Proliferating Reactive Astrocytes Are Regulated by Notch-1 in the Peri-Infarct Area After Stroke
Background and Purpose—The formation of reactive astrocytes is common after central nervous system injuries such as stroke. However, the signaling pathway(s) that control astrocyte formation and functions are poorly defined. We assess the effects of Notch 1 signaling in peri-infarct-reactive astrocytes after stroke.
Methods—We examined reactive astrocyte formation in the peri-infarct area 3 days after distal middle cerebral artery occlusion with or without γ-secretase inhibitor treatment. To directly study the effects of inhibiting a γ-secretase cleavage target in reactive astrocytes, we generated glial fibrillary acidic protein-CreERTM:Notch 1 conditional knockout mice.
Results—Gamma-secretase inhibitor treatment after stroke decreased the number of proliferative glial fibrillary acidic protein-positive reactive astrocytes and RC2-positive reactive astrocytes directly adjacent to the infarct core. The decrease in reactive astrocytes correlated with an increased number of CD45-positive cells that invaded into the peri-infarct area. To study the influence of reactive astrocytes on immune cell invasion, ex vivo immune cell invasion assays were performed. We found that a γ-secretase-mediated pathway in astrocytes affected Jurkat cell invasion. After tamoxifen treatment, glial fibrillary acidic protein-CreERTM:Notch 1 conditional knockout mice had a significantly decreased number of proliferating reactive astrocytes and RC2-positive reactive astrocytes. Tamoxifen treatment also led to an increased number of CD45-positive cells that invaded the peri-infarct area.
Conclusions—Our results demonstrate that proliferating and RC2-positive reactive astrocytes are regulated by Notch 1 signal transduction and control immune cell invasion after stroke.
- animal models
- basic science
- brain infarction
- brain ischemia
- cerebral infarct
- focal ischemia
- gene regulation
- glial cells
After brain injury, reactive astrocytes participate in the formation of a “glial scar.” Because the glial scar has been shown to inhibit axonal regeneration, reactive astrocytes have historically been considered as a barrier to neuronal repair.1,2 However, this negative view of reactive astrocytes is changing because they have been shown to play several important roles in preserving neural tissue during the early phase of brain injuries.2–4 Knockout mice lacking two intermediate filament proteins commonly expressed by astrocytes, glial fibrillary acidic protein (GFAP), and vimentin exhibited attenuated reactive astrocyte formation, reduced glutamate transport, and increased infarct volumes 7 days after stroke.5 Reactive astrocytes were also shown to limit cellular degeneration by maintaining and repairing the blood–brain barrier and decreasing immune cell infiltration after stab injury.2,6
The emergence of reactive astrocytes in the peri-infarct area is one of the most obvious events in the brain after stroke. However, the signaling pathway(s) that control astrocyte formation and their functions in the peri-infarct area are poorly defined. Gamma-secretase-targeted proteins such as Notch 1 and APP are expressed by reactive astrocytes after brain injury.7,8 Here we determine the effects of Notch 1 signaling on a unique subpopulation of proliferative reactive astrocytes that localize immediately adjacent to the infarct core and that regulate the peri-infarct area after stroke.
All animal work was approved by the University of Vermont College of Medicine's Office of Animal Care in accordance with American Association for Accreditation of Laboratory Animal Care and National Institutes of Health guidelines. Focal cerebral ischemia was produced by permanently occluding the middle cerebral artery.9–11
Full methods are available online (http://stroke.ahajournals.org).
RC2-Positive Reactive Astrocytes Express Gamma-Secretase Cleavage Products
To study reactive astrocytes in the brain after stroke, we performed distal middle cerebral artery occlusion in wild type C57bl6J mice (Taconic Farms, Hudson, NY). Reactive astrocytes expressed GFAP and/or RC2 in the cortical peri-infarct area after stroke (Figure 1A). We previously identified an RC2-positive subpopulation of reactive astrocytes that formed on the ipsilateral but not the contralateral side of the brain after stroke.11 At 3 days after stroke, the number of RC2-positive cells was significantly higher in the inner cortical layer adjacent to the infarct (200 μm radially from the infarct area) compared with the outer cortical layer (200 μm radially from the inner layer; n=3, P<0.05; Figure 1B). Similarly, the number of GFAP/RC2-positive cells was significantly greater in the inner cortical layer compared with the outer layer, demonstrating that RC2-positive reactive astrocytes appear directly adjacent to the infarct core after stroke (n=3, P<0.05; Figure 1B). Ki67/GFAP-positive proliferating reactive astrocytes were also observed adjacent to the infarct core after stroke (Figure 1C). The 3 day time point was chosen because the RC2 antigen and GFAP were highly expressed at this time relative to 1 day after stroke.
Gamma-secretase activity is known to increase in the brain early after stroke.7,12 We examined the presence of γ-secretase cleavage products: APP Intracellular Domain (AICD), and Notch 1 Intracellular Domain (NICD1) in the peri-infarct area 3 days after stroke. By immunohistochemistry, GFAP-positive reactive astrocytes and RC2-positive reactive astrocytes both expressed AICD and NICD1 in the peri-infarct area (Figure 2A). We also observed expression and nuclear localization of AICD and NICD1 in neuronal nuclei (NeuN)-positive neurons (data not shown). These data suggested that γ-secretase cleavage products may regulate reactive astrocytes and RC2-positive reactive astrocytes in the peri-infarct area after stroke.
Formation of Reactive Astrocytes Is Disrupted by Gamma-Secretase Inhibitor Treatment
To determine whether γ-secretase-mediated events controlled the formation of reactive astrocytes, we inhibited Type I intracellular membrane protein cleavage with the γ-secretase inhibitor, dibenzazepine (GSI; Figure 2B). In accordance with altered γ-secretase activity early after stroke, we observed increased AICD expression levels in the peri-infarct area 1 day after stroke compared with AICD levels in sham-operated brains (n=6, P<0.01; Figure 2C). After GSI treatment, we observed decreased levels of AICD expression in the peri-infarct area compared with dimethyl sulfoxide (DMSO) treatment (n=6, P<0.05; Figure 2C). Overall NICD1 expression levels in the peri-infarct area did not increase after stroke compared with levels in sham animals. We did, however, observe a trend of decreased NICD1 expression levels in the peri-infarct area after GSI treatment compared with DMSO treatment (sham, 106%± 18.9%; DMSO, 100%±15.4%; GSI, 79%±7.4%; mean± SEM, n=4 to 6 mice for each group). These data indicated that GSI treatment affected cleavage of Type I intramembrane proteins in the peri-infarct area after stroke. As expected, GFAP protein expression levels were increased in the peri-infarct area 3 days after stroke compared with sham-operated brains. In contrast, GFAP protein expression levels were decreased 3 days after stroke and GSI treatment compared with DMSO treatment (n=10 to 12, P<0.01; Figure 2D).
We quantified the number of GFAP-positive reactive astrocytes after stroke and GSI treatment. The number of GFAP/Ki67-positive proliferating reactive astrocytes was significantly decreased by GSI treatment (n=4, P<0.05; Figure 3A). We quantified also the number of RC2-positive reactive astrocytes in the peri-infarct area after stroke and DMSO or GSI treatment. We observed a decreased number of RC2-positive reactive astrocytes in the cortical peri-infarct area after GSI treatment compared with vehicle treatment (n=3, P<0.05; Figure 3B). These data indicated that γ-secretase cleavage products play an important role in regulating the proliferating reactive astrocytes and the subpopulation of RC2-positive reactive astrocytes in the peri-infarct area after stroke. Notably, we did not observe a significant difference in the total numbers of GFAP-positive cells after stroke with GSI treatment compared with vehicle controls (DMSO, 5255.1±437.9 cells/mm2; GSI, 5199.8± 231.8 cells/mm2; mean±SEM, n=4 mice for each group, P=0.91).
To examine whether dimethyl sulfoxide or GSI treatment affected reactive astrocyte apoptosis/necrosis after stroke, we performed terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling assays for mice treated with phosphate-buffered saline, DMSO, or GSI. We observed many terminal deoxynucleotidyltransferase-mediated dUTP nick end labeled (TUNEL)-positive cells in the infarct areas of mice treated with phosphate-buffered saline, DMSO, and GSI (data not shown). However, we did not observe TUNEL-positive/GFAP-positive reactive astrocytes in the peri-infarct area 3 days after stroke (data not shown).
Reduced Number of Proliferating and RC2-Positive Reactive Astrocytes and Immune Cell Invasion Into the Peri-Infarct Area
One possible role of reactive astrocytes is to prevent immune cell invasion as reported in spinal cord injury and brain stab injury.13,14 To determine whether this was the case in our model, we quantified CD45-positive monocytes at the border of the infarct region with or without GSI treatment. Because monocytes and microglia express similar proteins such as CD45 and CD11b, it is technically challenging to distinguish monocytes and microglia in the peri-infarct area. Therefore, we used CD45 as a marker and distinguished monocytes by their round morphology in the peri-infarct area. We observed a significant increase in the number of CD45-positive cells in the peri-infarct area after GSI treatment compared with the number in vehicle-treated controls (n=4 to 5, P<0.05; Figure 3C).
To study the effects of astrocytes on immune cell invasion, we isolated astrocytes from neonatal mouse brains. When cultured in serum-containing medium, the astrocytes were activated and expressed GFAP, Nestin, and RC2, similar to protein expression patterns of RC2-positive reactive astrocytes after stroke (Figure 4A). Jurkat cells (T cell line) are commonly used to study immune cell invasion.15 Jurkat cell invasion was determined with either 5% fetal bovine serum or 5% fetal bovine serum with astrocytes in the bottom wells of invasion chambers. In agreement with our observations in vivo, astrocytes cultured in 5% fetal bovine serum significantly decreased Jurkat cell invasion compared with positive migration control (5% fetal bovine serum medium alone; n=3, P<0.05; Figure 4B). Assays of lactate dehydrogenase activity showed that astrocyte conditioned medium did not induce cell death in Jurkat cells (n=4, P=0.96; Figure 4B). To understand the effects of γ-secretase-mediated signaling in cultured astrocytes, invasion studies were performed using GSI-treated astrocytes and DMSO-treated astrocytes. GSI treatment significantly reduced the proliferation of astrocytes compared with DMSO treatment (n=4, Cyquant assay, Day 4, P<0.01; Day 6, P<0.05; 5-bromodeoxyuridine incorporation assay, n=4, P<0.05; Figure 4C). Assays of lactate dehydrogenase acitivity showed that GSI treatment did not induce cell death in astrocytes (n=3, P=0.25; Figure 4C). After GSI treatment of astrocytes, the number of invading Jurkat cells increased compared with DMSO treatment (n=4, P<0.05; Figure 4C).
To determine whether NICD1 levels affect astrocyte proliferation, we transfected cultured astrocytes with NICD1-F2A-green fluorescent protein (GFP) or GFP control plasmid. Two days after transfection, we examined the number of proliferating astrocytes by immunocytochemistry for Ki67. We observed significantly more proliferative astrocytes after NICD1-F2A-GFP transfection compared with GFP transfection alone (n=6, P<0.05; Figure 4D).
Notch 1 Inhibition in GFAP-CreERTM:Notch 1 cKO Mice Reduces Reactive Astrocyte Formation
To determine whether Notch 1 regulates reactive astrocyte formation in vivo after stroke, we generated GFAP-CreER™:tdRFP mice (GR mice) and GFAP-CreERTM:Notch 1 conditional knockout (cKO) mice (GN cKO mice). In GR mice, tdRFP should be expressed exclusively in GFAP-positive reactive astrocytes after tamoxifen (TM) treatment. In GN cKO mice, Notch 1 should be knocked out exclusively in GFAP-positive reactive astrocytes after TM treatment (Figure 5). First, to confirm the expression patterns of tdRFP in reactive astrocytes after stroke, we administered TM for 3 consecutive days and performed distal middle cerebral artery occlusion surgery 7 days after the last TM administration. Three days after distal middle cerebral artery occlusion surgery, reactive astrocytes were analyzed by immunohistochemistry. In GR mice, 70% of GFAP-positive reactive astrocytes expressed tdRFP (Figure 6A). To confirm that NICD1 was no longer expressed in GFAP-positive cells after TM treatment of GN cKO mice, immunohistochemistry against NICD1 was performed. As expected, after TM treatment, but not after corn oil treatment, NICD1 was not expressed in GFAP-positive cells (Figure 6A). Quantifying GFAP-positive cells, we observed a significantly reduced number of proliferative reactive astrocytes in GN cKO mice after TM treatment (n=4, P<0.01; Figure 6B). Notably, the number of RC2-positive reactive astrocytes was also significantly decreased after TM treatment (n=4, P<0.01; Figure 6C).
These data demonstrated that Notch 1 plays an important role in reactive astrocyte formation in the peri-infarct area after stroke. To determine whether the decreased number of proliferating reactive astrocytes would affect immune cell invasion, we quantified CD45-positive cell invasion into the peri-infarct area. In agreement with our pharmacological GSI treatment data and ex vivo invasion studies, we also observed a significantly increased number of CD45-positive cells in the peri-infarct area in TM-treated mice (n=4 to 5, P<0.05; Figure 6C). These data showed that proliferating reactive astrocytes, which include RC2-positive reactive astrocytes, are important for suppressing immune cell invasion after stroke.
Elucidation of the signaling mechanism(s) that regulate reactive astrocyte formation is important to understand and to treat central nervous system injury. Gamma-secretase-mediated Notch signaling occurs through a conserved pathway that is important for stem cell proliferation and differentiation.16 We found that reactive astrocytes expressed the γ-secretase cleavage products, NICD1 and AICD, in the peri-infarct area after stroke. To examine the effects of γ-secretase activity and Notch 1 signaling on reactive astrocyte formation, we treated mice with GSI and generated GN cKO mice to exclusively knockout Notch 1 in GFAP-positive cells. The number of proliferating reactive astrocytes and RC2-positive reactive astrocytes were both significantly decreased after GSI treatment and also in GN cKO mice, demonstrating that Notch 1 plays a critical role in reactive astrocyte formation after stroke.
Reactive astrocytes were previously shown to prevent immune cell invasion and to reduce inflammation after brain stab injury and spinal cord injury.2,13 We found that the number of invading CD45-positive cells was increased in the peri-infarct area after GSI treatment and also in GN cKO mice. These data demonstrated that reactive astrocyte proliferation requires Notch 1 and that proliferating reactive astrocytes may have a specialized role in protecting the brain after stroke by decreasing immune cell invasion.
We did not observe an infarct size difference at 3 days after distal middle cerebral artery occlusion in TM-treated GN cKO mice compared with controls, suggesting that Notch 1 knockout in reactive astrocytes may not alter neural protection early after stroke (I.S.S. and J.L.S., unpublished data). Importantly, the effects of Notch signaling in stroke may differ based on the type of stroke, duration after onset of occlusion, cell type, and signaling through particular Notch ligands/receptors. Arumugam et al reported that GSI treatment decreased infarct size in a reperfusion model of stroke (intraluminal) by reducing neuronal apoptosis and decreasing the immune response.7 Notch 3 knockout mice were reported to have larger stroke volumes than controls after transient middle cerebral artery occlusion.17 Infusion of Delta-like 4, a Notch ligand, did not alter infarct volume after stroke.18 Additional studies will be necessary to fully understand the roles of Notch signaling in stroke and where and when it is beneficial or detrimental to outcome.
We have shown that Notch 1 is a key factor required for reactive astrocyte proliferation in the peri-infarct area after stroke. Our data also indicate that the proliferative pool of reactive astrocytes surrounding the infarct core suppress immune cell invasion into peri-infarct tissues. Proliferating reactive astrocytes and the RC2-positive subpopulation of reactive astrocytes adjacent to the infarct core may provide important targets for treatment of stroke.
Sources of Funding
I.S.S. is a recipient of the American Heart Association Postdoctoral Fellowship (10POST3730026). This work was supported by P20 RR016435 NIH/NCRR (R. Parsons, COBRE Principal Investigator; J.L.S., Principal Investigator, Project 3).
We thank Dr Suzanne Baker for GFAP-CreERTM mice (St Jude's Children Hospital, Memphis, TN) and Dr Hans Joerg Fehling for ROSA-tdRFP mice (University Clinics, Ulm, Germany). We thank Dr Diane Jaworski, University of Vermont, for assistance with primary astrocyte culture.
The online-only Data Supplement is available at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.623280/-/DC1.
- Received April 13, 2011.
- Accepted May 3, 2011.
- © 2011 American Heart Association, Inc.
- Nagai N,
- De Mol M,
- Lijnen HR,
- Carmeliet P,
- Collen D
- Herrmann JE,
- Imura T,
- Song B,
- Qi J,
- Ao Y,
- Nguyen TK,
- et al
- Pandey MK,
- Sung B,
- Kunnumakkara AB,
- Sethi G,
- Chaturvedi MM,
- Aggarwal BB
- Ables JL,
- Decarolis NA,
- Johnson MA,
- Rivera PD,
- Gao Z,
- Cooper DC,
- et al
- Arboleda-Velasquez JF,
- Zhou Z,
- Shin HK,
- Louvi A,
- Kim HH,
- Savitz SI,
- et al