Stroke produces a limited process of neural repair. Axonal sprouting in cortex adjacent to the infarct is part of this recovery process, but the signal that initiates axonal sprouting is not known. Growth and differentiation factor 10 (GDF10) is induced in peri-infarct neurons in mice, non-human primates and humans. GDF10 promotes axonal outgrowth in vitro in mouse, rat and human neurons through TGFβRI and TGFβRII signaling. Using pharmacogenetic gain- and loss-of-function studies, we found that GDF10 produced axonal sprouting and enhanced functional recovery after stroke; knocking down GDF10 blocked axonal sprouting and reduced recovery. RNA sequencing from peri-infarct cortical neurons revealed that GDF10 downregulated PTEN, upregulated PI3 kinase signaling and induced specific axonal guidance molecules. Using unsupervised genome-wide association analysis of the GDF10 transcriptome, we found that it was not related to neurodevelopment, but may partially overlap with other CNS injury patterns. Thus, GDF10 is a stroke-induced signal for axonal sprouting and functional recovery.
At a glance
- Extensive cortical rewiring after brain injury. J. Neurosci. 25, 10167–10179 (2005). et al.
- In vivo voltage-sensitive dye imaging in adult mice reveals that somatosensory maps lost to stroke are replaced over weeks by new structural and functional circuits with prolonged modes of activation within both the peri-infarct zone and distant sites. J. Neurosci. 29, 1719–1734 (2009). , , , &
- An age-related sprouting transcriptome provides molecular control of axonal sprouting after stroke. Nat. Neurosci. 13, 1496–1504 (2010). et al.
- A role for ephrin-A5 in axonal sprouting, recovery and activity-dependent plasticity after stroke. Proc. Natl. Acad. Sci. USA 109, E2230–E2239 (2012). et al.
- Upper limb recovery after stroke is associated with ipsilesional primary motor cortical activity: a meta-analysis. Stroke 45, 1077–1083 (2014). et al.
- Rewiring the brain: potential role of the premotor cortex in motor control, learning, and recovery of function following brain injury. Neurorehabil. Neural Repair 26, 282–292 (2012). , , &
- Structural and functional plasticity in the somatosensory cortex of chronic stroke patients. Brain 129, 2722–2733 (2006). , , , &
- Neuronal intrinsic mechanisms of axon regeneration. Annu. Rev. Neurosci. 34, 131–152 (2011). , , &
- Targeting the neural extracellular matrix in neurological disorders. Neuroscience 253, 194–213 (2013). , , &
- Neuronal repair. Asynchronous therapy restores motor control by rewiring of the rat corticospinal tract after stroke. Science 344, 1250–1255 (2014). et al.
- Growth/differentiation factor-10: a new member of the transforming growth factor-beta superfamily related to bone morphogenetic protein-3. Growth Factors 12, 99–109 (1995). et al.
- Comparative integromics on BMP/GDF family. Int. J. Mol. Med. 17, 951–955 (2006). &
- Bone morphogenetic proteins: facts, challenges, and future perspectives. J. Dent. Res. 93, 335–345 (2014). et al.
- Targeting the TGFβ signalling pathway in disease. Nat. Rev. Drug Discov. 11, 790–811 (2012). &
- Stem cell antigen-1 enhances tumorigenicity by disruption of growth differentiation factor-10 (GDF10)-dependent TGF-beta signaling. Proc. Natl. Acad. Sci. USA 108, 7820–7825 (2011). et al.
- Localized expression of BMP and GDF mRNA in the rodent brain. J. Neurosci. Res. 56, 482–492 (1999). &
- Characterization of GDF-10 expression patterns and null mice. Dev. Biol. 212, 68–79 (1999). , &
- Macrophage-derived factors stimulate optic nerve regeneration. J. Neurosci. 23, 2284–2293 (2003). et al.
- Losartan prevents acquired epilepsy via TGF-b signaling suppression. Ann. Neurol. 75, 864–875 (2014). et al.
- Pharmacological inhibition of transforming growth factor β signaling decreases infection and prevents heart damage in acute Chagas' Disease. Antimicrob. Agents Chemother. 53, 4694–4701 (2009). et al.
- Transforming growth factor-beta 1 and -beta 2 promote neurite sprouting and elongation of cultured rat hippocampal neurons. Brain Res. 639, 21–25 (1994). , &
- TGF-beta 1 enhances neurite outgrowth via regulation of proteasome function and EFABP. Neurobiol. Dis. 38, 395–404 (2010). et al.
- Secretory leukocyte protease inhibitor reverses inhibition by CNS myelin, promotes regeneration in the optic nerve, and suppresses expression of the transforming growth factor-beta signaling protein Smad2. J. Neurosci. 33, 5138–5151 (2013). et al.
- TGFbeta-Smad2 signaling regulates the Cdh1-APC/SnoN pathway of axonal morphogenesis. J. Neurosci. 28, 1961–1969 (2008). , , , &
- Human embryonic stem cells: an in vitro model to study mechanisms controlling pluripotency in early mammalian development. Stem Cell Rev. 1, 119–130 (2005). &
- Molecular analyses of human induced pluripotent stem cells and embryonic stem cells. Cell Stem Cell 7, 263–269 (2010). , , &
- Using iPSC-derived neurons to uncover cellular phenotypes associated with Timothy syndrome. Nat. Med. 17, 1657–1662 (2011). et al.
- Multimodal examination of structural and functional remapping in the mouse photothrombotic stroke model. J. Cereb. Blood Flow Metab. 33, 716–723 (2013). et al.
- AMPA receptor–induced local brain-derived neurotrophic factor signaling mediates motor recovery after stroke. J. Neurosci. 31, 3766–3775 (2011). et al.
- Growth factor and cytokine regulation of chondroitin sulfate proteoglycans by astrocytes. Glia 52, 209–218 (2005). &
- Transforming growth factor β-regulated microRNA-29a promotes angiogenesis through targeting the phosphatase and tensin homolog in endothelium. J. Biol. Chem. 288, 10418–10426 (2013). et al.
- The vermicelli and capellini handling tests: simple quantitative measures of dexterous forepaw function in rats and mice. J. Vis. Exp. 41, 2076 (2010). et al.
- Reducing excessive GABA-mediated tonic inhibition promotes functional recovery after stroke. Nature 468, 305–309 (2010). , , , &
- A lifespan analysis of intraneocortical connections and gene expression in the mouse I. Cereb. Cortex 21, 1311–1330 (2011). , &
- Sustained axon regeneration induced by co-deletion of PTEN and SOCS3. Nature 480, 372–375 (2011). et al.
- Translating the frontiers of brain repair to treatments: starting not to break the rules. Neurobiol. Dis. 37, 237–242 (2010).
- Inosine alters gene expression and axonal projections in neurons contralateral to a cortical infarct and improves skilled use of the impaired limb. J. Neurosci. 29, 8187–8197 (2009). et al.
- Cloning of human GAP-43: growth association and ischemic resurgence. Neuron 1, 133–139 (1988). , , , &
- Transforming growth factor-beta 1 promotes re-elongation of injured axons of cultured rat hippocampal neurons. Brain Res. 723, 206–209 (1996). , , &
- TGF-beta signaling specifies axons during brain development. Cell 142, 144–157 (2010). , , , &
- TGF-beta signaling is required for maintenance of retinal ganglion cell differentiation and survival. Neuroscience 189, 123–131 (2011). , &
- Transcriptome profiling of a TGF-beta-induced epithelial-to-mesenchymal transition reveals extracellular clusterin as a target for therapeutic antibodies. Oncogene 29, 831–844 (2010). et al.
- Proof of concept: pharmacological preconditioning with a Toll-like receptor agonist protects against cerebrovascular injury in a primate model of stroke. J. Cereb. Blood Flow Metab. 31, 1229–1242 (2011). et al.
- Cerebral microinfarcts associated with severe cerebral beta-amyloid angiopathy. Brain Pathol. 20, 459–467 (2010). et al.
- Isolation and culture of adult neurons and neurospheres. Nat. Protoc. 2, 1490–1498 (2007). &
- From skin biopsy to neurons through a pluripotent intermediate under Good Manufacturing Practice protocols. Stem Cells Transl. Med. 1, 36–43 (2012). et al.
- A modified technique for culturing primary fetal rat cortical neurons. J. Biomed. Biotechnol. 2012, 803930 (2012). , , , &
- Defining the nature of human pluripotent stem cell progeny. Cell Res. 22, 178–193 (2012). et al.
- let-7 miRNAs can act through notch to regulate human gliogenesis. Stem Cell Reports 3, 758–773 (2014). et al.
- Human neural stem cells enhance structural plasticity and axonal transport in the ischaemic brain. Brain 134, 1777–1789 (2011). et al.
- IGF-I specifically enhances axon outgrowth of corticospinal motor neurons. Nat. Neurosci. 9, 1371–1381 (2006). &
- The Mouse Brain in Stereotaxic Coordinates 2nd edn. (Academic, San Diego, 2001). &
- STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013). et al.
- edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010). , &
- Supplementary Figure 1: TGFβ superfamily (114 KB)
(a) TGFβ superfamily members are classified into TGFβ, Activin/inhibin/Nodal group and BMP groups. In the canonical signaling pathway, shown here, TGFβ family members signal through type I and type II receptors for each member to receptor-associated SMAD proteins (R-SMAD), SMAD2 and SMAD3. Activin/inhibin/Nodal proteins signal through distinct type I and II receptors: ACVR1B or ALK-4 acts as a transducer of Activin/inhibin/Nodal signals. These Activin ligands bind to either ACVR2A or ACVR2B and then bind ACVR1B and then activate (phosphorylate) SMAD2/SMAD3. BMP signals are transduced through their distinct type I and type II receptors for each member to R-SMAD proteins SMAD1, SMAD5 and SMAD8. Phosphorylated R-SMADs associate with SMAD4 and are then translocated to the nucleus to activate transcription of target genes. GDF family members signal through both Activin and BMP pathways, except for GDF10, which signals through TGFβ receptors. This summary is simplified and, for example, does not represent the full complexity of signaling that can come through heterotetrameric binding between TGFβ, Activin/inhibin/Nodal and BMP families. (b) The genetic relationship of GDF/BMP proteins as established by sequence homology. Figures modified from12,14.
- Supplementary Figure 2: GDF10 is upregulated in peri-infarct cortex. (842 KB)
(a) GDF10 protein expression in the ipsilateral hemisphere after stroke. This low magnification photomicrograph of the ipsilesional hemisphere shows axonal loss, as indicated by loss of MAP2, in the stroke core. GDF10 is induced in the cortex bordering the stroke core, as seen by the increased intensity of staining (arrows). The stroke core contains an abundance of microglia and macrophages, labeled by CD11b. (b) GDF10 expression in the peri-infarct cortex. Image is taken from the region of interest indicated by box in panel (a) GDF10 protein expression is upregulated compared to contralateral control in panel 1(c). (b) Higher magnification view of peri-infarct cortex shows that GDF10 is present in MAP2+ neuronal dendrites (arrows) in addition to neuronal cell bodies (Fig. 1) and not present in CD11b+ microglia/macrophages. (c) GDF10 expression in the non-injured cortex from the contralateral hemisphere. GDF10 has a sparse low level of endogenous expression. This indicates that GDF10 induction is specific to cortex immediately adjacent to the site of stroke injury. In the absence of stroke, CD11b is not induced in the control cortex. (d) GDF10 expression in the peri-infarct colocalizes to neurons labeled with MAP2 and excludes microglia and macrophages. GDF10 immunoreactivity is present in MAP2+ neuronal somas and in dendrites (arrows). GDF10 does not colocalize with CD11b+ microglia/ macrophages. (e) Secreted GDF10 is also found in the extracellular space in the peri-infarct cortex near surviving processes (MAP2).
- Supplementary Figure 3: GDF10 protein levels and siRNA knockdown in vivo, P4 neuron outgrowth, and siRNA protein knockdown (89 KB)
(a) GDF10 siRNA, vehicle or scrambled siRNA was delivered into the stroke cavity and peri-infarct tissue processed for Western blot. X axis shows days after stroke and siRNA delivery. * = p< 0.05, ** = p< 0.01 compared to the scrambled siRNA; # = p<0.05 compared to Stroke only. Lower panel shows representative Western blot images from (a). N = 3 for each experiment. (b) Both myelin and CSPGs inhibit the outgrowth or P4 cortical neurons. *** = p<0.001, multiple comparisons ANOVA, Tukey-Kramer post-hoc. (c) Western blot of GDF10 protein level with two-day treatment of siRNA in P4 neurons. Numbers in the columns indicate the distinct siRNA construct. Lower panel shows a representative blot from each experiment. Each experiment represents 4 samples and 2 technical replicates. (d) (b) Western blot results for knockdown of TGFβRI, II and Smads1, 2, 3, 5. Same conventions as in (a). ** = p<0.01 compared to the scrambled siRNA.
- Supplementary Figure 4: Primary cortical neurons plated in vitro express GDF10 (225 KB)
(a) Primary cortical neurons stained with GDF10 and MAP2 antibody demonstrates neuronal expression of GDF10 in vitro. This substantiates the in vitro GDF10 knockdown experiments using siRNAs to study axonal outgrowth, without a need for addition of GDF10 to the system to mimic induction. In contrast to the low endogenous expression of GDF10 in uninjured cortical neurons in vivo, GDF10 expression in vitro is likely induced by mechanical stress of dissection and plating of primary neurons. (b) High magnification of neuronal soma expressing GDF10. GDF10 immunoreactivity is also evident along neuronal processes (arrows).
- Supplementary Figure 5: pSmad2/3 quantification in peri-infarct cortex after TGFβ antagonism in vivo (79 KB)
Stroked animals were treated with two different TGFβ antagonists, SB431542 and Losartan at 10mg/kg and 100mg/kg, respectively, based on published i.p. doses16,17 for 5 days after stroke (n=3 mice per group). pSmad puncta of 0.45μm were quantified at 100x imaging fields in two sections per animal, and mean puncta values were statistically compared. As seen with in vitro axon outgrowth studies in primary cortical neurons from mouse (Fig. 2) and human iPSC-derived neurons (Fig. 3), in vivo administration of TGFβ antagonists, SB431542 (p=0.0056) and Losartan (p=0.0244), each significantly decreases pSmad2/3 signaling within 300µm of the infarct core. *=p<0.05.
- Supplementary Figure 6: Quantitative connectional mapping and axonal connections in peri-infarct cortex (81 KB)
(a) Left panel. Quantitative connectional map of neurons back-labeled from a retrograde tracer injection (cholera toxin b subunit) into cervical level 5 spinal cord (n = 5). This labels all neurons in motor, somatosensory and premotor cortex that send projections to the spinal cord. The location of all neurons is plotted in tangential sections and the x y coordinates are made relative to bregma49. Middle panel. Axonal label plotted from BDA injections into forelimb motor cortex in stroke+GDF10 and Cyto C+stroke (protein control). This is the same experiment as in Fig 4a. Right panel: alignment of corticospinal projections with GDF10/stroke map, to show location of areas of axonal sprouting relative to motor, premotor and somatosensory areas. Note that a substantial region of new connections formed in Stroke+GDF10 is in cortex rostral to the corticospinal populations of neurons. (b) The effect of stroke on motor cortex connections. Left panel: quantitative connectional map of connections from tracer injections into forelimb motor cortex in stroke only (red label) and in control, non-stroke (light blue label) mice. The cortical areas with dense overlap of connections in stroke and control are in dark blue. Note the presence of connections only in the stroke condition in posterior cortical areas that localize to motor cortex and second somatosensory cortex in registering to the corticospinal map and to primary somatosensory cortex by cytochrome oxidase stain (not shown). P value is significant (Hotellings T2 test). Middle panel: Polar plot of connections in stroke (red) and control (blue). Conventions as in Fig. 4b. Right panel: quantification of neuronal connections in linear array through cortical hemisphere.
- Supplementary Figure 7: Infarct volume, BDA volume and injection locations. (163 KB)
(a) Left Y axis: BDA injection volume, right Y axis infarct volume. Columns are for in vivo axonal tracing studies. There are no significant differences among groups. (b) Location of BDA injections for each group relative to the midline of the cortical hemisphere and to the rostral pole of the brain. There are no significant differences among groups. Error bars are SEM.
- Supplementary Figure 8: Quantitative connectional mapping and axonal connections in GDF10 controls compared with normal control. (75 KB)
(a-c) Cortical connections of forelimb motor cortex in normal control animals only compared to stroke+protein control, cytochrome C delivery from the infarct core via biopolymer hydrogel. (n = 8 for each group). Figure conventions are as in Supplementary Figure 8 and Figure 4. There is a significant difference in the cortical connectional map (a), polar plots (b) and linear measurement of neuronal connections (c) between normal control forelimb motor cortex and stroke+scrambled siRNA. This indicates that the normal mode axonal sprouting that is seen in stroke compared to control brains is also seen in the control conditions for GDF10 delivery. (d-f) Cortical connections of forelimb motor area in normal control compared to scrambled siRNA delivered into infarct core. siRNA GDF10 knocks down GDF10 protein levels (Supplementary Fig 5a) and blocks post-stroke axonal sprouting (Fig. 4d-f). To verify that the siRNA is not having effects outside of GDF10 knock down, scrambled sequence siRNA was tested. Figure conventions are as in Supplementary Fig. 8 and Fig. 4. There is a significant difference in cortical connection map (a), polar plots (b) or linear measurement of neuronal connections (c) between control and scrambled siRNA (n = 7-8 for each group). * = p<0.05. Note that Cyto C+stroke and scrambled siRNA+stroke produce very similar maps of forelimb motor cortex connections, and are comparable to stroke only (Supplementary Fig. 6b).
- Supplementary Figure 9: Quantitative connectional mapping and axonal connections in GDF10 controls compared with stroke only. (76 KB)
(a-c) Cortical connections of forelimb motor area in stroke only compared to scrambled siRNA delivered into infarct core. siRNA GDF10 knocks down GDF10 protein levels (Supplementary Fig 3a) and blocks post-stroke axonal sprouting (Fig. 4d-f). To verify that the siRNA is not having effects outside of GDF10 knock down, scrambled sequence siRNA was tested. Figure conventions are as in Supplementary Fig. 8 and Fig. 4. There is no significant difference in cortical connection map (a), polar plots (b) or linear measurement of neuronal connections (c) between stroke and scrambled siRNA. (d-f) Cortical connections of forelimb motor cortex in stroke only compared to stroke+protein control, cytochrome C delivery from the infarct core via biopolymer hydrogel (n = 8 each group). Figure conventions are as in Supplementary Fig 8 and Figure 4. There is no significant difference in cortical connection map (a), polar plots (b) or linear measurement of neuronal connections (c) between stroke and Cyto C. Cohorts (n=8) of mice treated with GDF10 (red) or control protein (Cytochrome C) (light blue).
- Supplementary Figure 10: Synaptic protein localization in peri-infarct cortex axons that have undergone axonal sprouting after GDF-10 treatment. (213 KB)
(a) Synaptic protein analysis was performed in the same peri-infarct tissues from animals used for BDA axonal sprouting maps in Fig. 4(a) and (b). Presynaptic VGLUT2 and postsynaptic Homer1 antibodies were used for identification of synaptic contacts. Marker colocalization analyses were performed on Imaris Imaging software to uniquely identify the synaptic connections formed by GDF10-induced sprouting cortical neurons after stroke (mapped in Fig. 4). (b) Video through a 10.5μm thick section of peri-infarct cortex taken at 100x. BDA surface is shown in light blue. VGLUT2 presynaptic marker is shown by green spots, and Homer1 postsynaptic marker in red spots.
- Supplementary Figure 12: Pipeline for incorporating microarray and RNA-seq data sets for neurodevelopmental and CNS injury experiments. (54 KB)
Raw data from different platforms are processed to have gene symbols which were subsequently used to merge the datasets. Merged datasets were normalized, then batch effect was adjusted. In one case, the datasets have very few common genes on different array platforms and these datasets were combined (unionized) instead of taking intersections.
- Supplementary Figure 13: Neuronal numbers in in vitro axonal outgrowth assays. (124 KB)
The numbers of neurons used for each in vitro experiment to generate the data on axonal outgrowth effect of GDF10 and other experimental manipulations in Figures 2 and 3. There is no significant difference in neuronal sampling number across experiments.
- Video 1: Synaptic proteins identified in axons that have undergone axonal sprouting in peri-infarct cortex after GDF10 treatment. (93.59 MB, Download)
- This movie shows the process of identifying pre- and post-synaptic connections that are present in axons that have undergone axonal sprouting with GDF10 after stroke. The movie is derived from a confocal image stack taken from a BDA-labelled area in premotor cortex that has axons only when axonal sprouting is induced with GDF10 delivery. The BDA-labelled axons are seen at the outset of the movie (white axonal label). This is then filamented (blue label). The tissue has been stained for Homer (post-synaptic, green) and VGlut2 (a subset of excitatory pre-synaptic terminals, red) and filtered so that only those VGlut2 positive pre-synaptic terminals that co-localize with BDA-filled axons are present. Further, the Homer positive post-synaptic boutons are filtered so that only those that co-localize with the VGlut2 that is within BDA-labelled axons are displayed. The movie that zooms and rotates to show the 3D distribution of this network of axonal connections.
- Supplementary Text and Figures (1,892 KB)
Supplementary Figures 1–13
- Supplementary Table 1 (30 KB)
Canonical Pathways Mult Comp Stroke+GDF10 vs Stroke
- Supplementary Table 2 (8 KB)
PI3K Pathway Mult Comp Stroke+GDF10 vs Stroke
- Supplementary Table 3 (8 KB)
PTEN Pathway Mult Comp Stroke+GDF10 vs Stroke
- Supplementary Table 4 (11 KB)
PI3K Pathway Mult Comp Stroke+GDF10 vs Stroke
- Supplementary Table 5 (11 KB)
PTEN Pathway Mult Comp Stroke+GDF10 vs P4
- Supplementary Table 6 (9 KB)
Axonal Guidance Molecules Differentially Regulated in Stroke+GDF10 vs Stroke
- Supplementary Table 7 (48 KB)
Genes Differentially Regulated in Stroke+GDF10 vs Stroke at FDR<0.1
- Supplementary Table 8 (11 KB)
Sources for Genome Wide Assocation Analysis
- Supplementary Table 9 (9 KB)
- Supplementary Table 10 (10 KB)
Multiple Comparsions ANOVA and Tukey-Kramer Test for Differences Between Means in Figure 5
- Supplementary Table 11 (11 KB)