TGFβ signaling in the brain increases with aging and signals to astrocytes and innate immune cells in the weeks after stroke
© Doyle et al. 2010
Received: 29 July 2010
Accepted: 11 October 2010
Published: 11 October 2010
Skip to main content
© Doyle et al. 2010
Received: 29 July 2010
Accepted: 11 October 2010
Published: 11 October 2010
TGFβ is both neuroprotective and a key immune system modulator and is likely to be an important target for future stroke therapy. The precise function of increased TGF-β1 after stroke is unknown and its pleiotropic nature means that it may convey a neuroprotective signal, orchestrate glial scarring or function as an important immune system regulator. We therefore investigated the time course and cell-specificity of TGFβ signaling after stroke, and whether its signaling pattern is altered by gender and aging.
We performed distal middle cerebral artery occlusion strokes on 5 and 18 month old TGFβ reporter mice to get a readout of TGFβ responses after stroke in real time. To determine which cell type is the source of increased TGFβ production after stroke, brain sections were stained with an anti-TGFβ antibody, colocalized with markers for reactive astrocytes, neurons, and activated microglia. To determine which cells are responding to TGFβ after stroke, brain sections were double-labelled with anti-pSmad2, a marker of TGFβ signaling, and markers of neurons, oligodendrocytes, endothelial cells, astrocytes and microglia.
TGFβ signaling increased 2 fold after stroke, beginning on day 1 and peaking on day 7. This pattern of increase was preserved in old animals and absolute TGFβ signaling in the brain increased with age. Activated microglia and macrophages were the predominant source of increased TGFβ after stroke and astrocytes and activated microglia and macrophages demonstrated dramatic upregulation of TGFβ signaling after stroke. TGFβ signaling in neurons and oligodendrocytes did not undergo marked changes.
We found that TGFβ signaling increases with age and that astrocytes and activated microglia and macrophages are the main cell types that undergo increased TGFβ signaling in response to post-stroke increases in TGFβ. Therefore increased TGFβ after stroke likely regulates glial scar formation and the immune response to stroke.
Transforming Growth Factor β1 (TGF-β1) is universally induced by acute and chronic brain injury, including stroke, trauma, seizure, multiple sclerosis, and Alzheimer's disease . TGFβ is highly conserved and in mammals exists as three isoforms that bind to the same receptors; TGF-β1, TGF-β2 and TGF-β3. TGF-β1 is the isoform typically induced by injury . In the brain, TGFβ receptors are present on all major cell types . However, both TGFβ activation and signaling are extensively regulated and due to this its effects are typically pleiotrophic and context-dependent. Therefore, after brain injury the biological effects of TGFβ signaling are likely influenced by both injury type and timing.
After brain injury, TGFβ signaling can be neuroprotective, but also promote glial scarring and fibrosis [1, 4–6]. While neuroprotection is undoubtedly desirable, glial scarring and fibrosis are less likely to be advantageous. In addition, TGFβ's have potent effects on the immune system. TGF-β1 can be stimulatory or inhibitory depending on the cell type, cytokine milieu and differentiation state of the responding cell and can have both pro- and anti-inflammatory effects . It is not known which subset of these roles TGF-β1 plays when it is induced by brain injury and how this varies in different types of brain injury.
In stroke, TGF-β1 mRNA is elevated for at least a week afterwards  and clearly exerts a neuroprotective role. When overexpressed, TGF-β1 limits stroke size [4, 9, 10]. And when TGFβ signaling is blocked, ischemic damage is exacerbated . Because of this, it may be an effective therapeutic agent for stroke. Developing new therapeutic agents to treat those affected by stroke is critical. The elderly segment of the American population is rapidly increasing and 70% of the deaths of individuals over 65 are attributable to cardiovascular disease . But in order to develop TGFβ as a therapeutic agent, it is necessary to understand if it plays any additional roles in the brain after stroke.
We therefore investigated the timecourse and cell-specificity of TGFβ signaling after stroke, and whether its signaling pattern is altered by gender and aging in a mouse model of stroke. We report here that TGFβ signaling increases 2 fold after the dMCAO model of stroke, beginning on day 1 and peaking on day 7. This pattern of increase is similar in each gender and preserved in old animals, although absolute TGFβ signaling was much higher in the brains of the aged animals. CD68+ activated microglia and macrophages were the predominant source of increased TGF-β1 after stroke and astrocytes and CD68+ cells were the main cell types that responded to the post-stroke increase in TGF-β1. Our results demonstrate that TGFβ signaling peaks at 7 days after dMCAO stroke, increases with aging, and that the cells with increased responses during this time period, astrocytes and microglia, are those involved in glial scarring and innate immune responses. This may indicate that TGFβ signaling plays a role in regulating both the immune response and the glial scar after stroke.
SBE-LucRT mice were a gift from Dr Tony Wyss-Coray . Nine females and 5 males (5 months old) were used to gauge variability in bioluminescence between individual mice, and to determine if bioluminescence is affected by the estrus cycle. Eighteen young females (5 months old) and eighteen old females (18 months old) were used to measure TGFβ signaling after stroke in real time by bioluminescence. 3 mice from each cohort were sacrificed at days 1, 3, 7, 14 and 21 after stroke for localization of TGF-β1 and TGFβ signaling by immunoflorescence. Three mice of each age group also underwent sham surgery and were imaged for 21 days prior to sacrifice for localization of TGFβ signaling in uninjured brains. This experimental paradigm was repeated in young and old male mice.
Distal middle cerebral artery occlusion (dMCAO) was induced as described previously . Briefly, mice were anesthetized by isoflurane inhalation (2% isoflurane in 100% oxygen) and the skin over the right temple was shaved. Skin was swabbed with chlorehexidine and an incision was made to expose the temporalis muscle. A pocket was created in the temporalis muscle to expose the skull underneath and the right middle cerebral artery (MCA) was identified. A microdrill was used to penetrate the skull and expose the underlying MCA. The meninges were cut and the vessel was ligated using a small vessel cauterizer. The temporalis muscle was replaced and the wound closed using surgical glue. Throughout surgery body temperature was maintained at 37°C using a feedback controlled heating blanket.
Bioluminescence was detected as described previously using the In Vivo Imaging System (IVIS; Xenogen, Alameda, CA) . Mice were injected intraperitoneally with 150 mg/kg D-luciferin (Xenogen) 10 min before imaging and anesthetized with isoflurane during imaging. Photons emitted from living mice were acquired as photons per second per cm2 per steradian (sr) by using LIVINGIMAGE software (Xenogen) and integrated over 5 min. For photon quantification, a region of interest was manually selected and kept constant within all experiments. Baseline imaging was performed 1 week before stroke was induced and bioluminescence was expressed as fold induction over baseline levels.
For determination of stage of estrus, sterile cotton swabs moistened with saline were inserted into the vagina and gently rotated. Swabs were then rolled onto glass slides and allowed to dry. Polymorphonuclear lymphocytes (PMNs) and epithelial cells were visualized with a Hema 3 Stat Pack (Fisher Cat#23-123-869) used according to the manufactuer's directions. Stage of estrus cycle was determined based on the ratio of PMN's to nucleated and cornified epithlieal cells. Diestrus was defined as the stage when cells consisted mainly of polymorphonuclear lymphocytes (PMNs). Proestrus was defined as the stage when nucleated and cornified epithelial cells were observed, along with a few PMNs. Estrus was classified as the stage when cornified epithelial cells were predominant, and metestrus was defined as the stage when cornified epithelial cells and PMNs were detected.
To visualize the region of infarction, representative animals from each age group (n ≥ 5) were sacrificed at 24 hours post stroke and their brains processed for TTC staining. Brains were sectioned into 1 mm thick slices and each slice was immersed in 1.5% TTC in PBS at 37°C for 15 minutes and then fixed in 10% formalin. The area of the ipsilateral hemisphere and the area of the infarct on each section was measured in a blinded fashion using NIH ImageJ 1.43u. The measurements were multiplied by the distance between sections (1 mm) and then summed over the entire brain to yield volume measurements.
Immunoflorescence was performed on PFA fixed free-floating coronal brain sections (40 μm) using standard techniques. The following primary antibodies were used: anti-TGF-β1 (1:1000; Torrey Pines Biolabs, East Orange, NJ), anti-pSmad2 (1:1000; Cat# AB3849, Millipore, Billerica, MA); Milli-Mark Pan Neuronal marker (1:500; Millipore, Billerica, MA); anti-GFAP (1:1000; Dako, Carpinteria, CA); anti-CD68 (1:1000; ABD Serotec, Raleigh, NC); anti CAII (1:2000; ABD Serotec, Raleigh, NC) and anti-β-dystroglycan (1:1000; Novacastra/Leica Microsystems, Bannockburn, IL). For quantification of Milli-Mark Pan Neuronal marker, CAII, CD68 and GFAP colocalization with pSmad2, two sections from each mouse spaced 300 μm apart were immunostained. From each section four random fields of view in the penumbra (40× objective) were analyzed. The total number of cells positive for each cell marker was averaged and the percentage of each population that was also positive for pSmad2 was calculated.
Unstroked and stroked brains were lysed in glo lysis buffer (Promega). Protein concentrations were equalized and lysates mixed with 4× NuPage LDS loading buffer (Invitrogen). Samples were loaded onto 4-12% NuPage Bis-Tris gels (Invitrogen) and subsequently transferred onto PVDF membranes. Blots were incubated with rabbit polyclonal antibodies (1:1000) against pSmad2 (cat# AB3849 from Millipore and cat# 3101 from Cell Signaling) and actin (cat# A5060 Sigma-Aldritch) and horseradish peroxidase-conjugated goat anti-rabbit IgG as secondary antibody. Protein signals were detected using an ECL kit (Amersham Pharmacia Biotech) and band intensities were quantified on the original digital image using UN-Scan-It Gel 6.1.
Data was acquired in a blinded and unbiased fashion. Data are expressed as mean +/- standard error of the mean (SEM). Statistical analyses were performed with Prism 5 software (GraphPad, San Diego, CA). Means between two groups were compared with two-tailed, unpaired Student's t tests; comparisons of means from multiple groups with one control were analyzed with one-way ANOVA and Dunnet's post hoc test.
To confirm the increase in TGFβ signaling we detected by luciferase activity, we performed a western blot on brain extracts from the ipsilateral hemisphere of mice sacrificed before, and 7 days after stroke, for phosphorylated Smad2 (pSmad2), a downstream mediator of TGFβ signaling. The pSmad2 antibody we used for immunoflorescence (Millipore AB3849) detected a protein with the correct molecular weight for pSmad2 (58 kDa) and bound the same band as an alternative pSmad2 antibody (Cell Signaling 3101; Fig 2G). This verifies that the pSmad2 antibody we used for subsequent immunoflorescence is specific for pSmad2. In accordance with our luciferase findings pSmad2 levels were increased 7 days post stroke (Fig 2G).
In summary, TGFβ signaling increases 2 fold after dMCAO stroke, beginning on day 1 and peaking on day 7, this pattern of increase is preserved in old animals, and absolute TGFβ signaling in the non-injured brain increases with age. We did not observe a gender difference in these findings or in infarct volume.
We show here for the first time that TGFβ signaling in the brain increases in the first week after stroke in both young and old mice. We incorporated 18 month old mice into this study because nearly three quarters of all strokes occur in people over the age of 65 and there is a large knowledge gap regarding how mechanisms of recovery change with age. Similarly, we performed this study on mice of both genders in order to determine if there is a sex difference in TGFβ signaling after stroke. In 5 month old mice, TGFβ signaling increased 2 fold over baseline in the first week after stroke, before beginning to return to pre-stroke levels. This profile of increased TGFβ signaling was similar in 18 month old mice, although the absolute level of TGFβ signaling was significantly higher in the older animals. This could be in response to greater damage in the older animals as we found that lesion size was over double the volume compared to the younger mice. Baseline TGFβ signaling was also higher in the aged animals and so increased TGFβ signaling after stroke may also be a reflection of their higher baseline level. Our finding that infarct volume is increased in older animals conflicts with other studies that report that infarct volume is not increased in older subjects [18, 20]. However, these studies used the suture model of stroke, which causes a larger lesion than the dMCAO model of stroke used here. Studies that use a similar cortical model of stroke to the dMCAO model find that lesion size is increased in aged rats relative to young . Therefore the impact of age on lesion size appears to be stroke model dependent.
Every cell type in the brain has been shown to be capable of making TGF-β1 and increases in TGF-β1 mRNA have been demonstrated after stroke [1, 21]. To discover which cell type is predominantly responsible for post stroke production of TGF-β1 we co-localized an anti-TGF-β1 antibody with markers of different brain cell types. We found that TGF-β1 predominantly co-localized with CD68, a marker of activated microglia and macrophages, indicating that they are likely to be the cells that produce it after stroke.
While there is ample evidence that TGFβ production is increased after stroke, as it is after many kinds of brain injury, it was not known whether increased TGFβ production translates to an increased response, because all TGFβ isoforms are excreted as an inactive form that lies inert in the extracellular matrix . TGFβ receptor insertion into the membrane is highly regulated and the intracellular Smad pathways that transmit TGFβ signals interact with a complex array of other kinase substrates . Therefore, although many of the proteins that activate TGFβs are upregulated after brain injury, such as reactive oxygen species, metalloproteases, plasmin, and thrombospondin , it was not known if increases in TGF-β1 mRNA would correlate directly with increased TGFβ signaling after stroke. And in fact we found that TGFβ signaling was regulated differently in different cell types after stroke.
TGF-β1 is strongly neuroprotective, at least in part due to direct effects in neurons [11, 24–28]. It is also involved in synapse formation, the balance of excitatory and inhibitory transmission in the hippocampus and plasticity in multiple circuits . We found that TGFβ signaling in neurons is widespread and consistent before stroke and that there are no obvious quantitative or qualitative differences after stroke. This suggests that while TGFβ signaling plays an important constitutive function in neurons, the physiological role of increased TGF-β1 after stroke may not be to signal to neurons.
Surprisingly, we also found that the majority of oligodendrocytes in the brain are responding to TGFβ in the absence of injury and after stroke. This suggests that TGFβ signaling plays a physiological role in oligodendrocytes that was heretofore unappreciated. As with neurons, co-localization of pSmad2 with oligodendrocytes did not increase after stroke suggesting that the role of increased TGF-β1 after stroke is not to communicate to this cell type. Whether constitutive TGFβ signaling in oligodendrocytes is important for their function or survival remains to be elucidated.
We rarely found co-localization of pSmad2 with endothelial cells. Signal transduction by TGFβ family members is mediated via specific heteromeric complexes of type I and type II serine/threonine kinase receptors. In most cells TGFβ signals via the type I receptor ALK5 but in endothelial cells it can also signal via the type I receptor ALK1 [7, 30]. ALK5 induces phosphorylation of Smads 2 and 3 and ALK1 mediates phosphorylation of Smads 1, 5 and 8. Endothelial cells may not co-localize with pSmad2 due to the activation of the ALK1 pathway. Thus endothelial cells may still respond to TGFβ in the brain at baseline and after stroke but by activation of the Smads downstream of ALK1.
Our data demonstrates that CD68+ activated microglia and macrophages respond in an autocrine manner to the TGF-β1 they produce. We found that co-localization of CD68 with pSmad2 increased in parallel with TGFβ signaling reporter gene expression after stroke. TGFβ can exert either anti-inflammatory or pro-inflammatory effects in a context-dependent fashion. However, in the majority of contexts TGF-β1 appears to play an anti-inflammatory role and moderate the production of neurotoxic pro-inflammatory cytokines. It drives the differentiation of regulatory T cells and M2c macrophages that resolve immune responses, and resolves the inflammatory process in myocardial ischemia [7, 31, 32]. A function of TGF-β1 after stroke may be to drive the differentiation of monocytic lineage cells - both activated microglia and macrophages - into an M2c phenotype to enable a similar wound healing response to that identified in myocardial ischemia.
Our data also demonstrates that co-localization of pSmad2 with activated astrocytes increases after stroke. There is strong in vitro and in vivo evidence that TGFβ plays an important role in generating reactive astrogliosis, and that reactive astrocytes participate in innate immune responses. Overexpression of TGFβ chronically leads to reactive astrogliosis, and these mice display exacerbated reactive astrogliosis in a stab wound model . Injection of TGF-β1 also increases reactive astrogliosis after stab wound , while injection of neutralizing antibodies inhibits expression of fibrogenic proteins, including proteoglycans, and reduces reactive astrogliosis [5, 34]. Finally, Smad3 knockout mice display faster wound healing and decreased scar formation after brain stab injury . Prior to this study no investigation addressed whether it is TGFβ signaling in astrocytes that leads to reactive astrogliosis, but we show here that astrocytes do respond to post stroke increases in TGFβ. Together this supports the hypothesis that TGFβ signaling in astrocytes directly mediates astrogliosis after stroke. Astrogliosis may have a detrimental effect on recovery due to astrocytes extending processes in the infarct border to encompass the developing infarct, thereby inhibiting neuronal plasticity and the formation of new axons and blood vessels in the infarcted region .
As the brain ages it becomes characterized by a low grade chronic pro-inflammatory state and responds with an earlier and stronger innate immune response following stroke . We found that there is increased TGFβ signaling in the aged brain in the absence of injury and this may be part of an anti-inflammatory strategy employed by the aging brain to counter the primed pro-inflammatory state. Our findings are consistent with others who have shown that in aging the expression of TGFβ receptors increases, as does the expression of TGF-β1 [2, 20, 36, 37]. Conversely, since the data we present here strongly suggests that TGFβ signaling is important for neuronal function, increased TGFβ signaling in the aged brain may be required to maintain normal neuronal function in the face of age related changes. More research is needed to determine which of these scenarios is correct.
These results show that TGFβ signaling in the brain does not oscillate with the estrus cycle and that it increases 2 fold after the dMCAO model of stroke in young and old animals of each gender. Astrocytes and CD68+ innate immune cells are the cell types that respond to increased levels of TGFβ after stroke and they do so during the first week after stroke, coincident with glial scar formation. Given TGFβs effects on monocytic lineage cells and astrocytes we hypothesize that the role of increased TGFβ signaling after stroke is to regulate glial scar formation and the polarization of invading innate immune cells and resident microglia.
We thank Jullet Han and Geoffrey Stanley for valuable technical assistance and Dr Tony Wyss-Coray for providing us with SBE-LucRT mice. This work was supported by an Ellison Medical Foundation/AFAR Postdoctoral Fellowship to KPD and KO8 NS050304-01A2 to MSB.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.