB-67™), murine M-1 renal TECs (CRL-2038™), and 293 T cells were obtained from the American Type Culture Collection. Mouse peritoneal macrophages48, BMDCs47, and glomerular PMCs49 were prepared from NLRP3 KO and WT mice as described previously. All cells were cultured at 37 °C in a 5% CO2 incubator. | other | 34.88 |
Urine samples were collected in metabolic cages and albuminuria determined by the urine albumin/Cr ratio in samples taken at 0, 7, 14, 28, and 36 days as described previously50. The mice were sacrificed at day 14 and day 36 after IgAN induction, and renal cortical tissue and blood samples taken, respectively. Serum levels of BUN and Cr were measured using BUN kits or Cr kits (both from Fuji Dry-Chem Slide, Fuji Film Medical). Renal tissues were fixed in 10% formalin and paraffin-embedded, then 4 um thick sections were cut and stained with hematoxylin and eosin. Scoring of the severity of renal histopathology were performed as described previously20. The percentage of glomeruli showing glomerular proliferation, glomerular sclerosis, or periglomerular mononuclear leukocyte infiltration was determined by counting in 50 randomly sampled glomeruli by light microscopy at a magnification of x400. | other | 44.16 |
Methyl Carnoy’s solution-fixed, paraffin-embedded renal sections were used for the detection of CD3+ cells (pan-T cells, Dako) and F4/80+ cells (macrophages; Serotec), while frozen sections of renal tissues were used for the detection of CD11c+ cells (DCs, BD Biosciences). The number of CD3+, F4/80+, and CD11c+ cells were counted at a magnification of x400 in 50 randomly selected glomeruli or in 20 randomly selected fields of the tubulointerstitial compartment in the renal cortex using Pax-it quantitative image analysis software (Pax-it; Paxcam) as described previously51. | other | 44.3 |
Protein lysates from renal cortical tissues and cultured cells, and supernatants from cultured cells were run on 10% or 15% SDS–PAGE gels. Mouse against NLRP3 (AdipoGen), rabbit against IL-1β, IL-18, caspase-1 or goat anti-β-actin antibodies (all from Santa Cruz Biotechnology) were used as primary antibodies and horseradish peroxidase-conjugated rabbit anti-goat, goat anti-rabbit or goat anti-mouse IgG antibodies as secondary antibodies (all from Santa Cruz Biotechnology). The membrane-bound antibody detected was incubated Enhanced Reagent Plus (PerkinElmer Life Sciences) and PVDF membranes were scanned with an UVP BioSpectrum Imaging Systems (Financial HealthCare) as described previously52. | other | 39.97 |
IL-1β, TNF-α, IL-17A, IL-4, and IFN-γ levels in serum or supernatants of cultured cells were measured using ELISA kits (all from R&D Systems), while caspase-1 activity in lysates was measured using a caspase-1 fluorometric kit (R&D Systems) according to the manufacturer’s instructions. | other | 42.53 |
RNA was extracted using TriZOL reagent (Invitrogen) and cDNA prepared as described previously50. The primers used were: for mouse caspase-11, forward 5-GCCACTTGCCAGGTCTACGAG-3 and reverse 5-AGGCCTGCACAATGATGACTTT-3; for mouse GAPDH, forward 5-TCCGCC CCTTCTGCCGATG-3 and reverse 5-CACGGAAGGCCATGCCAGTGA-3. Mitochondrial DNA in the cytosol was extracted as described previously27. The primers used were: for mouse cytochrome oxidase I, forward 5-GCCCCAGATATAGCATTCCC-3, and reverse 5-GTTCATCCTGTTCCC-3, and for mouse 18S, forward 5-TAGAGGGACAAGTGGCGTTC-3, and reverse 5-CGCTGAGCCAGTCAGTGT-3. Real-time PCR was performed on an ABI Prism 7700 Sequence Detection System (Applied Biosystems) using SYBR Green mix (Thermo Scientific). | other | 38.9 |
Isolated splenocytes were stained to measure activation of T cells using allophycocyanin (APC)-conjugated anti-CD4 (RM4-5) or CD8 (53-6.7) antibodies, phycoerythrin (PE)-conjugated anti-CD62 antibodies (MEL-14), or fluorescein isothiocyanate (FITC)-conjugated anti-CD44 antibodies (IM7) (all from BD Biosciences). The cells were stained for Treg cells using the Mouse Regulatory T cell Staining Kit (eBioscience) according to the manufacturer’s instructions. Maturation of BMDCs was determined by the upregulation of CD11c+ and costimulatory molecules expression, the cells were stained with PE-conjugated anti-CD11c antibody, then with FITC-conjugated anti-CD40 or anti-CD86 antibodies (all from BD Biosciences), respectively, as described previously47. To measure percentage of CD4+ IL-17A+ T cells, intracellular cytokine staining was performed and the cells were stained with APC-conjugated anti-CD4 antibody, then with PE-conjugated anti-IL-17 antibody (BD Biosciences) as described previously53. Cells were analyzed by a FACSCalibur (BD Biosciences) as described previously20. | other | 38.28 |
Production of total ROS in J774A.1 macrophages was determined by measuring the intensity of fluorescence of 2′, 7′-dichlorofluorescein, the oxidation product of 2′, 7′-dichlorofluorescein diacetate (Molecular Probes), while mt ROS production was assessed by measuring the fluorescence of MitoSOX (Invitrogen) as described previously33. | other | 42.75 |
Lentivirus transduction particles carrying shNLRP3 or caspase-11 (National RNAi Core Facility, Academia Sinica, Taipei, Taiwan) in 293 T cells were constructed. J774A.1 macrophages or TECs were then infected with lentivirus-bearing specific shRNAs of NLRP3 or caspase-11, respectively, and incubated with puromycin (Invitrogen) to select stably-infected cells for further experiments as described previously54. | other | 36.3 |
BMDCs from NLRP3 KO or WT mice were incubated with IgA ICs for 24 h, then incubated 18 h with OVA323–339 peptide (1 μg/ml) (Genomics) and CD4+ T cells derived from OT-II mice, and T cell proliferation was measured by [3H] thymidine incorporation as described previously47. | other | 39.34 |
Plasmids of shNLRP3-luciferase was generated by conjugating shNLPR3 (Santa Cruz) with luciferase as described previously55. The plasmids in 200 μl of saline was mixed with 200 μl of SonoVue® microbubbles (Bracco), and the mixture injected into mice via the tail vein, followed by application of transcutaneous ultrasound to the back at the level of the kidney using a Sonopuls 590 at 1 MHz (Ernaf-Nonius) as described previously50. To determine the time-course of shNLRP3 expression, and the interval at which shNLRP3 plasmids were injected into WT mice, the mice received the shNLRP3 plasmids (200 μg) and monitored for the presence of luciferase using an real-time imaging system IVIS 100 series (Xenogen Corp.) on days 0, 1, 2, 3, and 7 after the delivery. Control mice were injected with shSC plasmids. To evaluate the renoprotective effect of shNLRP3 on IgAN mice, the plasmids were given every 3 days starting one day before induction of IgAN, while, to evaluate the therapeutic effect on diseased mice, the treatment was given every 3 days starting seven days after induction of the disease. | other | 38.16 |
Results were presented as the mean ± SEM. Comparisons between two groups were performed using Mann-Whitney U test. The significance of differences in urinary albumin/Cr levels and the level of bioluminescence activity of injected shNLRP3 in vivo assessed by IVIS was examined using Kruskal-Wallis test. For the in vitro experiments, data analysis involved Kruskal-Wallis test. A value of p < 0.05 was considered statistically significant. | other | 39.25 |
Satb2 is a transcriptional regulator that binds to matrix attachment regions in the DNA and recruits chromatin-modifying complexes at the anchorage sites (Baranek et al., 2012; Britanova et al., 2005; Gyorgy et al., 2008; Szemes et al., 2006). Furthermore, similarly to its homologue Satb1 (Wang et al., 2014, 2012), Satb2 modifies higher-order chromatin structure by mediating the formation of intra-chromosomal DNA loops (Zhou et al., 2012). | study | 28.81 |
Recent genome-wide association studies of schizophrenia have identified SATB2 as a genetic risk locus (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014). Moreover, patients with mutations or deletions within the SATB2 locus, a condition referred to as ‘SATB2-associated syndrome (SAS)’, exhibit severe learning difficulties and profound mental retardation, providing further indication for a potential role of SATB2 in higher brain function (Liedén et al., 2014; Zarate et al., 2015; Zarate and Fish, 2016; Marshall et al., 2008). So far the neuropsychiatric symptoms of SAS have been discussed in the context of the established role of Satb2 during embryonic development of the cerebral cortex. In the embryonic cortex Satb2 is restricted to upper layer neurons where it inhibits the corticospinal motor neuron fate and promotes callosal neuron identity (Alcamo et al., 2008; Britanova et al., 2008; Leone et al., 2015; Srinivasan et al., 2012; Srivatsa et al., 2014). Thus, deficits in cortico-cortical connections could account for the reported neurological defects in SAS patients. However, patients with SATB2 haploinsufficiency have no apparent corpus callosum abnormalities (Lee et al., 2015; Rosenfeld et al., 2009). In heterozygous Satb2 knockout mice, resembling the genetic condition of SAS patients, the corpus callosum is also intact (Alcamo et al., 2008). This suggests a function of Satb2 in adult brain independent from its developmental role. The function of Satb2 in the adult central nervous system (CNS) is completely unknown since germ-line Satb2-deficient mice die perinatally (Dobreva et al., 2006). In contrast to the layer-specific embryonic expression, adult CNS Satb2 is expressed in pyramidal neurons of all layers of the cerebral cortex and in the hippocampal CA1 area (Huang et al., 2013). As both brain regions are tightly linked to cognition, Satb2 is well-positioned to regulate cognitive processes. | study | 31.6 |
(A) Satb2 is mainly expressed in the adult forebrain. Immunostaining for Satb2 of sagittal brain sections from adult mice. Scale bar: 1000 μm. (B) Immunobloting analysis of the Satb2 protein level in cortical or hippocampal lysates from adult Satb2 cKO mice (cKO) or Satb2flox/flox mice (Ctrl). Erk2 was used as loading control. Representative images are shown. (C) Satb2 immunostaining of coronal brain sections from 3-month old Satb2 cKO mice or Satb2flox/flox littermate controls. Nuclei were counterstained with DAPI. Representative images are shown. Scale bar: 200 μm. (D) Nissl-stained coronal brain sections from 3-month old Satb2 cKO mice and littermate controls, demonstrating normal gross brain morphology of Satb2 cKO animals. Scale bar: 200 μm. High magnification views of boxed areas reveal the normal cyto-architecture of the cortex and hippocampus of Satb2 cKO mice. Representative images are shown. Scale bar: 50 μm. (E) Immunohistochemical labeling against the CA1 specific marker Wfs1 in hippocampus of Satb2 cKO mice and littermate controls reveals normally developed hippocampal CA1 area in Satb2 mutants. Nuclei were counterstained with DAPI. Representative images are shown. Scale bar: 150 μm. | other | 31.89 |
10.7554/eLife.17361.003Figure 1—figure supplement 1.Satb2 is expressed in the cortex and hippocampus of both Satb2 conditional mutants and littermate controls at postnatal day 15.Immunostaining for Satb2 of coronal brain sections from juvenile, postnatal day 15 Satb2 cKO mice (cKO) and Satb2flox/flox littermate controls (Ctrl). Nuclei were counterstained with DAPI. Representative images are shown. Scale bar: 200 μm.DOI: http://dx.doi.org/10.7554/eLife.17361.003 | study | 33.84 |
In this study, we investigated the role of Satb2 in the mature mouse brain by selectively deleting Satb2 from forebrain excitatory neurons after the third postnatal week. Our results demonstrate deficient long-term potentiation (LTP) and long-term memory in Satb2 conditional mutants. At a mechanistic level, we establish Satb2 as a nuclear component of two main pathways implicated not only in cognition but also in schizophrenia pathophysiology, i.e. BDNF signaling and miRNA-mediated post-transcriptional regulation of gene expression. | other | 27.6 |
Given the highly specific expression pattern of Satb2 in the adult brain (Figure 1A) as well as the severe learning disabilities and mental retardation observed in SAS patients, we hypothesized that Satb2 is critical for learning and memory. To circumvent the perinatal and early postnatal lethality of the existing constitutive and conditional Satb2 mutants (Dobreva et al., 2003; Srinivasan et al., 2012) and to be able to perform behavioral experiments, we generated a novel conditional Satb2 mutant line by crossing mice bearing a floxed allele of Satb2 (Satb2flox/flox) with mice that express Cre recombinase under the Camk2a promoter (Minichiello et al., 1999). The expression of the Camk2a-Cre transgene allowed for a forebrain-specific deletion of Satb2 from the third postnatal week on, thus bypassing the confounding effects of early Satb2 inactivation on the formation of cortical neuronal circuits (Alcamo et al., 2008; Britanova et al., 2008; Harb et al., 2016; Leone et al., 2015; Srinivasan et al., 2012; Srivatsa et al., 2014). The absence of Satb2 protein in the cortex and hippocampus of adult but not juvenile Satb2flox/flox::Camk2a-Cre mice (Satb2 cKO) was confirmed by immunoblotting (Figure 1B) and immunohistochemistry (Figure 1C, Figure 1—figure supplement 1). Satb2 cKO mice were viable, fertile, and reached the same age and body weights as their littermate controls (Figure 1—figure supplement 2A). Gross morphological examination revealed no abnormalities in the brain of Satb2 cKO mutants (Figure 1—figure supplement 2B). Corpus callosum and the cellular layers of the neocortex and hippocampus appeared intact (Figure 1D). Immunoreactivity for the CA1-specific marker Wfs1 (Figure 1E) and the cortical layer markers Cux1, Ctip2, Tbr1 (Figure 1—figure supplement 2C) was undistinguishable from control mice, suggesting normal cortical and hippocampal morphology in Satb2 conditional mutants.10.7554/eLife.17361.002Figure 1.Characterization of Satb2 conditional mutants.(A) Satb2 is mainly expressed in the adult forebrain. Immunostaining for Satb2 of sagittal brain sections from adult mice. Scale bar: 1000 μm. (B) Immunobloting analysis of the Satb2 protein level in cortical or hippocampal lysates from adult Satb2 cKO mice (cKO) or Satb2flox/flox mice (Ctrl). Erk2 was used as loading control. Representative images are shown. (C) Satb2 immunostaining of coronal brain sections from 3-month old Satb2 cKO mice or Satb2flox/flox littermate controls. Nuclei were counterstained with DAPI. Representative images are shown. Scale bar: 200 μm. (D) Nissl-stained coronal brain sections from 3-month old Satb2 cKO mice and littermate controls, demonstrating normal gross brain morphology of Satb2 cKO animals. Scale bar: 200 μm. High magnification views of boxed areas reveal the normal cyto-architecture of the cortex and hippocampus of Satb2 cKO mice. Representative images are shown. Scale bar: 50 μm. (E) Immunohistochemical labeling against the CA1 specific marker Wfs1 in hippocampus of Satb2 cKO mice and littermate controls reveals normally developed hippocampal CA1 area in Satb2 mutants. Nuclei were counterstained with DAPI. Representative images are shown. Scale bar: 150 μm.DOI: http://dx.doi.org/10.7554/eLife.17361.00210.7554/eLife.17361.003Figure 1—figure supplement 1.Satb2 is expressed in the cortex and hippocampus of both Satb2 conditional mutants and littermate controls at postnatal day 15.Immunostaining for Satb2 of coronal brain sections from juvenile, postnatal day 15 Satb2 cKO mice (cKO) and Satb2flox/flox littermate controls (Ctrl). Nuclei were counterstained with DAPI. Representative images are shown. Scale bar: 200 μm.DOI: http://dx.doi.org/10.7554/eLife.17361.00310.7554/eLife.17361.004Figure 1—figure supplement 2.Postnatal Satb2 deletion does not cause alterations in body weight, gross brain morphology and cortical layer-specific marker expression.(A) Satb2 cKO mice (cKO) are visually indistinguishable from Satb2flox/flox mice (Ctrl) and have similar body weights to their control littermates (Ctrl, n = 12; cKO, n = 17; Student's t test, t27 = 0.0551, p = 0.9665). Left panel: Representative images of 3-month old male mice from both genotypes; Right panel: Bar graph of mean body weights. Data are presented as mean ± SEM. (B) Representative images of freshly dissected brains of 3-month old male Satb2 cKO and control mice (Left panel). Brain weights normalized to the body weights showed no difference between Satb2 cKO mice (n = 17) and littermate controls (n = 12); Student's t test, t27 = 0.1664, p = 0.8691. Data are presented as mean ± SEM (Right panel). (C) Representative images showing Ctip2, Cux1 and Tbr1 immunoreactivity in the cortex of 3-month old Satb2 cKO and Satb2flox/flox mice. Nuclei were counterstained with DAPI. Scale bar: 100 μm.DOI: http://dx.doi.org/10.7554/eLife.17361.004 | review | 29.78 |
10.7554/eLife.17361.004Figure 1—figure supplement 2.Postnatal Satb2 deletion does not cause alterations in body weight, gross brain morphology and cortical layer-specific marker expression.(A) Satb2 cKO mice (cKO) are visually indistinguishable from Satb2flox/flox mice (Ctrl) and have similar body weights to their control littermates (Ctrl, n = 12; cKO, n = 17; Student's t test, t27 = 0.0551, p = 0.9665). Left panel: Representative images of 3-month old male mice from both genotypes; Right panel: Bar graph of mean body weights. Data are presented as mean ± SEM. (B) Representative images of freshly dissected brains of 3-month old male Satb2 cKO and control mice (Left panel). Brain weights normalized to the body weights showed no difference between Satb2 cKO mice (n = 17) and littermate controls (n = 12); Student's t test, t27 = 0.1664, p = 0.8691. Data are presented as mean ± SEM (Right panel). (C) Representative images showing Ctip2, Cux1 and Tbr1 immunoreactivity in the cortex of 3-month old Satb2 cKO and Satb2flox/flox mice. Nuclei were counterstained with DAPI. Scale bar: 100 μm.DOI: http://dx.doi.org/10.7554/eLife.17361.004 | study | 33.94 |
(A) Satb2 cKO mice (cKO) are visually indistinguishable from Satb2flox/flox mice (Ctrl) and have similar body weights to their control littermates (Ctrl, n = 12; cKO, n = 17; Student's t test, t27 = 0.0551, p = 0.9665). Left panel: Representative images of 3-month old male mice from both genotypes; Right panel: Bar graph of mean body weights. Data are presented as mean ± SEM. (B) Representative images of freshly dissected brains of 3-month old male Satb2 cKO and control mice (Left panel). Brain weights normalized to the body weights showed no difference between Satb2 cKO mice (n = 17) and littermate controls (n = 12); Student's t test, t27 = 0.1664, p = 0.8691. Data are presented as mean ± SEM (Right panel). (C) Representative images showing Ctip2, Cux1 and Tbr1 immunoreactivity in the cortex of 3-month old Satb2 cKO and Satb2flox/flox mice. Nuclei were counterstained with DAPI. Scale bar: 100 μm. | other | 30.16 |
To test whether Satb2 is required for learning and memory we used contextual fear conditioning, a hippocampus-dependent paradigm of associative learning. Satb2 cKO mice showed a normal response to electric foot shock exposure (Figure 2—figure supplement 1) and normal fear acquisition during fear conditioning (Figure 2A). Short-term memory for contextual fear, measured 1 hr following training, was also not affected in Satb2 cKO mice. However, Satb2 cKO mice exhibited a significant decrease in freezing when compared to control littermates 24 hr after training (Figure 2A), indicating a specific deficit in the consolidation of associative memory. Next, we subjected Satb2 conditional mutants to the object location memory (OLM) and novel object recognition memory (ORM) tasks. Again, Satb2 cKO mice demonstrated normal short-term (1 hr) memory but significant deficits in long-term (24 hr) OLM (Figure 2B) and ORM (Figure 2C), providing evidence for requirement of Satb2 for long-term object discrimination/placement memories.10.7554/eLife.17361.005Figure 2.Satb2 is required for long-term memory formation.(A) In a contextual fear conditioning paradigm, Satb2 cKO mice (cKO), showed (i) similar levels of freezing to Satb2flox/flox mice (Ctrl) during the fear-acquisition phase (cKO, n = 7; Ctrl, n = 6; repeated measures ANOVA, F3,33 = 0.76, p = 0.52) and at the 1 hr fear expression test (Student's t test, t11 = 0.19, p = 0.86) but (ii) froze significantly less than their littermate controls at the 24 hr fear expression test (cKO, n = 8; Ctrl, n = 8; repeated measures ANOVA, F3,42 = 0.36, p = 0.778; Student's t test, t14 = 4.88, p = 0.0002). Data are presented as mean ± SEM, n values refer to the number of mice per group, *p < 0.05. (B) Object location memory test. (i) Scheme of the experiment. (ii) Satb2 cKO mice (n = 11) and control mice (n = 10) exhibited similar preference for the novel location over the familiar location at the 1 hr memory retention test (Student's t test, t19 = 0.46, p = 0.65). (iii) Satb2 cKO mice (n = 8) showed reduced preference for the novel location over the familiar location at the 24 hr memory retention test (Student's t test, t14 = 2.89, p = 0.011) compared to Satb2flox/flox mice (n = 8). The relative exploration time is expressed as a percent discrimination index (D.I. = (tnovel location − tfamiliar location) / (tnovel location + tfamiliar location) × 100%). Data are presented as mean ± SEM, n values refer to the number of mice, **p < 0.01. (C) Novel object recognition test. (i) Scheme of the experiment. (ii) Satb2 cKO mice (n = 10) and control mice (n = 9) exhibited a similar preference for the novel over the familiar object at the 1 hr memory retention test (Student's t test, t19 = 1.11, p = 0.28). (iii) Satb2 cKO mice (n = 8) spent less time exploring the novel object at the 24 hr memory retention test (Student's t test, t14 = 3.0, p = 0.009) compared to Satb2flox/flox mice (n = 8). The relative exploration time is expressed as a percent discrimination index (D.I. = (tnovel object− tfamiliar object) / (tnovel object+ tfamiliar object) × 100%). Data are presented as mean ± SEM, n values refer to the number of mice, *p < 0.05.DOI: http://dx.doi.org/10.7554/eLife.17361.00510.7554/eLife.17361.006Figure 2—figure supplement 1.Satb2 cKO mice show normal responses to electric foot shock.Flinch-Jump test revealed similar general responses to electric foot shock between Satb2 cKO mice (n = 5) and control littermates (n = 5), Student’s t-test, flinch: t8 = 0, p = 1.0, run/jump: t8 = 0.25, p = 0.81, vocalize: t8 = 0.24, p = 0.82. Mice were scored for their first visible response to the shock (flinch), their first pronounced motor response (run or jump), and their first vocalized distress as described in Materials and Methods. Data are presented as mean ± SEM.DOI: http://dx.doi.org/10.7554/eLife.17361.006 | clinical case | 26.64 |
(A) In a contextual fear conditioning paradigm, Satb2 cKO mice (cKO), showed (i) similar levels of freezing to Satb2flox/flox mice (Ctrl) during the fear-acquisition phase (cKO, n = 7; Ctrl, n = 6; repeated measures ANOVA, F3,33 = 0.76, p = 0.52) and at the 1 hr fear expression test (Student's t test, t11 = 0.19, p = 0.86) but (ii) froze significantly less than their littermate controls at the 24 hr fear expression test (cKO, n = 8; Ctrl, n = 8; repeated measures ANOVA, F3,42 = 0.36, p = 0.778; Student's t test, t14 = 4.88, p = 0.0002). Data are presented as mean ± SEM, n values refer to the number of mice per group, *p < 0.05. (B) Object location memory test. (i) Scheme of the experiment. (ii) Satb2 cKO mice (n = 11) and control mice (n = 10) exhibited similar preference for the novel location over the familiar location at the 1 hr memory retention test (Student's t test, t19 = 0.46, p = 0.65). (iii) Satb2 cKO mice (n = 8) showed reduced preference for the novel location over the familiar location at the 24 hr memory retention test (Student's t test, t14 = 2.89, p = 0.011) compared to Satb2flox/flox mice (n = 8). The relative exploration time is expressed as a percent discrimination index (D.I. = (tnovel location − tfamiliar location) / (tnovel location + tfamiliar location) × 100%). Data are presented as mean ± SEM, n values refer to the number of mice, **p < 0.01. (C) Novel object recognition test. (i) Scheme of the experiment. (ii) Satb2 cKO mice (n = 10) and control mice (n = 9) exhibited a similar preference for the novel over the familiar object at the 1 hr memory retention test (Student's t test, t19 = 1.11, p = 0.28). (iii) Satb2 cKO mice (n = 8) spent less time exploring the novel object at the 24 hr memory retention test (Student's t test, t14 = 3.0, p = 0.009) compared to Satb2flox/flox mice (n = 8). The relative exploration time is expressed as a percent discrimination index (D.I. = (tnovel object− tfamiliar object) / (tnovel object+ tfamiliar object) × 100%). Data are presented as mean ± SEM, n values refer to the number of mice, *p < 0.05. | other | 27.61 |
10.7554/eLife.17361.006Figure 2—figure supplement 1.Satb2 cKO mice show normal responses to electric foot shock.Flinch-Jump test revealed similar general responses to electric foot shock between Satb2 cKO mice (n = 5) and control littermates (n = 5), Student’s t-test, flinch: t8 = 0, p = 1.0, run/jump: t8 = 0.25, p = 0.81, vocalize: t8 = 0.24, p = 0.82. Mice were scored for their first visible response to the shock (flinch), their first pronounced motor response (run or jump), and their first vocalized distress as described in Materials and Methods. Data are presented as mean ± SEM.DOI: http://dx.doi.org/10.7554/eLife.17361.006 | study | 29.47 |
Flinch-Jump test revealed similar general responses to electric foot shock between Satb2 cKO mice (n = 5) and control littermates (n = 5), Student’s t-test, flinch: t8 = 0, p = 1.0, run/jump: t8 = 0.25, p = 0.81, vocalize: t8 = 0.24, p = 0.82. Mice were scored for their first visible response to the shock (flinch), their first pronounced motor response (run or jump), and their first vocalized distress as described in Materials and Methods. Data are presented as mean ± SEM. | other | 31.77 |
Next, we investigated the effect of loss of Satb2 on long-term potentiation (LTP), an electrophysiological correlate of memory formation (Mayford et al., 2012). To this aim, we prepared acute slices from Satb2 cKO mice and littermate control animals and tested LTP at hippocampal Schaffer collateral-CA1 synapses. Field-excitatory post-synaptic potential recordings from the apical dendritic layer of the CA1 region showed that the early phase of LTP (up to 40 min post-theta burst stimulation) did not differ from control values; however late-LTP (45–180 min post-theta burst stimulation) was significantly attenuated in Satb2 cKO mice (Figure 3A). The slopes of the input-output curves (Figure 3B) and the paired-pulse facilitation ratios across different inter-pulse intervals (Figure 3C) did not differ between mutant and control mice indicating normal basal synaptic transmission and presynaptic function in Satb2 cKO mice. Hence, Satb2 is not required for short-term plasticity at CA3–CA1 synapses but is essential for late-LTP maintenance.10.7554/eLife.17361.007Figure 3.Late-LTP maintenance is impaired in Satb2 conditional mutants.(A) Schaffer collateral-CA1 late-LTP was significantly impaired in Satb2 cKO mice (Student's t test, t28 = 4.92, p < 0.0001 for the interval 180–185 min post-theta burst stimulation, TBS). Shown are field EPSP slopes in Satb2flox/flox (Ctrl, n = 17 slices, 6 mice) vs. Satb2 cKO mice (cKO, n = 13 slices, 6 mice) recorded before and after TBS (100 Hz, repeated three times in a 10 s interval). Data are presented as mean ± SEM. (B) Input-output curves comparing the amplitudes of the presynaptic fiber volley to the field EPSP amplitude across a range of stimulation currents showed that basal synaptic transmission did not differ in hippocampal slices from Satb2 cKO mice (cKO, n = 17 slices, 6 mice) and littermate controls (Ctrl, n = 13 slices, 5 mice); Student's t test, p > 0.05 for all data points. Data are presented as mean ± SEM. (C) Paired-pulse facilitation studies across different inter-stimulus intervals revealed no difference between Satb2 cKO mice (cKO, n = 17 slices, 6 mice) and littermate controls (Ctrl, n = 13 slices, 5 mice); Student's t test, p > 0.05 for all data points. Data are presented as mean ± SEM.DOI: http://dx.doi.org/10.7554/eLife.17361.007 | review | 27.66 |
(A) Schaffer collateral-CA1 late-LTP was significantly impaired in Satb2 cKO mice (Student's t test, t28 = 4.92, p < 0.0001 for the interval 180–185 min post-theta burst stimulation, TBS). Shown are field EPSP slopes in Satb2flox/flox (Ctrl, n = 17 slices, 6 mice) vs. Satb2 cKO mice (cKO, n = 13 slices, 6 mice) recorded before and after TBS (100 Hz, repeated three times in a 10 s interval). Data are presented as mean ± SEM. (B) Input-output curves comparing the amplitudes of the presynaptic fiber volley to the field EPSP amplitude across a range of stimulation currents showed that basal synaptic transmission did not differ in hippocampal slices from Satb2 cKO mice (cKO, n = 17 slices, 6 mice) and littermate controls (Ctrl, n = 13 slices, 5 mice); Student's t test, p > 0.05 for all data points. Data are presented as mean ± SEM. (C) Paired-pulse facilitation studies across different inter-stimulus intervals revealed no difference between Satb2 cKO mice (cKO, n = 17 slices, 6 mice) and littermate controls (Ctrl, n = 13 slices, 5 mice); Student's t test, p > 0.05 for all data points. Data are presented as mean ± SEM. | study | 26.81 |
Given our findings of impaired hippocampal late-LTP in Satb2 conditional mutants, we examined whether neuronal activity or the neurotrophin BDNF, a mediator of structural and functional plasticity at synapses (Zagrebelsky and Korte, 2014), regulate Satb2 in primary hippocampal neurons. | review | 27.2 |
We treated hippocampal cultures with bicuculline (Bic) and 4-aminopyridine (4AP), a combination that causes robust action potential (AP) bursting (Hardingham et al., 2002). Bic/4AP application for 24 hr resulted in a strong up-regulation of both Satb2 mRNA (2.5-fold change, n = 7, Student's t test, p = 0.003) and protein levels (6.3-fold change, ANOVA, Ctrl vs. Bic/4AP, p = 0.002, Figure 4A). Next, we applied MK-801, a NMDA receptor antagonist, or nimodipine, an L-type VGCC blocker, together with Bic/4AP for 24 hr to determine which source of calcium (Bengtson et al., 2013) is required for Bic/4AP-triggered Satb2 induction. Nimodipine, but not MK-801, blocked the increase of Satb2 following synaptic activity indicating that Satb2 induction after synaptic stimulation depends on calcium influx through L-type VGCC and not through NMDA receptors (Figure 4A). Also BDNF and NT4, which both bind to the tyrosine kinase TrkB receptor, caused a significant increase in Satb2 mRNA (1.8-fold change, n = 7, Student's t test, p = 0.003) and protein (3-fold change, ANOVA, Ctrl vs. BDNF, p = 0.0004; Ctrl vs. NT4, p = 0.0002) levels 24 hr after treatment (Figure 4B).10.7554/eLife.17361.008Figure 4.Synaptic activity and BDNF up-regulate Satb2 in primary hippocampal neurons.(A) Increased synaptic activity up-regulates Satb2 protein depending on calcium influx through L-type VGCC. Representative Western blot (top) and quantification (bottom) of the Satb2 protein level 24 hr after Bic/4AP treatment in the presence or absence of L-VGCC blocker nimodipine or NMDAR antagonist MK-801 (n = 10, 6, 7, 5; ANOVA followed by Hochberg post hoc test; F3,24 = 9.171, Ctrl vs. Bic/4AP, p = 0.002; Ctrl vs. Bic/4AP+MK-801, p = 0.006; Bic/4AP vs. Bic/4AP+Nim, p = 0.011). (B) BDNF and NT4 significantly increase Satb2 protein 24 hr after treatment. Representative Western blot image (top) and quantification of the Satb2 protein level (bottom) are shown, n = 4; ANOVA followed by Tukey post hoc test, F4,15 = 15.4, Ctrl vs. BDNF, p = 0.0004; Ctrl vs. NT4, p = 0.0002. (C) TTX application does not prevent Satb2 induction by BDNF but abolishes Satb2 up-regulation by synaptic activity. Representative image of immunoblot analysis (top) and quantification of the Satb2 protein level (bottom) are shown, n = 3–7; ANOVA followed by Hochberg post hoc test, F4,22 = 12.5, Ctrl vs. BDNF, p = 0.002, Ctrl vs. BDNF+ TTX, p = 0.001, Ctrl vs. Bic/4AP, p = 0.0004, Bic/4AP vs. Bic/4AP+TTX, p = 0.002. Treatments with Bic/4AP also contained MK-801. (D) Treatment with the Trk inhibitor K252a completely blocks the up-regulation of Satb2 by both BDNF and Bic/4AP Representative image of Western blot (top) and quantification of the Satb2 protein level (bottom) are shown, n = 4–6; ANOVA followed by Hochberg post hoc test, F4,20 = 15.6 Ctrl vs. BDNF, p = 0.001, Ctrl vs. Bic/4AP, p = 0.0001, BDNF vs. BDNF+K252a, p = 0.002, Bic/4AP vs. Bic/4AP+K252a, p = 0.0002. Treatment with Bic/4AP also contained MK-801. (E) Blockade of the ERK1/2 signaling pathway with UO126 inhibits the induction of Satb2 by BDNF. Representative image of Western blot (top) and quantification of Satb2 protein levels (bottom) are shown, n = 4; ANOVA followed by Tukey post hoc test, F3,12 = 26.7, Ctrl vs. BDNF, p = 0.003; BDNF vs. BDNF+UO126, p = 0.001. (F) Inhibition of ERK1/2-downstream kinase MSK1 prevents BDNF-induced Satb2 up-regulation. Representative image of Western blot (top) and quantification (bottom) are shown, n = 4–5; ANOVA followed by Hochberg post hoc test, F2,11 = 11.4, Ctrl vs. BDNF, p = 0.002, BDNF vs. BDNF+SB747651A, p = 0.018. In (A–F), data are presented as mean ± SEM of the indicated number of experiments, n values refer to the number of independent hippocampal cultures, *p < 0.05; **p < 0.01; ***p < 0.001, compared with Ctrl; #p < 0.05; ##p < 0.01; ###p < 0.001, compared with BDNF; +p < 0.05; ++p < 0.01; +++p < 0.001, compared with Bic/4AP.DOI: http://dx.doi.org/10.7554/eLife.17361.00810.7554/eLife.17361.009Figure 4—figure supplement 1.BDNF-induced Satb2 expression requires gene transcription.Representative image of Western blot (top) and quantification (bottom) of Satb2 protein levels 24 hr after stimulation with BDNF in the presence or absence of actinomycin D, (n = 4, ANOVA followed by Tukey post hoc test, F2,9 = 70.7, Ctrl vs. BDNF, p = 0.000009, BDNF vs. BDNF+ ActD, p = 0.000006). Data are presented as mean ± SEM.DOI: http://dx.doi.org/10.7554/eLife.17361.00910.7554/eLife.17361.010Figure 4—figure supplement 2.Time-course analysis of Satb2 expression after BDNF treatment.(A) Representative image of Western blot (top) and quantification (bottom) showing that Satb2 protein is significantly induced 6 hr and peaks at around 12 hr after BDNF application, (n = 4; ANOVA followed by Tukey post hoc test, F5,18 = 15.3, Ctrl vs. BDNF 12 hr, p = 0.004, Ctrl vs. BDNF 6 hr, p = 0.0001, Ctrl vs. BDNF 24 hr, p = 0.00006). (B) Representative image of immunoblot analysis (top) and quantification (bottom) of Satb2 protein level following treatment with K252a to block TrkB signaling. Satb2 was induced by a 24 hr period of BDNF stimulation before K252a application (n = 4; ANOVA followed by Tukey post hoc test, F4,15 = 7.6, Ctrl vs. BDNF, p = 0.001, Ctrl vs. BDNF+K252a (6 hr), p = 0.021, Ctrl vs. BDNF+K252a (12 hr), p = 0.045, Ctrl vs. BDNF+K252a (24 hr), p = 0.022). Data in A and B are presented as mean ± SEM of the indicated number of experiments, n values refer to number of independent hippocampal cultures, *p < 0.05; **p < 0.01; ***p < 0.001, compared with Ctrl; #p < 0.05; ##p < 0.01; ###p < 0.001, compared with BDNF.DOI: http://dx.doi.org/10.7554/eLife.17361.010 | review | 26.75 |
(A) Increased synaptic activity up-regulates Satb2 protein depending on calcium influx through L-type VGCC. Representative Western blot (top) and quantification (bottom) of the Satb2 protein level 24 hr after Bic/4AP treatment in the presence or absence of L-VGCC blocker nimodipine or NMDAR antagonist MK-801 (n = 10, 6, 7, 5; ANOVA followed by Hochberg post hoc test; F3,24 = 9.171, Ctrl vs. Bic/4AP, p = 0.002; Ctrl vs. Bic/4AP+MK-801, p = 0.006; Bic/4AP vs. Bic/4AP+Nim, p = 0.011). (B) BDNF and NT4 significantly increase Satb2 protein 24 hr after treatment. Representative Western blot image (top) and quantification of the Satb2 protein level (bottom) are shown, n = 4; ANOVA followed by Tukey post hoc test, F4,15 = 15.4, Ctrl vs. BDNF, p = 0.0004; Ctrl vs. NT4, p = 0.0002. (C) TTX application does not prevent Satb2 induction by BDNF but abolishes Satb2 up-regulation by synaptic activity. Representative image of immunoblot analysis (top) and quantification of the Satb2 protein level (bottom) are shown, n = 3–7; ANOVA followed by Hochberg post hoc test, F4,22 = 12.5, Ctrl vs. BDNF, p = 0.002, Ctrl vs. BDNF+ TTX, p = 0.001, Ctrl vs. Bic/4AP, p = 0.0004, Bic/4AP vs. Bic/4AP+TTX, p = 0.002. Treatments with Bic/4AP also contained MK-801. (D) Treatment with the Trk inhibitor K252a completely blocks the up-regulation of Satb2 by both BDNF and Bic/4AP Representative image of Western blot (top) and quantification of the Satb2 protein level (bottom) are shown, n = 4–6; ANOVA followed by Hochberg post hoc test, F4,20 = 15.6 Ctrl vs. BDNF, p = 0.001, Ctrl vs. Bic/4AP, p = 0.0001, BDNF vs. BDNF+K252a, p = 0.002, Bic/4AP vs. Bic/4AP+K252a, p = 0.0002. Treatment with Bic/4AP also contained MK-801. (E) Blockade of the ERK1/2 signaling pathway with UO126 inhibits the induction of Satb2 by BDNF. Representative image of Western blot (top) and quantification of Satb2 protein levels (bottom) are shown, n = 4; ANOVA followed by Tukey post hoc test, F3,12 = 26.7, Ctrl vs. BDNF, p = 0.003; BDNF vs. BDNF+UO126, p = 0.001. (F) Inhibition of ERK1/2-downstream kinase MSK1 prevents BDNF-induced Satb2 up-regulation. Representative image of Western blot (top) and quantification (bottom) are shown, n = 4–5; ANOVA followed by Hochberg post hoc test, F2,11 = 11.4, Ctrl vs. BDNF, p = 0.002, BDNF vs. BDNF+SB747651A, p = 0.018. In (A–F), data are presented as mean ± SEM of the indicated number of experiments, n values refer to the number of independent hippocampal cultures, *p < 0.05; **p < 0.01; ***p < 0.001, compared with Ctrl; #p < 0.05; ##p < 0.01; ###p < 0.001, compared with BDNF; +p < 0.05; ++p < 0.01; +++p < 0.001, compared with Bic/4AP. | clinical case | 29.66 |
10.7554/eLife.17361.009Figure 4—figure supplement 1.BDNF-induced Satb2 expression requires gene transcription.Representative image of Western blot (top) and quantification (bottom) of Satb2 protein levels 24 hr after stimulation with BDNF in the presence or absence of actinomycin D, (n = 4, ANOVA followed by Tukey post hoc test, F2,9 = 70.7, Ctrl vs. BDNF, p = 0.000009, BDNF vs. BDNF+ ActD, p = 0.000006). Data are presented as mean ± SEM.DOI: http://dx.doi.org/10.7554/eLife.17361.009 | study | 30.44 |
Representative image of Western blot (top) and quantification (bottom) of Satb2 protein levels 24 hr after stimulation with BDNF in the presence or absence of actinomycin D, (n = 4, ANOVA followed by Tukey post hoc test, F2,9 = 70.7, Ctrl vs. BDNF, p = 0.000009, BDNF vs. BDNF+ ActD, p = 0.000006). Data are presented as mean ± SEM. | other | 31.81 |
10.7554/eLife.17361.010Figure 4—figure supplement 2.Time-course analysis of Satb2 expression after BDNF treatment.(A) Representative image of Western blot (top) and quantification (bottom) showing that Satb2 protein is significantly induced 6 hr and peaks at around 12 hr after BDNF application, (n = 4; ANOVA followed by Tukey post hoc test, F5,18 = 15.3, Ctrl vs. BDNF 12 hr, p = 0.004, Ctrl vs. BDNF 6 hr, p = 0.0001, Ctrl vs. BDNF 24 hr, p = 0.00006). (B) Representative image of immunoblot analysis (top) and quantification (bottom) of Satb2 protein level following treatment with K252a to block TrkB signaling. Satb2 was induced by a 24 hr period of BDNF stimulation before K252a application (n = 4; ANOVA followed by Tukey post hoc test, F4,15 = 7.6, Ctrl vs. BDNF, p = 0.001, Ctrl vs. BDNF+K252a (6 hr), p = 0.021, Ctrl vs. BDNF+K252a (12 hr), p = 0.045, Ctrl vs. BDNF+K252a (24 hr), p = 0.022). Data in A and B are presented as mean ± SEM of the indicated number of experiments, n values refer to number of independent hippocampal cultures, *p < 0.05; **p < 0.01; ***p < 0.001, compared with Ctrl; #p < 0.05; ##p < 0.01; ###p < 0.001, compared with BDNF.DOI: http://dx.doi.org/10.7554/eLife.17361.010 | study | 32 |
(A) Representative image of Western blot (top) and quantification (bottom) showing that Satb2 protein is significantly induced 6 hr and peaks at around 12 hr after BDNF application, (n = 4; ANOVA followed by Tukey post hoc test, F5,18 = 15.3, Ctrl vs. BDNF 12 hr, p = 0.004, Ctrl vs. BDNF 6 hr, p = 0.0001, Ctrl vs. BDNF 24 hr, p = 0.00006). (B) Representative image of immunoblot analysis (top) and quantification (bottom) of Satb2 protein level following treatment with K252a to block TrkB signaling. Satb2 was induced by a 24 hr period of BDNF stimulation before K252a application (n = 4; ANOVA followed by Tukey post hoc test, F4,15 = 7.6, Ctrl vs. BDNF, p = 0.001, Ctrl vs. BDNF+K252a (6 hr), p = 0.021, Ctrl vs. BDNF+K252a (12 hr), p = 0.045, Ctrl vs. BDNF+K252a (24 hr), p = 0.022). Data in A and B are presented as mean ± SEM of the indicated number of experiments, n values refer to number of independent hippocampal cultures, *p < 0.05; **p < 0.01; ***p < 0.001, compared with Ctrl; #p < 0.05; ##p < 0.01; ###p < 0.001, compared with BDNF. | other | 32.53 |
Since BDNF expression and secretion in the CNS are controlled by neuronal activity we reasoned that the induction of Satb2 following AP bursting might be due to enhanced Bdnf transcription, translation and/or BDNF release (Hardingham et al., 2002; Kuczewski et al., 2009; Tao et al., 1998). To further investigate this possibility we inhibited Trk signaling with K252a during Bic/4AP stimulation. As a control experiment, we pharmacologically blocked AP bursting with the sodium channel blocker tetrodotoxin (TTX) during BDNF treatment. We found that silencing the neuronal activity with TTX did not affect BDNF-induced Satb2 expression, even though it abolished the synaptic activity-driven Satb2 induction (Figure 4C). In contrast, synaptic activity-triggered increase of Satb2 was blocked by Trk antagonism (Figure 4D), suggesting that synaptic activity-triggered Satb2 induction is indeed mediated via BDNF/TrkB signaling. De novo gene transcription is necessary for this process since actinomycin D, an inhibitor of gene transcription, blocked BDNF-driven Satb2 up-regulation (Figure 4—figure supplement 1). | other | 26.83 |
To determine if MEK1/2 - ERK1/2 pathway (Minichiello, 2009) contributes to BDNF-triggered Satb2 induction we applied the MEK1/2 inhibitor UO126 1 hr prior to BDNF stimulation. UO126 blocked the induction of Satb2 by BDNF (Figure 4E). Furthermore, inhibition of mitogen/stress-activated kinase 1 (MSK1), a major regulator of activity- and experience-dependent synaptic adaptation downstream of MEK1/2 (Corrêa et al., 2012), had the same effect as the inhibition of MEK1/2 (Figure 4F), indicating that in hippocampal neurons BDNF up-regulates Satb2 through a pathway that requires ERK1/2 and MSK1. | other | 29.6 |
Finally, we examined the temporal pattern of Satb2 induction by BDNF. Satb2 protein was increased at 6 hr, reached a maximum within 12–24 hr, and remained at this level for 48 hr following BDNF stimulation (Figure 4—figure supplement 2A). Loss of Satb2 occurred with a similar kinetic after antagonizing Trk receptor signaling in cultures previously stimulated with BDNF for 24 hr (Figure 4—figure supplement 2B). The relatively slow kinetics of Satb2 induction and elimination in hippocampal neurons is consistent with a potential role of Satb2 in slow, long-lasting adaptive neuronal processes (Zagrebelsky and Korte, 2014). | other | 30.11 |
To gain insights into the molecular mechanisms by which Satb2 contributes to LTP maintenance and memory consolidation we mapped Satb2 genomic binding sites by ChIP-seq. Mouse primary hippocampal neurons were transduced with an AAV encoding V5-tagged Satb2 and a V5 antibody was used for chromatin immunoprecipitation. The specificity of the anti-V5 antibody was verified by ChIP-qPCR in non-transduced primary neurons (Figure 5—figure supplement 1). Out of 8414 Satb2 binding sites identified, 4496 were located within less than 1 kb distance from a transcriptional start site (TSS), indicative of Satb2 enrichment at proximal promoters (Figure 5A and B). Promoter sequences in eukaryotic genomes are marked by histone tail modifications associated with active or inactive state of the downstream gene (Barski et al., 2007; Wang et al., 2008). To examine the chromatin states of Satb2-bound promoters we compared our Satb2 ChIP-seq data with previously published datasets reporting genome-wide histone modifications and RNA Pol II recruitment in cortical neurons (GSE63271, GSE66701 and GSE21161) or in hippocampal tissue (GSE65159). Statistical testing of the overlap among the peaks of these datasets and our Satb2 ChIP-seq data revealed highly significant enrichment (adjusted p<0.00024) of active chromatin states (H4K16ac, H3K27ac, H3K4me1, H3K4me2, and H3K4me3 peaks) and PolII peaks at Satb2-bound promoters (Figure 5C, Figure 5—figure supplement 2A). Conversely, Satb2 genome occupancy and the Polycomb-associated H3K27me3 repressive mark showed no correlation (Figure 5—figure supplement 2B). GO enrichment analysis of the genes with Satb2 peaks within their promoters revealed significant over-representation of the following GO categories: ‘transcription factor activity’, ‘transcription corepressor activity’, ‘chromatin remodelling complex’, ‘dendritic spine’ and the KEGG pathway ‘long-term potentiation’ (Figure 5—source data 1)10.7554/eLife.17361.011Figure 5.Satb2 binding sites are enriched on active gene promoters including miRNA promoters.(A) Pie-chart illustrating the genomic annotation of Satb2 binding sites. (B) Heatmap of Satb2 binding to TSS (±3 Kb) regions. (C) Average profiles of H4K16ac, H3K27ac, H3K4me1, H3K4me2, H3K4me3 and PolII peaks (GEO: GSE63271, GSE66701, GSE21161, and GSE65159) at Satb2 bound promoters. (D) Average tag density profiles (ChIP/Input, left panel) and heat map depicting Satb2 ChIP-seq tag density at predicted miRNA promoter regions (right panel). ‘L’ – 5’ left, ‘R’ – 3’ right of the miRNA promoters. The tick marks represent distance of −3 kb, −1.5 kb, +1.5 kb, +3 kb relative to the miR promoters.DOI: http://dx.doi.org/10.7554/eLife.17361.01110.7554/eLife.17361.012Figure 5—source data 1.Major GO terms and KEGG pathways, revealed by ChIP-Enrich bioinformatics tool, found to be enriched among the genes having Satb2 peaks within their promoters.DOI: http://dx.doi.org/10.7554/eLife.17361.01210.7554/eLife.17361.013Figure 5—figure supplement 1.ChIP-qPCR validation of Satb2 enrichment at various identified target regions using chromatin from AAV-Satb2-V5-transduced and non-transduced primary hippocampal neurons.Each bar represents the ratio of precipitated DNA (bound) to the total input DNA (% Input). Each target region was tested in at least two independent chromatin samples derived from non-transduced hippocampal cultures and in at least one ChIP sample from AAV-Satb2-V5-transduced neurons.DOI: http://dx.doi.org/10.7554/eLife.17361.01310.7554/eLife.17361.014Figure 5—figure supplement 2.Satb2 binding sites are enriched on active gene promoters and do not correlate with the Polycomb-associated H3K27me3 repressive mark (PcR).(A) Heat maps depicting H3K4me3 (GSM530197), H3K4me2 (GSM1544912), H3K4me1 (GSM1544908), H3K27Ac (GSM1629392), H4K16Ac (GSM1629377), PolII (GSM1544942), and Satb2 tag densities (read counts/million mapped reads) at TSS/genebody/TES (±3 Kb) regions. (B) Average profile and heat maps of Satb2 genome occupancy centered on TSS (left panel) and the Polycomb-associated H3K27me3 peaks (GSE65159, PcR, ±3 Kb) (right panel).DOI: http://dx.doi.org/10.7554/eLife.17361.01410.7554/eLife.17361.015Figure 5—figure supplement 3.Satb2 is deposited at CpGs.Average tag density profiles (ChIP/Input) and heat maps depicting Satb2 ChIP-seq tag density at CpG islands localized at proximal promoters (A), gene body (B), intergenic regions (C), and miRNA proximal promoters (D), ‘L’ – 5’ left, ‘R’ – 3’ right of the CpG islands, the tick marks represent distance of −3 kb, −1.5 kb, +1.5 kb, +3 kb relative to the CpG islands.DOI: http://dx.doi.org/10.7554/eLife.17361.01510.7554/eLife.17361.016Figure 5—figure supplement 4.ChIP-qPCR validation of Satb2 targets in vivo.ChIP-qPCR analysis of chromatin derived from CA1 hippocampal tissue of control and Satb2 cKO mice immunoprecipitated with an antibody to Satb2 at several target sites identified by ChIP-seq in primary hippocampal cultures. Each bar represents the ratio of precipitated DNA (bound) to the total input DNA (% Input) of the region tested, normalized to the % Input in ChIP samples derived from Satb2 cKO mice. Each target region was tested in at least two independent chromatin samples (a pool of CA1 tissue from 8–10 mice).DOI: http://dx.doi.org/10.7554/eLife.17361.016 | other | 26.61 |
(A) Pie-chart illustrating the genomic annotation of Satb2 binding sites. (B) Heatmap of Satb2 binding to TSS (±3 Kb) regions. (C) Average profiles of H4K16ac, H3K27ac, H3K4me1, H3K4me2, H3K4me3 and PolII peaks (GEO: GSE63271, GSE66701, GSE21161, and GSE65159) at Satb2 bound promoters. (D) Average tag density profiles (ChIP/Input, left panel) and heat map depicting Satb2 ChIP-seq tag density at predicted miRNA promoter regions (right panel). ‘L’ – 5’ left, ‘R’ – 3’ right of the miRNA promoters. The tick marks represent distance of −3 kb, −1.5 kb, +1.5 kb, +3 kb relative to the miR promoters. | other | 33.3 |
10.7554/eLife.17361.014Figure 5—figure supplement 2.Satb2 binding sites are enriched on active gene promoters and do not correlate with the Polycomb-associated H3K27me3 repressive mark (PcR).(A) Heat maps depicting H3K4me3 (GSM530197), H3K4me2 (GSM1544912), H3K4me1 (GSM1544908), H3K27Ac (GSM1629392), H4K16Ac (GSM1629377), PolII (GSM1544942), and Satb2 tag densities (read counts/million mapped reads) at TSS/genebody/TES (±3 Kb) regions. (B) Average profile and heat maps of Satb2 genome occupancy centered on TSS (left panel) and the Polycomb-associated H3K27me3 peaks (GSE65159, PcR, ±3 Kb) (right panel).DOI: http://dx.doi.org/10.7554/eLife.17361.014 | study | 31.84 |
(A) Heat maps depicting H3K4me3 (GSM530197), H3K4me2 (GSM1544912), H3K4me1 (GSM1544908), H3K27Ac (GSM1629392), H4K16Ac (GSM1629377), PolII (GSM1544942), and Satb2 tag densities (read counts/million mapped reads) at TSS/genebody/TES (±3 Kb) regions. (B) Average profile and heat maps of Satb2 genome occupancy centered on TSS (left panel) and the Polycomb-associated H3K27me3 peaks (GSE65159, PcR, ±3 Kb) (right panel). | other | 30.3 |
10.7554/eLife.17361.013Figure 5—figure supplement 1.ChIP-qPCR validation of Satb2 enrichment at various identified target regions using chromatin from AAV-Satb2-V5-transduced and non-transduced primary hippocampal neurons.Each bar represents the ratio of precipitated DNA (bound) to the total input DNA (% Input). Each target region was tested in at least two independent chromatin samples derived from non-transduced hippocampal cultures and in at least one ChIP sample from AAV-Satb2-V5-transduced neurons.DOI: http://dx.doi.org/10.7554/eLife.17361.013 | study | 31.44 |
Each bar represents the ratio of precipitated DNA (bound) to the total input DNA (% Input). Each target region was tested in at least two independent chromatin samples derived from non-transduced hippocampal cultures and in at least one ChIP sample from AAV-Satb2-V5-transduced neurons. | other | 32.28 |
10.7554/eLife.17361.015Figure 5—figure supplement 3.Satb2 is deposited at CpGs.Average tag density profiles (ChIP/Input) and heat maps depicting Satb2 ChIP-seq tag density at CpG islands localized at proximal promoters (A), gene body (B), intergenic regions (C), and miRNA proximal promoters (D), ‘L’ – 5’ left, ‘R’ – 3’ right of the CpG islands, the tick marks represent distance of −3 kb, −1.5 kb, +1.5 kb, +3 kb relative to the CpG islands.DOI: http://dx.doi.org/10.7554/eLife.17361.015 | study | 32.53 |
Average tag density profiles (ChIP/Input) and heat maps depicting Satb2 ChIP-seq tag density at CpG islands localized at proximal promoters (A), gene body (B), intergenic regions (C), and miRNA proximal promoters (D), ‘L’ – 5’ left, ‘R’ – 3’ right of the CpG islands, the tick marks represent distance of −3 kb, −1.5 kb, +1.5 kb, +3 kb relative to the CpG islands. | other | 33.97 |
Given that the majority of annotated gene promoters are associated with CpG islands (Deaton and Bird, 2011) we examined if Satb2 localizes to CpG islands. ChIP-seq tag distribution profiles revealed that Satb2 was deposited at CpGs associated with proximal promoters, intragenic (gene body associated) or intergenic CpGs (Figure 5—figure supplement 3A–C). Evidence suggests that most of the latter two CpG classes represent alternative promoters of nearby annotated genes or TSSs for non-coding RNAs (Monteys et al., 2010; Wang et al., 2010). Indeed, we found Satb2 enrichment on miRNA-associated CpGs (Figure 5—figure supplement 3D). | other | 26.81 |
To further explore if Satb2 binds to miRNA TSSs, we used a recently developed miRNA promoter prediction method (Marsico et al., 2013) to assess potential enrichment of Satb2 on miRNA promoters. The miRNA TSS identification algorithm that we applied detected at least one TSS for about 82% of the miRBase-annotated miRNAs (Marsico et al., 2013). Furthermore, it is particularly suited for detection of intronic miRNA promoters, which often act independently from the host gene promoters (Monteys et al., 2010). Our analysis revealed significant association of Satb2 with miRNA promoters (Figure 5D). | other | 27.86 |
10.7554/eLife.17361.016Figure 5—figure supplement 4.ChIP-qPCR validation of Satb2 targets in vivo.ChIP-qPCR analysis of chromatin derived from CA1 hippocampal tissue of control and Satb2 cKO mice immunoprecipitated with an antibody to Satb2 at several target sites identified by ChIP-seq in primary hippocampal cultures. Each bar represents the ratio of precipitated DNA (bound) to the total input DNA (% Input) of the region tested, normalized to the % Input in ChIP samples derived from Satb2 cKO mice. Each target region was tested in at least two independent chromatin samples (a pool of CA1 tissue from 8–10 mice).DOI: http://dx.doi.org/10.7554/eLife.17361.016 | study | 32.06 |
ChIP-qPCR analysis of chromatin derived from CA1 hippocampal tissue of control and Satb2 cKO mice immunoprecipitated with an antibody to Satb2 at several target sites identified by ChIP-seq in primary hippocampal cultures. Each bar represents the ratio of precipitated DNA (bound) to the total input DNA (% Input) of the region tested, normalized to the % Input in ChIP samples derived from Satb2 cKO mice. Each target region was tested in at least two independent chromatin samples (a pool of CA1 tissue from 8–10 mice). | other | 32.3 |
To study the role of Satb2 in transcriptional control in vivo, we performed global transcriptome profiling analyses (RNA-seq and small RNA-seq) using CA1 hippocampal tissue from Satb2 cKO mice and littermate controls. RNA-seq analysis identified a number of protein-coding genes that were differentially expressed between control and Satb2 cKO mice (25 up-regulated and 15 down-regulated, Figure 6—source data 1). Amongst them we found genes that have previously been identified as highly relevant for learning and memory or directly implicated in memory formation such as Adra2a, Penk, Htr5b, and Ghsr (Diano et al., 2006; Galeotti et al., 2004; Ghersi et al., 2015; Peppin and Raffa, 2015). Pathway analysis revealed significant enrichment (p=0.018) of the ‘neuroactive ligand-receptor interaction pathway’ amongst the regulated genes. The differential expression of selected genes, including Adra2a, Penk, Htr5b, and Ghsr was validated by real-time qPCR (Figure 6—figure supplement 1). Notably, Satb2 was bound to the promoters of the majority of the identified differentially expressed genes as shown by Satb2 ChIP-seq (Figure 6—source data 1). We also investigated differential splicing between the CA1 transcriptomes of Satb2 cKO mice and littermate controls, since global de-regulation of RNA-splicing has been linked to impaired synaptic plasticity and memory function (Benito et al., 2015). However, no major difference in RNA-spicing was observed comparing Satb2 cKO and control mice (Figure 6—source data 2). | other | 28.83 |
Next, we performed small RNA-seq analysis to examine miRNA transcriptome changes in the CA1 hippocampal area of Satb2 cKO vs control mice. We detected the expression of 476 miRNAs in the CA1 tissue, similar to the number reported for this hippocampal subfield in a previous study (Stilling et al., 2014). Of these 464 miRNAs 43.9% showed significant differential expression between Satb2 cKO and control mice (Figure 6A, Figure 6—source data 3). Principal component analysis, a method for visualizing gene expression patterns, revealed a clear separation between Satb2 cKO and control samples (Figure 6B). Furthermore, hierarchical clustering using a Pearson correlation-based method demonstrated two main clusters of miRNAs (up- and down regulated) based on their expression levels in the CA1 field of Satb2 cKO vs control mice (Figure 6C). The differential expression of selected miRNAs in the CA1 tissue of control vs. Satb2 cKO mice was validated by qPCR (Figure 6—figure supplement 2). 44.4% of the miRNAs found to be deregulated in Satb2 cKO mice had at least one predicted promoter occupied by Satb2 as demonstrated by Satb2 Chip-seq data (Figure 6—figure supplement 3 and Figure 6—source data 4). Moreover, in vivo Chip-qPCR analysis using an antibody against Satb2 and chromatin from CA1 tissue demonstrated that Satb2 can bind to the miR-22 promoter (Figure 5—figure supplement 4). Importantly, amongst the miRNAs identified as deregulated in Satb2 cKO mice there were miRNAs with well-documented synaptic regulatory functions, including miR-125b, miR-132, miR-212, miR-124 (McNeill and Van Vactor, 2012), or miRNAs which have been directly implicated in learning and memory, such as miR-132 or miR-124 (Hansen et al., 2010; Malmevik et al., 2016).10.7554/eLife.17361.017Figure 6.Satb2 regulates miRNA expression in CA1 hippocampal area.(A) Pie-chart showing the percentage of differentially expressed miRNAs (up-and down-regulated, Figure 6—source data 1) in the CA1 region of Satb2 cKO mice vs. littermate controls as assessed by small RNA-seq analysis (base mean above 10 counts, 1.5-fold change, and adjusted p < 0.05). (B) PCA plot of miRNA counts analyzed by small RNA-seq of Satb2 cKO vs. control CA1 hippocampal tissue. The first two PCs explained 33.3% and 21.4% of the variance, respectively. (C) Heat map from hierarchical clustering of differentially expressed miRNAs (base mean above 100 counts) in the CA1 hippocampal area of Satb2 cKO mice (n = 9) vs. littermate controls (n = 9).DOI: http://dx.doi.org/10.7554/eLife.17361.01710.7554/eLife.17361.018Figure 6—source data 1.Differentially expressed genes between Satb2 cKO mice and Satb2flox/flox littermate controls in the CA1 region as assessed by RNA-seq.Shown are the genomic coordinates of the identified Satb2 binding sites and the distance to the TSS of the corresponding gene.DOI: http://dx.doi.org/10.7554/eLife.17361.01810.7554/eLife.17361.019Figure 6—source data 2.No major significant differences in splicing were observed between the CA1 region of Satb2 cKO mice (n = 9) and littermate controls (n = 9).Only one gene showed differential exon usage at FDR corrected p-value significance level 0.01 and 12 genes at FDR 0.1 (based on DEXSeq R-package with default parameters).DOI: http://dx.doi.org/10.7554/eLife.17361.01910.7554/eLife.17361.020Figure 6—source data 3.List of differentially expressed miRNAs between Satb2 cKO mice and littermate controls (Satb2flox/floxmice) in the CA1 region as assessed by small RNA-seq analysis (base mean above 10 counts, 1.5-fold change, and adjusted p < 0.05).DOI: http://dx.doi.org/10.7554/eLife.17361.02010.7554/eLife.17361.021Figure 6—source data 4.List of differentially expressed miRNAs in the CA1 region of Satb2 cKO mice vs. littermate controls that contain Satb2 peak(s) within their promoter.DOI: http://dx.doi.org/10.7554/eLife.17361.02110.7554/eLife.17361.022Figure 6—figure supplement 1.Validation of the differential expression of selected genes by qPCR.qPCR quantification of selected gene mRNA levels in the CA1 field of Satb2 cKO mice compared to the control group (Oxr1: Ctrl, n = 11; cKO, n = 11; Ghsr: Ctrl, n = 10; cKO, n = 10; Adra2a: Ctrl, n = 8; cKO, n = 12; Penk: Ctrl, n = 8; cKO, n = 12; Htr5b: Ctrl, n = 8; cKO, n = 12; Sema3A: Ctrl, n = 5; cKO, n = 6; Crym: Ctrl, n = 5; cKO, n = 6; Nt5dc3: Ctrl, n = 10; cKO, n = 10; Zfp423: Ctrl, n = 10; cKO, n = 10; Nrip3: Ctrl, n = 10; cKO, n = 10; Student's t test, Oxr1: t24 = 7.978, p < 0.0001; Ghsr: t18 = 2.648, p = 0.0164; Adra2a: t18 = 3.269, p = 0.0043; Penk: t18 = 3.9634, p = 0.0009; Htr5b: t18 = 3.5517, p = 0.0023; Calb1: t18 = 3.435, p = 0.0030; Sema3A: t9 = 4.5756, p = 0.0013; Crym: t9 = 3.4984, p = 0.0067; Nt5dc3: t18 = 4.5276, p = 0.0003; Zfp423: t18 = 2.5701, p = 0.0193; Nrip3: t18 = 3.7965, p = 0.0013), ***p < 0.001, **p < 0.01, *p < 0.05. Data are presented as mean ± SEM.DOI: http://dx.doi.org/10.7554/eLife.17361.02210.7554/eLife.17361.023Figure 6—figure supplement 2.Validation of the differential expression of selected miRNAs by qPCR.qPCR quantification of miRNA levels in the CA1 field of Satb2 cKO mice compared to the control group (miR-124: Ctrl, n = 9; cKO, n = 9; miR-125b: Ctrl, n = 9; cKO, n = 9; miR-22: Ctrl, n = 7; cKO, n = 9; miR-143: Ctrl, n = 9; cKO, n = 9). Student's t test, miR-124: t = 4.391, p = 0.005; miR-125b: t18 = 4.876, p = 0.0002; miR-22: t18 = 3.806, p = 0.0019; miR-143: t18 = 1.611, p = 0.1266). Data are presented as mean ± SEM.DOI: http://dx.doi.org/10.7554/eLife.17361.02310.7554/eLife.17361.024Figure 6—figure supplement 3.Satb2 binds to miRNA’s promoters.IGV genome browser images revealing overlapping between Satb2 peaks and predicted promoters (depicted as red rectangles) of selected miRNAs: miR124a (a), miR125a (b), mir132 (c), mir182 (d), found to be deregulated in Satb2 cKO mice. The complete list of all differentially expressed miRNA with Satb2 peak(s) within their promoter is shown in Figure 6—source data 4.DOI: http://dx.doi.org/10.7554/eLife.17361.024 | other | 27.64 |
(A) Pie-chart showing the percentage of differentially expressed miRNAs (up-and down-regulated, Figure 6—source data 1) in the CA1 region of Satb2 cKO mice vs. littermate controls as assessed by small RNA-seq analysis (base mean above 10 counts, 1.5-fold change, and adjusted p < 0.05). (B) PCA plot of miRNA counts analyzed by small RNA-seq of Satb2 cKO vs. control CA1 hippocampal tissue. The first two PCs explained 33.3% and 21.4% of the variance, respectively. (C) Heat map from hierarchical clustering of differentially expressed miRNAs (base mean above 100 counts) in the CA1 hippocampal area of Satb2 cKO mice (n = 9) vs. littermate controls (n = 9). | other | 34.28 |
10.7554/eLife.17361.018Figure 6—source data 1.Differentially expressed genes between Satb2 cKO mice and Satb2flox/flox littermate controls in the CA1 region as assessed by RNA-seq.Shown are the genomic coordinates of the identified Satb2 binding sites and the distance to the TSS of the corresponding gene.DOI: http://dx.doi.org/10.7554/eLife.17361.018 | study | 32.47 |
To confirm that endogenous Satb2 in the adult hippocampus binds to the same regions identified by ChIP-seq in hippocampal cultures we performed ChIP-qPCR using a Satb2-specific antibody and chromatin derived from control and Satb2 cKO CA1 hippocampal tissue. The results revealed Satb2 enrichment at various identified target promoters and/or Satb2 binding sites in chromatin samples from control but not Satb2 cKO mice, thus validating the in vitro Satb2 genomic binding patterns (Figure 5—figure supplement 4). | other | 30.5 |
Taken together, our results provide evidence for association of Satb2 with active promoter regulatory sequences in the genome of hippocampal neurons, including miRNA promoters, suggesting a potential role of Satb2 in the transcription of active neuronal chromatin. | other | 32.8 |
10.7554/eLife.17361.019Figure 6—source data 2.No major significant differences in splicing were observed between the CA1 region of Satb2 cKO mice (n = 9) and littermate controls (n = 9).Only one gene showed differential exon usage at FDR corrected p-value significance level 0.01 and 12 genes at FDR 0.1 (based on DEXSeq R-package with default parameters).DOI: http://dx.doi.org/10.7554/eLife.17361.019 | study | 31.3 |
10.7554/eLife.17361.020Figure 6—source data 3.List of differentially expressed miRNAs between Satb2 cKO mice and littermate controls (Satb2flox/floxmice) in the CA1 region as assessed by small RNA-seq analysis (base mean above 10 counts, 1.5-fold change, and adjusted p < 0.05).DOI: http://dx.doi.org/10.7554/eLife.17361.020 | study | 32.88 |
10.7554/eLife.17361.023Figure 6—figure supplement 2.Validation of the differential expression of selected miRNAs by qPCR.qPCR quantification of miRNA levels in the CA1 field of Satb2 cKO mice compared to the control group (miR-124: Ctrl, n = 9; cKO, n = 9; miR-125b: Ctrl, n = 9; cKO, n = 9; miR-22: Ctrl, n = 7; cKO, n = 9; miR-143: Ctrl, n = 9; cKO, n = 9). Student's t test, miR-124: t = 4.391, p = 0.005; miR-125b: t18 = 4.876, p = 0.0002; miR-22: t18 = 3.806, p = 0.0019; miR-143: t18 = 1.611, p = 0.1266). Data are presented as mean ± SEM.DOI: http://dx.doi.org/10.7554/eLife.17361.023 | study | 33.5 |
qPCR quantification of miRNA levels in the CA1 field of Satb2 cKO mice compared to the control group (miR-124: Ctrl, n = 9; cKO, n = 9; miR-125b: Ctrl, n = 9; cKO, n = 9; miR-22: Ctrl, n = 7; cKO, n = 9; miR-143: Ctrl, n = 9; cKO, n = 9). Student's t test, miR-124: t = 4.391, p = 0.005; miR-125b: t18 = 4.876, p = 0.0002; miR-22: t18 = 3.806, p = 0.0019; miR-143: t18 = 1.611, p = 0.1266). Data are presented as mean ± SEM. | other | 32.03 |
10.7554/eLife.17361.026Figure 7—source data 1.List of miRNAs predicted to target mouse Arc 3’UTR by the bioinformatics tools TargetScan, PITA and miRanda.The miRNAs identified as up-regulated in the CA1 region of Satb2 cKO vs. control mice are shown in red.DOI: http://dx.doi.org/10.7554/eLife.17361.026 | study | 32.3 |
To investigate if restoring Satb2 expression can rescue the observed reduction in Arc level and/or the impaired long-term fear memory of Satb2 cKO mice, we used rAAV-mediated gene delivery to re-introduce Satb2 into the hippocampus of Satb2 mutants. Viral vectors encoding either Satb2 (rAAV8-hSyn-Satb2-V5) or GFP (rAAV8-hSyn-GFP) were injected into the dorsal hippocampus of Satb2 cKO mice four weeks prior to immunoblotting analysis or contextual fear conditioning training. Littermate Satb2flox/flox mice, injected with rAAV8-hSyn-GFP, served as controls. Immunohistochemistry analysis confirmed robust and specific expression of V5-tagged Satb2 in the hippocampal CA1 region of conditional mutants after stereotaxic injection of rAAV8-hSyn-Satb2-V5 (Figure 7C). Immunoblotting results revealed that rAAV8-hSyn-Satb2-V5 injection was able to restore Arc protein levels in the CA1 area of Satb2 cKO mice to control levels (Figure 7D). Moreover, the reinstatement of Satb2 in the dorsal hippocampus not only reversed the decreased Arc protein level, but also rescued the long-term contextual fear memory impairment. We observed indistinguishable freezing behavior between control and rAAV8-hSyn-Satb2-V5-transduced Satb2 cKO mice 24 hr after training (Figure 7E). By contrast, Satb2 cKO mice injected with rAAV8-hSyn-GFP showed significantly reduced freezing, reproducing our initial findings with un-injected animals (Figure 2A). | other | 29.73 |
10.7554/eLife.17361.022Figure 6—figure supplement 1.Validation of the differential expression of selected genes by qPCR.qPCR quantification of selected gene mRNA levels in the CA1 field of Satb2 cKO mice compared to the control group (Oxr1: Ctrl, n = 11; cKO, n = 11; Ghsr: Ctrl, n = 10; cKO, n = 10; Adra2a: Ctrl, n = 8; cKO, n = 12; Penk: Ctrl, n = 8; cKO, n = 12; Htr5b: Ctrl, n = 8; cKO, n = 12; Sema3A: Ctrl, n = 5; cKO, n = 6; Crym: Ctrl, n = 5; cKO, n = 6; Nt5dc3: Ctrl, n = 10; cKO, n = 10; Zfp423: Ctrl, n = 10; cKO, n = 10; Nrip3: Ctrl, n = 10; cKO, n = 10; Student's t test, Oxr1: t24 = 7.978, p < 0.0001; Ghsr: t18 = 2.648, p = 0.0164; Adra2a: t18 = 3.269, p = 0.0043; Penk: t18 = 3.9634, p = 0.0009; Htr5b: t18 = 3.5517, p = 0.0023; Calb1: t18 = 3.435, p = 0.0030; Sema3A: t9 = 4.5756, p = 0.0013; Crym: t9 = 3.4984, p = 0.0067; Nt5dc3: t18 = 4.5276, p = 0.0003; Zfp423: t18 = 2.5701, p = 0.0193; Nrip3: t18 = 3.7965, p = 0.0013), ***p < 0.001, **p < 0.01, *p < 0.05. Data are presented as mean ± SEM.DOI: http://dx.doi.org/10.7554/eLife.17361.022 | study | 32.66 |
qPCR quantification of selected gene mRNA levels in the CA1 field of Satb2 cKO mice compared to the control group (Oxr1: Ctrl, n = 11; cKO, n = 11; Ghsr: Ctrl, n = 10; cKO, n = 10; Adra2a: Ctrl, n = 8; cKO, n = 12; Penk: Ctrl, n = 8; cKO, n = 12; Htr5b: Ctrl, n = 8; cKO, n = 12; Sema3A: Ctrl, n = 5; cKO, n = 6; Crym: Ctrl, n = 5; cKO, n = 6; Nt5dc3: Ctrl, n = 10; cKO, n = 10; Zfp423: Ctrl, n = 10; cKO, n = 10; Nrip3: Ctrl, n = 10; cKO, n = 10; Student's t test, Oxr1: t24 = 7.978, p < 0.0001; Ghsr: t18 = 2.648, p = 0.0164; Adra2a: t18 = 3.269, p = 0.0043; Penk: t18 = 3.9634, p = 0.0009; Htr5b: t18 = 3.5517, p = 0.0023; Calb1: t18 = 3.435, p = 0.0030; Sema3A: t9 = 4.5756, p = 0.0013; Crym: t9 = 3.4984, p = 0.0067; Nt5dc3: t18 = 4.5276, p = 0.0003; Zfp423: t18 = 2.5701, p = 0.0193; Nrip3: t18 = 3.7965, p = 0.0013), ***p < 0.001, **p < 0.01, *p < 0.05. Data are presented as mean ± SEM. | other | 33.28 |
10.7554/eLife.17361.024Figure 6—figure supplement 3.Satb2 binds to miRNA’s promoters.IGV genome browser images revealing overlapping between Satb2 peaks and predicted promoters (depicted as red rectangles) of selected miRNAs: miR124a (a), miR125a (b), mir132 (c), mir182 (d), found to be deregulated in Satb2 cKO mice. The complete list of all differentially expressed miRNA with Satb2 peak(s) within their promoter is shown in Figure 6—source data 4.DOI: http://dx.doi.org/10.7554/eLife.17361.024 | study | 35.88 |
IGV genome browser images revealing overlapping between Satb2 peaks and predicted promoters (depicted as red rectangles) of selected miRNAs: miR124a (a), miR125a (b), mir132 (c), mir182 (d), found to be deregulated in Satb2 cKO mice. The complete list of all differentially expressed miRNA with Satb2 peak(s) within their promoter is shown in Figure 6—source data 4. | other | 31.12 |
Among the miRNAs up-regulated in the CA1 of Satb2 cKO mice, 24 miRNAs were predicted to target the 3’UTR of ‘activity–regulated cytoskeletal associated protein’ (Arc) by at least one of the miRNA-target prediction tools TargetScan, miRanda, and PITA (Figure 7—source data 1). Moreover, regulation of Arc translation by multiple miRNAs has already been demonstrated in primary hippocampal neurons (Wibrand et al., 2012). Given the crucial role of Arc in experience-dependent synaptic plasticity and long-term memory (Korb and Finkbeiner, 2011; Shepherd and Bear, 2011) and the deficits in LTP and fear memory observed in Satb2 cKO mice, we examined whether Arc protein is altered in vivo in the CA1 of Satb2 mutants. Indeed, immunoblotting analysis revealed significantly reduced Arc protein level in Satb2 cKO mice compared to littermate controls (Figure 7A). At the same time, Arc mRNA level was not altered (Figure 7B).10.7554/eLife.17361.025Figure 7.Hippocampal Satb2 re-expression rescues Arc levels and long-term fear memory deficits.(A) Representative Western blot (left) and quantification (right) of Arc protein level in the CA1 field of Satb2 cKO mice and littermate controls. β3-tubulin was used as a loading control (Ctrl, n = 9; cKO, n = 11; Student's t test, t18 = 4.52, p = 0.0003). Data are presented as mean ± SEM, n values refer to the number of mice per group, ***p < 0.001 compared to Ctrl. (B) qPCR quantification of Arc mRNA level in the CA1 field of Satb2 cKO mice and littermate controls (Ctrl, n = 3; cKO, n = 4; Student's t test, t5 = 1.14, p = 0.303). Data are presented as mean ± SEM, n values refer to the number of mice used. (C) Satb2/V5-immunoreactivity in the CA1 hippocampal area of Satb2 cKO mice after stereotaxic injection of rAAV-hSyn-Satb2-V5 into the dorsal hippocampus. Nuclei were counterstained with RedDot. Representative images are shown. Scale bar: 150 μm. (D) Representative Western blot (left) and quantification (right) of Arc protein level in the CA1 field of Satb2flox/flox mice injected with rAAV-hSyn-eGFP (Ctrl:AAV-GFP, n = 7), Satb2 cKO mice injected with rAAV-hSyn-eGFP (cKO:AAV-GFP, n = 3) and Satb2 cKO mice injected with rAAV-hSyn-Satb2-V5 (cKO:AAV-Satb2, n = 6). β3-tubulin was used as a loading control. Re-expression of Satb2 in the dorsal hippocampus rescued the reduction in Arc protein in the CA1 area of Satb2 cKO mice, bringing it up to control levels (ANOVA followed by Fischer LSD post hoc test, F2,13 = 5.011, cKO:AAV-GFP vs. Ctrl:AAV-GFP, p = 0.014, cKO:AAV-Satb2 vs. cKO:AAV-GFP, p = 0.011, Ctrl:AAV-GFP vs. cKO:AAV-Satb2, p = 0.793). Data are presented as mean ± SEM, n values refer to the number of mice used, *p < 0.05 compared with Ctrl:AAV-GFP, #p < 0.05, compared with cKO:AAV-GFP. (E) In a contextual fear conditioning test, Ctrl:AAV-GFP (n = 9), cKO:AAV-GFP (n = 9) and cKO:AAV-Satb2 (n = 14) mice showed similar levels of freezing during the fear-acquisition phase (repeated measures ANOVA, F6,87 = 0.12, p = 0.99). Freezing behavior, analyzed 24 hr after the training, was significantly impaired in cKO:AAV-GFP mice, however the fear memory deficit was completely rescued in cKO:AAV-Satb2 mice (ANOVA followed by Fischer LSD post hoc test, F2,29 = 12.8, cKO:AAV-GFP vs. Ctrl:AAV-GFP, p = 0.00005, cKO:AAV-Satb2 vs. cKO:AAV-GFP, p = 0.0004, Ctrl:AAV-GFP vs. cKO:AAV-Satb2, p = 0.22). Data are presented as mean ± SEM, n values refer to the number of mice per group, ***p < 0.001 compared with cKO:AAV-Satb2 and Ctrl:AAV-GFP.DOI: http://dx.doi.org/10.7554/eLife.17361.02510.7554/eLife.17361.026Figure 7—source data 1.List of miRNAs predicted to target mouse Arc 3’UTR by the bioinformatics tools TargetScan, PITA and miRanda.The miRNAs identified as up-regulated in the CA1 region of Satb2 cKO vs. control mice are shown in red.DOI: http://dx.doi.org/10.7554/eLife.17361.026 | other | 27.7 |
(A) Representative Western blot (left) and quantification (right) of Arc protein level in the CA1 field of Satb2 cKO mice and littermate controls. β3-tubulin was used as a loading control (Ctrl, n = 9; cKO, n = 11; Student's t test, t18 = 4.52, p = 0.0003). Data are presented as mean ± SEM, n values refer to the number of mice per group, ***p < 0.001 compared to Ctrl. (B) qPCR quantification of Arc mRNA level in the CA1 field of Satb2 cKO mice and littermate controls (Ctrl, n = 3; cKO, n = 4; Student's t test, t5 = 1.14, p = 0.303). Data are presented as mean ± SEM, n values refer to the number of mice used. (C) Satb2/V5-immunoreactivity in the CA1 hippocampal area of Satb2 cKO mice after stereotaxic injection of rAAV-hSyn-Satb2-V5 into the dorsal hippocampus. Nuclei were counterstained with RedDot. Representative images are shown. Scale bar: 150 μm. (D) Representative Western blot (left) and quantification (right) of Arc protein level in the CA1 field of Satb2flox/flox mice injected with rAAV-hSyn-eGFP (Ctrl:AAV-GFP, n = 7), Satb2 cKO mice injected with rAAV-hSyn-eGFP (cKO:AAV-GFP, n = 3) and Satb2 cKO mice injected with rAAV-hSyn-Satb2-V5 (cKO:AAV-Satb2, n = 6). β3-tubulin was used as a loading control. Re-expression of Satb2 in the dorsal hippocampus rescued the reduction in Arc protein in the CA1 area of Satb2 cKO mice, bringing it up to control levels (ANOVA followed by Fischer LSD post hoc test, F2,13 = 5.011, cKO:AAV-GFP vs. Ctrl:AAV-GFP, p = 0.014, cKO:AAV-Satb2 vs. cKO:AAV-GFP, p = 0.011, Ctrl:AAV-GFP vs. cKO:AAV-Satb2, p = 0.793). Data are presented as mean ± SEM, n values refer to the number of mice used, *p < 0.05 compared with Ctrl:AAV-GFP, #p < 0.05, compared with cKO:AAV-GFP. (E) In a contextual fear conditioning test, Ctrl:AAV-GFP (n = 9), cKO:AAV-GFP (n = 9) and cKO:AAV-Satb2 (n = 14) mice showed similar levels of freezing during the fear-acquisition phase (repeated measures ANOVA, F6,87 = 0.12, p = 0.99). Freezing behavior, analyzed 24 hr after the training, was significantly impaired in cKO:AAV-GFP mice, however the fear memory deficit was completely rescued in cKO:AAV-Satb2 mice (ANOVA followed by Fischer LSD post hoc test, F2,29 = 12.8, cKO:AAV-GFP vs. Ctrl:AAV-GFP, p = 0.00005, cKO:AAV-Satb2 vs. cKO:AAV-GFP, p = 0.0004, Ctrl:AAV-GFP vs. cKO:AAV-Satb2, p = 0.22). Data are presented as mean ± SEM, n values refer to the number of mice per group, ***p < 0.001 compared with cKO:AAV-Satb2 and Ctrl:AAV-GFP. | clinical case | 25.17 |
Our results establish Satb2 as an important determinant of memory consolidation in the adult hippocampus. At molecular level we show in primary hippocampal neurons that calcium influx through L-type VGCC as well as BDNF up-regulate Satb2 in the nucleus, where it binds to promoters of coding and non-coding loci. Altered expression of Satb2-dependent miRNAs on a genome-wide scale is likely to cause changes in the posttranscriptional regulation of synaptic plasticity proteins, exemplified by Arc. Consistent with these molecular mechanisms, the lack of Satb2 in the forebrain causes impairments in late-phase LTP and long-term memory in adult mice. | study | 30.77 |
Given the potential of Satb2 to mediate DNA looping and local association of gene promoter regions (Zhou et al., 2012), our findings suggest that some of the key functions of the hippocampus may depend on changes in the higher-order chromatin architecture. Long-range looping interactions between promoters and distal regulatory elements are considered to recruit transcription factors and chromatin-modifying complexes to gene promoters (Bharadwaj et al., 2014; Sanyal et al., 2012). Our data demonstrating specific binding of Satb2 to active coding or non-coding gene promoters, co-occupied by RNA Pol II, are consistent with this model. Based on our results in primary hippocampal cultures, it can be proposed that this type of higher-order chromatin rearrangement is an activity- and BDNF-dependent process that involves changes in Satb2 expression levels. Of note, the reported schizophrenia risk allele for SATB2 (rs6704641) is intronic and likely affects SATB2 mRNA levels rather than protein function (Schizophrenia Working Group of the Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014). This finding strengthens the hypothesis that quantitative changes in Satb2, which otherwise is expressed in all CA1 pyramidal neurons, impact on Satb2-chromatin interactions. The relevance of chromatin looping in cognition or psychiatric diseases has already been shown for the regulation of individual genes: the NMDA receptor locus GRIN2B46 (Bharadwaj et al., 2014), and two schizophrenia-risk genes, encoding the GABA synthesis enzyme GAD1 and the calcium channel alpha subunit CACNA1C (Nestler et al., 2016; Rajarajan et al., 2016). Noteworthy, the homologue of Satb2, Satb1, a known chromatin structure organizer, affects the formation of dendritic spines and modulates the expression of multiple immediate early genes in the cortex (Balamotis et al., 2012), thus corroborating the importance of Satb family members in neuronal plasticity. | study | 28.95 |
We found that a large fraction of all miRNAs expressed in the adult CA1 hippocampal field is deregulated when Satb2 is genetically ablated. This establishes Satb2 as a novel regulator of the miRNA transcriptome in CA1 pyramidal neurons. Notably, several miRNAs with altered expression in Satb2 mutants, e.g. miR-124, miR-125b, miR-132, miR-212, miR-381, miR-326, miR-19b have already been shown to affect the translation of proteins important in various aspects of synaptic plasticity or memory formation (Aksoy-Aksel et al., 2014; Ryan et al., 2015). At synaptic sites translation of locally synthesized proteins is at least in part repressed by miRNAs. The miRNA machinery interacts with fragile X mental retardation protein (FMRP), which acts as translational repressor. Intriguingly, almost all of the miRs found to be associated with FMRP (miR-100, miR-124, miR-125a, miR-127, miR-128, miR-132, miR-143) (Edbauer et al., 2010) were up-regulated in Satb2 mutants. This finding implies a decreased expression of synaptic proteins after Satb2 ablation and it also provides interesting candidates for future investigations. In our study, we have already demonstrated a decreased level of Arc in the CA1 of Satb2 cKO mice. The identification and validation of additional Satb2-dependent miRNA–mRNA interactions in vivo in the adult brain will be important to elucidate the mechanisms, by which Satb2 regulates memory formation. Recent studies have demonstrated enhanced spatial learning and working memory capacity after inhibition of miR-124 in the hippocampus or restored spatial memory, social interaction and LTP impairments in adult mice carrying a null mutation for EPAC protein (Malmevik et al., 2016; Yang et al., 2012). Our data showing up-regulation of miR-124 and impaired LTP and long-term memory in Satb2 mutants are in full agreement with these reports. Interestingly, amongst the miRNAs deregulated in Satb2-deficient CA1 we also found miRNAs shown to be deregulated in schizophrenia animal models or in schizophrenia patients (Beveridge and Cairns, 2012). | study | 30.52 |
Although the precise mechanism(s) of Satb2-dependent miRNA regulation remain to be determined, two lines of argument support the view that Satb2 has a direct influence on miRNA transcription. First, we find individual miRNAs to be up- as well as down-regulated, arguing against a general effect of Satb2 on miRNA biogenesis or processing by the regulation of Dicer as has been reported for BDNF (Huang et al., 2012). Second, we find that miRNA promoter elements are bound by Satb2 in hippocampal neurons. Our ChIP-seq data in primary cultures demonstrate a potential of Satb2 to occupy both sets of promoters (protein-coding and non-coding). The preferential regulation of miRNA expression over protein-coding genes that we observed in vivo might for example be explained by the presence of Satb2 co-interactors in mature neurons that determine selective binding and control of miR promoters. | study | 28.16 |
In conclusion, we provide evidence that Satb2 is required for synaptic plasticity and long-term memory formation in the adult CNS, likely via the regulation of miRNAs and protein-coding genes controlling synaptic structure and function. Our findings offer a plausible mechanism explaining the intellectual disability and severe learning difficulties observed in SAS patients. Furthermore, the Satb2 function in the adult brain unsuspectedly interconnects several individual components that have been discussed in the context of various psychiatric syndromes: BDNF signaling, epigenetic chromatin modifications, miRNA dysregulation and cognitive dysfunction. Therefore, in the future, it will be intriguing to establish Satb2flox/flox::Camk2a-Cre conditional mutant as an animal model of neuropsychiatric diseases. | study | 32.5 |
Mice carrying an allele of Satb2 in which exon 4 is flanked by loxP sites (Satb2flox/flox) were generated by microinjection of ES cells carrying a Satb2tm1a(KOMP)Wtsi Knockout First (Promoter driven) allele (clone Satb2_G07, JM8.N4 subline; KOMP repository) into blastocysts from albino C57BL/6J donor mice. The germline-positive mice were further crossed with FLPo deleter mice (RRID:MMRRC-032247-UCD) (Kranz et al., 2010) to excise the FRT cassette and to establish a conditional allele. To generate Satb2 conditional mutants Satb2flox/flox mice were crossed with Camk2a-Cre mice (Minichiello et al., 1999) on a C57BL/6 background. In all experiments, mice that carry the floxed exon 4 but do not express the Cre transgene (Satb2flox/flox) were used as littermate controls. Unless otherwise stated, adult male mice at the age of 3–4 months were used for behavioral and molecular analyses. All experimental procedures were approved by the Austrian Animal Experimentation Ethics Board (Bundesministerium für Wissenschaft und Verkehr, Kommission für Tierversuchsangelegenheiten). | other | 33.44 |
Primary antibodies. Satb2 ab92446 (AB_10563678), V5 tag antibody ChIP grade ab15828 (AB_443253), Satb2 ab34735 (AB_2301417), Ctip2 ab18465 (AB_2064130), Tbr1 ab31940 (AB_2200219) were purchased from Abcam (Cambridge, MA); Erk2 sc-154-G (AB_631459), Arc sc-17839 (AB_626696), Cux1 sc-13024 (AB_2261231) were obtained from Santa Cruz (Dallas, TX); Arc 156 003 (AB_887694) was ordered from Synaptic Systems (Germany); V5 epitope tag antibody R960-25 (AB_2556564) was obtained from Thermo Fisher Scientific (Waltham, MA); Satb2 AMAb90682 was purchased from Atlas Antibodies (Sweden); Wfs1 11558–1-AP (AB_2216046) was ordered from Proteintech (Rosemont, IL); and beta-III Tubulin antibody NB100-1612 (AB_10000548) was obtained from Novus Biologicals (UK). | other | 29.3 |
Secondary antibodies. Goat anti-mouse Alexa-488 A11001 (AB_2534069), donkey anti-mouse Alexa-488 A21202 (AB_2535788), goat anti-rabbit Alexa-555 A21428 (AB_10561552), donkey anti-rabbit Alexa-488 A21206, donkey anti-rat Alexa 488 A21208 (AB_2535792) were all purchased from Thermo Fisher Scientific; and goat anti-mouse CF633 20121 (AB_10854245) was obtained from Biotium (Fremont, CA). | other | 30.08 |
Hippocampi were dissected from newborn C57BL/6J mice at postnatal day P0 to P1. Hippocampal tissue was trypsinized and dissociated by trituration as described previously (Kaech and Banker, 2006). Neurons were plated on 35 mm tissue culture dishes or on 18 mm coverslips, previously coated with poly-L-ornithine (Sigma, St. Louis, MO) and laminin (Thermo Fischer Scientific), at a density of 8×104 cells/cm2. Cells were cultured in Neurobasal medium (Thermo Fisher Scientific) containing B-27 supplement (Thermo Fisher Scientific), 100 mg/ml penicillin G and 60 mg/ml streptomycin sulfate in a humidified atmosphere of 5% CO2 at 37°C for 8–10 days. Medium was replaced 1.5 hr after plating; thereafter one third of the culture medium was replaced with fresh medium at day in vitro (DIV) 3 or 4. Glial cell proliferation was inhibited by adding 5 µM cytosine arabinoside (Sigma) to the culture medium at DIV1. Cells were collected for immunoblotting analysis 24 hr after treatment. The following growth factors were used: BDNF (50 ng/ml, Peprotech, Rocky Hill, NJ), NT4/5 (50 ng/ml, a kind gift from Amgen-Regeneron). The following pharmacological substances were applied as described in the Results section: 50 μM bicuculline (Tocris Bioscience, UK), 500 μM 4-aminopyridine (Tocris Bioscience). Pharmacological inhibitors were added 1 hr prior to treatments with BDNF or Bic/4AP in the following concentrations: 10 μM nimodipine (Tocris Bioscience), 10 μM MK-801 (Tocris Bioscience), 10 μM SB202190 (Cell Signaling, Danvers, MA), 10 μM UO126 (Cell Signaling), 0.2 μM K252a (Santa Cruz), 333 ng/ml actinomycin D (Sigma) and 20 µM SB747651A (Axon Medchem, The Netherlands). Control cultures were treated with equal volumes of vehicle. | other | 29 |
Subsets and Splits