Abstract
Multipotent adult neural precursor cells (NPCs) have tremendous intrinsic potential to repair the damaged spinal cord. However, evidence shows that the regenerative capabilities of endogenous and transplanted NPCs are limited in the microenvironment of spinal cord injury (SCI). We previously demonstrated that injury-induced upregulation of matrix chondroitin sulfate proteoglycans (CSPGs) restricts the survival, migration, integration, and differentiation of NPCs following SCI. CSPGs are long-lasting components of the astroglial scar that are formed around the lesion. Our recent in vivo studies demonstrated that removing CSPGs from the SCI environment enhances the potential of transplanted and endogenous adult NPCs for spinal cord repair; however, the mechanisms by which CSPGs regulate NPCs remain unclear. In this study, using in vitro models recapitulating the extracellular matrix of SCI, we investigated the direct role of CSPGs in modulating the properties of adult spinal cord NPCs. We show that CSPGs significantly decrease NPCs growth, attachment, survival, proliferation, and oligodendrocytes differentiation. Moreover, using genetic models, we show that CSPGs regulate NPCs by signaling on receptor protein tyrosine phosphate sigma (RPTPσ) and leukocyte common antigen-related phosphatase (LAR). Intracellularly, CSPGs inhibitory effects are mediated through Rho/ROCK pathway and inhibition of Akt and Erk1/2 phosphorylation. Downregulation of RPTPσ and LAR and blockade of ROCK in NPCs attenuates the inhibitory effects of CSPGS. Our work provide novel evidence uncovering how upregulation of CSPGs challenges the response of NPCs in their post-SCI niche and identifies new therapeutic targets for enhancing NPC-based therapies for SCI repair. Stem Cells 2015;33:2550–2563
Introduction
Spinal cord injury (SCI) results in limited spontaneous tissue regeneration, despite the existence of tissue specific neural precursor cells (NPCs) residing inside the adult spinal cord [1, 2]. Although showing multipotentiality in vitro, recent studies in SCI models demonstrated that without manipulation, NPCs predominantly give rise to astrocytes, and their potential for oligodendrocyte differentiation is restricted following injury [3-8]. Also challenging is the limited migration and long-term survival of NPCs in the injured spinal cord [3, 4]. This evidence suggests that the properties of NPCs are negatively modulated by changes in their post-SCI niche. Following injury, extracellular matrix (ECM) undergoes drastic changes mainly driven by activated astrocytes [9-11]. Upregulation of inhibitory chondroitin sulfate proteoglycans (CSPGs) in the postinjury matrix potently impede axon regeneration and plasticity in chronic stages of injury [12]. Removal of CSPGs with chondroitinase ABC (ChABC) improves axonal regeneration, preservation, sprouting, remyelination, conduction, and functional recovery following SCI [4, 13-21].
We have recently shown that SCI-induced upregulation of CSPGs additionally pose a barrier to cell replacement repair strategies for SCI [4, 8]. Targeting CSPGs by ChABC treatment promoted activities of endogenous precursor cells and their potential for oligodendrocyte differentiation [8] and optimized the long-term survival and migration of transplanted NPCs in subacute [22] and chronic [4] stages of SCI. In rat chronic SCI, we demonstrated that ChABC treatment prior to NPCs transplantation resulted in extensive migration of grafted NPCs and allowed their integration with the host tissue suggesting an inhibitory role for CSPGs in modulating the regenerative response of NPCs following injury [4]. Currently, the underlying mechanisms of CSPGs effects on spinal cord NPCs remain to be identified. Upregulation of CSPGs is a hallmark of Central Nervous System (CNS) injuries including SCI that persists chronically in the NPCs niche [4], thus it is imperative to understand the impact of CSPGs on NPCs activities. The recent discovery in identifying CSPG specific receptors, including protein tyrosine phosphatase receptor sigma, RPTPσ [23] and leukocyte common antigen-related phosphatase, LAR [24], allows us to uncover CSPGs cellular mechanisms.
Here, using comprehensive in vitro assays mimicking the matrix of postinjury milieu, we demonstrate that CSPGs directly inhibit NPCs properties including growth, attachment, survival, proliferation, and oligodendrocyte differentiation. CSPGs inhibitory effects are mediated through signaling on both LAR and RPTPσ and intracellualrly by the Rho/ROCK pathway. Additionally, CSPGs reduce Akt and Erk1/2 phosphorylation in NPCs which can be overcome by downregulation of LAR and RPTPσ or ROCK inhibition. Our new findings unravel a key role for CSPGs in regulating NPCs and thereby identify new potential therapeutic options to effectively optimize NPC-based cell replacement strategies for CNS repair.
Materials and Methods
Animals
All animal procedures were approved by the Animal Ethics Committee of the University of Manitoba in accordance with the policies established by the Canadian Council of Animal Care (CCAC). NPCs were harvested from five C57BL/6, seven BALB/c mice (8 weeks of age, Central Animal Facility, University of Manitoba, Canada), and eight RPTPσ−/− mice (12 weeks of age, provided by Michel L. Tremblay, McGill University) [25]. Genotypes of RPTPσ−/− mice were verified using polymerase chain reaction (PCR) (products of 781 or 1,000 bp). Three adult Sprague Dawley rats (250 g) were used for immunohistological analysis.
Isolation and Culturing of Adult NPCs
Adult NPCs were isolated from the spinal cord and brain of adult mice as we described previously [3, 4, 26] (Supporting Information).
Plating NPCs on Laminin and CSPGs Substrates
Neurospheres were dissociated into single cells and plated onto coated tissue culture surfaces (12,000 cells/cm2 Sigma, St. Louis, MO, https://www.sigmaaldrich.com) under different conditions including (a) laminin (Sigma), (b) laminin+CSPGs (Millipore, Billerica, MA, http://www.emdmillipore.com), (c) laminin+CSPGs pretreated with chondroitinase ABC (ChABC, Sigma), (d) laminin+ChABC in the media (Supporting Information). The cells were grown in NPC growth medium containing epidermal growth factor and fibroblast growth factor-2 (Sigma) for 2 days assessment or switched to serum medium (SFM plus 2% fetal bovine serum (FBS, Invitrogen, Carlsbad, CA, http://www.invitrogen.com) 24 hours following cell plating to induce differentiation (7 days assessments). All in vitro assays in this study were conducted in four independent experiments (N = 4).
Immunocytochemistry and Western Blotting Procedures
Dissociated NPCs were plated on laminin and/or CSPG-coated multichamber glass slides (25,000 cells per chamber) (LabTek II Thermo Fisher, Waltham, MA, http://www.thermofisher.com) either in growth medium for 2 days assessment or in 2% FBS medium for 7 days differentiation assay. For proliferation assay, bromodeoxyuridine (BrdU 20 µM, Sigma) was added to the cultures 3 hours before processing NPCs for immunocytochemistry (Supporting Information). For Western blotting, cells were harvested from culture plates and homogenized in RIPA buffer (Thermo Fisher, Waltham, MA, http://www.thermofisher.com) containing SigmaFast Protease Inhibitor (Sigma). A total of 10–50 µg of protein was loaded onto the gel and transferred to a nitrocellulose membrane (BioRad, Hercules, CA, http://www.bio-rad.com) for electrophoresis. Samples from the same experiment were processed and run on the same gel, and bands were quantified accordingly. β-Actin was used as an internal control for protein loading (Supporting Information; Supporting Information Table 1 for details and list of antibodies).
Dicer Substrate Interference RNA Procedures to Knockdown RPTPσ and LAR
We used Dicer substrate interference RNA (DsiRNA) from Integrated DNA Technologies (IDT, Coralville, IA, http://www.idtdna.com) to downregulate RPTPσ and LAR expression in NPCs. NC1 nontarget DsiRNA was used as a control. Transfection protocol, primer, and DsiRNA sequences are provided in Supplementary Information and in Supporting Information Tables 2 and 3. DsiRNA experimental conditions were: (a) NC1, (b) LAR, (c) RPTPσ, (d) LAR + RPTPσ. Five days following DsiRNA administration, neurospheres were dissociated and harvested for RNA and protein analysis for transfection efficiency (Supporting Information) or plated onto laminin and/or CSPGs growth substrate.
Inhibition of Rho/ROCK Pathway with Y-27632
ROCK inhibitor Y-27632σ was used in NPC cultures. Neurospheres were dissociated into single cells and pretreated with Y-27632 (10 μM) for 1 hour before plating on laminin and/or CSPG-coated tissue culture.
Assessment of NPCs Attachment, Growth, and Survival
Following 48 hours of NPCs plating on laminin and/or CSPGs, cells were fixed with 3% paraformaldehyde. Using StereoInvestigator Cavalieri probe (MBF Bioscience, Williston, VT, http://www.mbfbioscience.com/), we randomly measured the total area of cells and their processes in 8–10 separate bright field images (under ×20 objective) containing an average of 300 cells for each treatment condition (Supporting Information). Survival assessment was completed through both MTT (3-(4, 5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide) and LIVE/DEAD assays at both 2 days and 7 days following cell plating (Supporting Information).
Image Processing and Analysis
For immunocytochemistry quantification, 8–10 separate fields (under ×20 objective) containing an average of 300 cells for each condition was randomly imaged (Zeiss AxioObserverZ1 or Imager2 microscope). For each condition, the total number of DAPI (4',6-diamidino-2-phenylindole) positive cells was first assessed, and the number of positive cells for Olig2, GFAP, Ki67, and BrdU (containing a DAPI positive nucleus) were then counted. The percentage of abundance for each cell type was calculated by dividing the number of positive cells for the marker by the total number of DAPI positive cells under each experimental condition. Values were then normalized to control condition for relative comparison.
Statistical Analysis
Data are reported as means ± SEM, and p ≤ .05 was considered significant. Statistical analyses of intensity measurements and cell counts were tested by one-way ANOVA comparing conditions followed by post hoc pairwise multiple comparison testing by the Holm–Sidak method. For DsiRNA knockdown experiments, where two conditions were compared, a statistical t test was used.
Results
CSPGs Negatively Modulate the Matrix Attachment, Spreading, Survival, and Proliferation of Spinal Cord NPCs
Dissociated primary adult mouse spinal cord NPCs were grown onto substrates containing either laminin or a combination of laminin and CSPGs (laminin + CSPG) for 2 days (Fig. 1A–1D). Laminin and CSPGs are highly upregulated in the ECM of the injured spinal cord and therefore laminin+CSPGs condition would more closely represent the milieu of SCI [27, 28]. We used CSPGs substrate containing a mixture of neurocan, phosphocan, versican, and aggrecan that are present in the ECM of SCI [28]. NPCs exposed to laminin+CSPGs showed a significant decrease in their ability to attach and extend their cell processes (Fig. 1B) compared to NPCs grown on control laminin substrate (Fig. 1A). Our quantifications showed a 47% decrease in the total number of attached DAPI+ cells in laminin+CSPGs condition (Fig. 1F). Moreover, we found a 36% decrease in the total occupied area per NPC on laminin+CSPGs substrate indicating that CSPGs not only limit cell attachment but also cell spreading and growth of the attached NPCs (Fig. 1E). The specificity of CSPGs effects was confirmed by ChABC pretreatment that is known to remove CSPGs functional properties in vitro and in vivo [4, 8, 13, 29].
CSPGs limit the attachment, growth, survival, and proliferation of spinal cord neural precursor cells (NPCs). (A–D): Bright field images of spinal cord NPCs grown on laminin (10 μg/ml) or laminin (10 μg/ml) + CSPGs (5 µg/ml) substrate 2 days following cell plating. NPCs grown on laminin+CSPGs substrate (B) showed limited attachment compared to laminin (A). (C): Effects of CSPGs were shown to be specific with ChABC digestion of CSPGs (0.1 U/ml) prior to cell plating. (E): A significant decrease in the total cell process area of NPCs on laminin+CSPGs substrate was observed in comparison to all other experimental groups. (F): Additionally, a significant decrease in the number of DAPI+ cells in laminin+CSPGs experimental groups demonstrates that CSPGs limit NPCs attachment. (G): Survival of NPCs on CSPGs substrate was assessed using an MTT assay 2 days following cell plating showing a significant decrease in NPCs survival on CSPGs substrate. (H): Assessment of NPC survival was complimented with a LIVE/DEAD assay 2 days following cell plating. Green cells were labeled as LIVE cells (Calcein) and red as DEAD cells (EthD-1). There was a significant decrease in the percentage of live NPCs when grown on laminin+CSPGs substrate. Inhibitory effect of CSPGs on proliferation of spinal cord NPCs was determined by BrdU (I–M) and Ki67 immunostaining (N–R). A significant downregulation in the percentage of BrdU+ and Ki67+ cells was observed when NPCs were grown on laminin+CSPGs substrate in comparison to laminin alone. N = 4 independent experiments. The data show the mean ± SEM. *, p < .05, one-way ANOVA. Abbreviations: BrdU, bromodeoxyuridine; ChABC, chondroitinase ABC; CSPG, chondroitin sulfate proteoglycan; DAPI, 4',6-diamidino-2-phenylindole.
We next assessed whether CSPGs affect the capacity of adult spinal cord NPCs to survive and proliferate. Using an MTT assay, we examined the metabolic activity of NPCs which were attached to the growth substrate as well as those unattached and floating in the media at 2 days (Fig. 1G) and 7 days (Supporting Information Fig. 1A) post-plating. NPCs cultured on laminin+CSPGs substrate demonstrated a significant 38% and 49% decrease in survival at 2 days and 7 days, respectively, in comparison to NPCs plated on laminin alone. Using a complementary LIVE/DEAD assay, we examined the viability of attached cells on the culture substrate (Supporting Information Fig. 1C–1J). NPCs under laminin condition showed 86.31% and 93.45% viability at 2 days (Fig. 1H) and 7 days (Supporting Information Fig. 1B), respectively, whereas NPCs on laminin+CSPGs substrate demonstrated a significant decrease in NPCs survival at 2 days (68.71%). However, we found no difference at 7 days in the percentage of viable attached cells (92.75%) in our LIVE/DEAD assay, which may be due to the fact that dead cells may not remain attached to the growth substrate and float in the medium in CSPGs condition.
Ability of NPCs to proliferate following an injury is a prerequisite for their successful regenerative response. Therefore, we examined how CSPGs influence NPCs proliferation using complementary quantitative BrdU incorporation and Ki67 immunostaining assays (Fig. 1I–1R). Our data revealed a greater number of proliferating BrdU+/DAPI+ cells among NPCs grown on control laminin (43.55%), which was significantly reduced in NPCs plated on laminin+CSPGs (32.73%). Similarly, a significant reduction in the percentage of Ki67+/DAPI+ cells was observed among NPCs grown on laminin+CSPGs substrate (29.19%) in comparison to our laminin control group (42.71%). CSPGs effects on survival and proliferation of NPCs were reversed by ChABC confirming specificity of our data. All values are provided in Supporting Information Table 4.
CSPGs Limit the Capacity of Spinal Cord NPCs for Oligodendrocyte Differentiation
Following SCI, oligodendrocytes are subject to degeneration, and their replacement is essential for axon remyelination [3, 30]. We show that CSPGs drive NPCs to an astrocytic fate and limit their differentiation along an oligodendrocytic lineage. We focused on astrocyte and oligodendrocyte differentiation of spinal NPCs as previous in vivo studies by our group and others showed that neuronal differentiation is a rare event in the milieu of spinal cord [3-6, 26]. Exposure to CSPGs resulted in a significant 19% increase in the ratio of glial fibrillary acidic protein (GFAP) positive astrocytes and instead a 47% decrease in the ratio of Olig2 positive oligodendrocytes compared to NPCs were grown on laminin (Fig. 2A–2E, 2G–2K 2',3'-Cyclic-nucleotide 3'-phosphodiesterase (CNPase, marking mature oligodendrocytes)). Complementary Western blotting for GFAP and CNPase (marking mature oligodendrocytes) also verified our immunohistochemical data showing that CSPGs restrict oligodendrocyte differentiation while promoting astrocyte differentiation (Fig. 2F, 2L). ChABC treatment removed CSPGs inhibitory effects on NPCs differentiation. All values are provided in Supporting Information Table 4.
CSPGs inhibit oligodendrocyte differentiation in neural precursor cells (NPCs) and drive their fate to astrocyte. Undifferentiated NPCs were allowed to attach to growth substrates containing laminin (10 μg/ml) or laminin (10 μg/ml) + CSPGs (5 μg/ml). Then they exposed to 2% serum media for 7 days to allow for differentiation. (A): Using quantitative immunostaining on NPCS, we measured the percent of GFAP positive astrocytes to the total number of DAPI positive cells. NPCs grown on laminin+CSPGs substrate (C) showed a significant increase in GFAP+/DAPI+ astrocytes in comparison to NPCs grown on laminin substrate alone (B). Degradation of CSPGs with ChABC (D) removed CSPGs effects on astrocyte differentiation suggesting specificity of CSPGs effects. (F): Western blotting verified a significant increase in the expression of GFAP protein in NPCs grown on laminin+CSPGs substrate in comparison to all other experimental conditions. (G): Percent of Olig2+/DAPI+ oligodendrocytes was quantified using the same immunostaining assay. NPCs grown on laminin+CSPGs substrate (I) showed a significant decrease in the percent of oligodendrocytes in comparison to laminin only control group (H). (J): Specificity of CSPGs effects was confirmed with ChABC pretreatment. (K): ChABC alone had no effects on oligodendrocyte differentiation of NPCs grown on laminin. (L): Complementary Western blotting confirmed CSPGs effects on oligodendrocyte differentiation, showing a significant decrease in CNPase protein expression in NPCs grown on laminin+CSPGs substrate compared to control laminin. N = 4 independent experiments. The data show the mean ± SEM. *, p < .05, one-way ANOVA. Abbreviations: ChABC, chondroitinase ABC; CNPase, 2',3'-cyclic-nucleotide 3'-phosphodiesterase; CSPG, chondroitin sulfate proteoglycan; DAPI, 4',6-diamidino-2-phenylindole; GFAP, glial fibrillary acidic protein.
CSPGs Modulate the Properties of NPCs by Signaling Through LAR and RPTPσ
We next investigated the cellular mechanisms involved in CSPGs effects on NPCs. We focused on two recently identified receptors for CSPGs, LAR, and RPTPσ, members of the transmembrane protein tyrosine phosphatase receptor family [23, 24]. Recent studies have implicated both LAR and RPTPσ in mediating the inhibitory effects of CSPGs on axonal growth and regeneration [23, 24, 31, 32]. Using immunostaining, we first confirmed that spinal cord NPCs express LAR and RPTPσ in vitro (Fig. 3A–3F) and in the ependymal and subependymal region of the rat spinal cord where NPCs reside (Fig. 3G–3J). We used DsiRNA strategy to downregulate LAR and RPTPσ in NPCs. Using NC1 scrambled DsiRNA as a control, we verified successful downregulation of LAR and RPTPσ transcripts and protein with quantitative PCR and Western blotting 5 days after DsiRNA transfection (Supporting Information Fig. 2A–2E). There was 88% and 71% reduction in LAR mRNA and protein expression, respectively, in NPCs transfected with LAR DsiRNA compared to control NC1 DsiRNA with no changes in RPTPσ expression. Similarly, RPTPσ DsiRNA resulted in 94% and 76% reduction in RPTPσ mRNA and protein levels, respectively, in NPCs treated with RPTPσ DsiRNA compared to NC1 DsiRNA with no changes in LAR expression. In our double knockdown condition of LAR+RPTPσ, successful downregulation of both LAR and RPTPσ mRNA and protein expression was achieved comparable to our single knockdown data. Of note, for each experimental setting that we discuss in following sections we confirmed transfection efficacy and downregulation of LAR and RPTPσ protein. We additionally used NPCs harvested from adult RPTPσ knockout mice to confirm our DsiRNA results. Our characterization of RPTPσ−/− derived NPCs confirmed lack of RPTPσ expression compared to their wild-type counterparts without any effects on LAR expression in NPCs (Supporting Information Fig. 3F–3H). Moreover, our in vitro characterization of RPTPσ−/− NPCs showed that deletion of RPTPσ itself had no apparent changes in the capacity of NPCs for self-renewal, proliferation, and differentiation compared to wild-type NPCs (Supporting Information Fig. 3A–3E).
LAR and RPTPσ mediate CSPGs inhibitory effects on attachment, growth, survival, proliferation, and oligodendrocyte differentiation of neural progenitor cells (NPCs). In vitro immunostaining verifies the expression of LAR (A–C) and RPTPσ (D–F) by nestin positive NPCs. In vivo immunohistochemical analysis of the spinal cord tissue also demonstrated the expression of LAR (G, H) and RPTPσ (I, J) in the ependymal layer of the spinal cord, where NPCs reside. (K, L): Downregulation of LAR and/or RPTPσ significantly increased the ability of NPCs to grow and attach on laminin (10 μg/ml) + CSPGs (5 μg/ml) substrate in comparison to Dicer substrate interference RNA NC1 control condition. (M): Survival of NPCs on CSPGs substrate was assessed using MTT assay 2 days following cell plating. Downregulation of RPTPσ or dual downregulation of LAR + RPTPσ significantly improved the survival of NPCs when grown on CSPGs substrate in comparison to NC1 control group. (N): Downregulation of LAR and/or RPTPσ also promoted the proliferation of NPCs on laminin+CSPGs substrate in comparison to NC1 control; however, this was only statistically significant higher in RPTPσ and LAR + RPTPσ experimental groups. (O, P): Immunostainings of NPCs grown on laminin+CSPGs substrate demonstrate both LAR and RPTPσ mediate CSPGs inhibitory effects on NPC differentiation. Downregulation of LAR or RPTPσ significantly increased the percentage of Olig2 positive cells when grown on laminin+CSPGs substrate in comparison to NC1 control group. In (M), results from each knockdown experimental condition were normalized to their respective laminin group. N = 4 independent experiments. The data show the mean ± SEM. *, p < .05, Student's t test. Abbreviations: BrdU, bromodeoxyuridine; CSPG, chondroitin sulfate proteoglycan; DAPI, 4',6-diamidino-2-phenylindole; GFAP, glial fibrillary acidic protein; LAR, RPTP, receptor protein tyrosine phosphate sigma; siRNA, dicer substrate interference RNA.
Using these two genetic approaches, we show that downregulation of LAR and RPTPσ remarkably reduces the inhibitory effects of CSPGs on the properties of NPCs. Assessment of the cell attachment and spreading of NPCs (Fig. 3K–3L, Supporting Information Fig. 4A–4H) showed significant reduction (47.7%) in the number of attached DAPI+ cells in our control NC1-treated NPCs grown on laminin+CSPGs whereas only a 24%, 15%, and 11% reduction was observed in LAR, RPTPσ, and LAR + RPTPσ DsiRNA experimental groups, respectively. Downregulation of LAR and/or RPTPσ also significantly promoted the ability of NPCs to grow and extend their processes on CSPGs substrate. Notably, dual downregulation of LAR and RPTPσ had additive effects on enhancing NPCs attachment and growth on CSPGs compared to their individual knockdown. All values are provided in Supporting Information Table 5.
Knockdown of LAR and RPTPσ receptors promoted survival of NPCs exposed to CSPGs. Complementary MTT and LIVE/DEAD assays confirmed a significant improvement in the survival of NPCs treated with LAR, RPTPσ, and LAR + RPTPσ DsiRNA when cells exposed to laminin+CSPGs substrate for 2 days and 7 days (data not shown) compared to our control DsiRNA NC1 condition (Fig 3M; Supporting Information Fig. 4K). Similarly, BrdU and Ki67 proliferation assays revealed significant improvement in proliferation of NPCs on CSPGs in conditions where LAR and/or RPTPσ were downregulated (Fig. 3N, data not shown for Ki67). We found that NPCs treated with LAR, RPTPσ, and LAR + RPTPσ DsiRNA showed comparable proliferation capacity compared to their respective counterparts grown on permissive laminin. Notably, proliferation of NPCs on CSPGs was statistically increased in RPTPσ and LAR + RPTPσ DsiRNA conditions compared to NC1 DsiRNA group (Fig. 3N) while LAR DsiRNA displayed nonsignificant increase. Moreover, using NPCs isolated from RPTPσ−/− mice, we confirmed our findings with DsiRNA approach showing significant improvement in the ability of NPCs to attach, grow, and proliferate when exposed to CSPGs (Fig. 4A–4D). All values are provided in Supporting Information Table 6.
RPTPσ KO neural precursor cells (NPCs) are less susceptible to CSPGs inhibitory effects. NPCs harvested from wild-type and RPTPσ−/− mice were grown on laminin (10 μg/ml) or laminin (10 μg/ml) + CSPG (5 μg/ml) substrate in NPCs growth media for 2 days or differentiated in 2% serum media for 7 days. Lack of RPTPσ in NPCs showed less susceptibility to CSPGs inhibitory effects on cell spreading (A) and attachment (B) in comparison to the wild-type control condition confirming our Dicer Substrate interference RNA findings. Similarly, RPTPσ−/− NPCs were less susceptible to CSPGs inhibitory effect on survival (C) as was shown through an MTT assay 2 days following cell plating. (D): RPTPσ−/− NPCs showed a significantly higher percentage of BrdU positive cells in comparison to wild-type NPCs when grown on laminin+CSPGs substrate. (E–N): Using quantitative immunostaining on NPCs, we measured the percent of GFAP+ astrocytes and Olig2+ oligodendrocytes to the total number of DAPI+ cells under each condition. Wild-type NPCs grown on laminin+CSPGs substrate were preferentially differentiated into astrocytes compared to the laminin control while RPTPσ−/− NPCs showed lesser degree of astrocyte differentiation compared to their respective control condition. Conversely, RPTPσ−/− showed a higher degree of oligodendrocyte differentiation compared to their wild-type counterparts when exposed to CSPGs demonstrating a role for RPTPσ in mediating CSPGs inhibitory effects on oligodendrocyte differentiation. N = 4 independent experiments. The data show the mean ± SEM. *, p < .05, Student's t test. Abbreviations: BrdU, bromodeoxyuridine; CSPG, chondroitin sulfate proteoglycan; DAPI, 4',6-diamidino-2-phenylindole; GFAP, glial fibrillary acidic protein; KO, knockout; RPTPσ, receptor protein tyrosine phosphate sigma.
By signaling on LAR and RPTPσ, CSPGs regulate the differentiation of NPCs (Fig. 3O–3P; Supporting Information Fig. 4I, 4J). Our complementary immunohistochemical and Western blotting using lineage markers for astrocytes (GFAP) and oligodendrocytes (Olig2 and CNPase) revealed that downregulation of both LAR and RPTPσ significantly increased the potential of NPCs for oligodendrocyte differentiation while decreasing astrocyte-derived NPCs when plated on CSPGs in comparison to NPCs treated with control NC1 DsiRNA. The modulatory role of CSPGs/RPTPσ signaling on NPCs differentiation was further confirmed using NPCs from RPTPσ−/− mice in similar experiments (Fig. 4E–4N). These results indicate that both CSPG/LAR and CSPG/RPTPσ signaling influence NPCs differentiation. Notably, our experiments demonstrated that downregulation of LAR and/or RPTPσ had no apparent effect on the ability of NPCs to attach, grow, and proliferate on laminin indicating the specific role of LAR and RPTPσ in mediating CSPGs functions.
CSPGs Effects on NPCs Are Mediated Through the Rho/ROCK Pathway
We next focused on identifying the mechanisms by which CSPGs signaling pathway modulate the properties of NPCs. CSPG activation of the Rho/ROCK pathway has been shown to induce growth cone collapse thereby limiting axonal growth [33-35]. Thus, we asked whether the Rho/ROCK pathway is also involved in exerting the effects of CSPGs on NPCs. Using Western blotting, we observed more than twofold increase in RhoA protein expression in NPCs after exposure to CSPGs (Fig. 5A–5E). Specificity of CSPGs effects on RhoA expression was confirmed by pretreatment of CSPG with ChABC prior to NPC plating. To investigate the functional impact of CSPGs-induced upregulation of RhoA in NPCs, we blocked Rho signaling by inhibition of the downstream ROCK in NPCs with a well-known specific ROCK inhibitor, Y-27632 [34]. We treated dissociated adult spinal cord NPCs with 10 μM of Y-27632 for 1 hour prior to NPC plating. Y-27632 has been commonly used and shown to efficiently block the ROCK pathway in vitro at 10 μM concentration [35-37]. ROCK inhibition was able to entirely overcome the inhibitory effects of CSPGs on NPCs growth and attachment (Fig. 5F, 5G; Supporting Information Fig. 5A–5D). Similarly, blockade of the Rho/ROCK pathway reversed CSPGs inhibition of NPC survival and proliferation. Our MTT, LIVE/DEAD, BrdU, and Ki67 assays collectively confirmed that pretreatment of NPCs with Y-27632 significantly removed CSPGs inhibitory properties on NPC survival and proliferation restoring it to levels near that of our control laminin group at 2 days following cell plating (Fig. 5H, 5I, data not shown for Ki67, Supporting Information Fig. 5G).
Activation of the Rho/ROCK pathway mediates CSPGs inhibitory effects on spinal cord neural precursor cells (NPCs). NPCs were grown on laminin (10 μg/ml) or laminin (10 μg/ml) + CSPG (5 μg/ml) substrate and grown in growth media for 2 days or differentiated in 2% serum media for 7 days. (A–D): Immunostaining of NPCs revealed upregulation in the expression of RhoA in NPCs when grown on laminin+CSPGs substrate (B) in comparison to laminin alone. (E): A significant increase in the expression of RhoA protein was observed in NPCs grown on laminin+CSPGs compared to all other experimental conditions 7 days following cell plating. Involvement of the Rho/ROCK pathway in CSPGs negative effects on NPC growth, attachment, and survival was demonstrated using Y-27632 (10 μM), a ROCK inhibitor. (F, G): CSPGs inhibitory effects on cell spreading (F), attachment (G), and survival (H) was significantly overcome by Y-27632 treatment. (I): Y-27632 treatment overcame CSPGs inhibitory effect on NPC proliferation. (J, K): Immunostaining of NPCs was completed measuring the percent of glial fibrillary acidic protein positive and Olig2 positive cells to the total number of DAPI positive cells. Y-27632 pretreatment of NPCs reversed CSPGs effect on oligodendrocyte differentiation to a level closer to laminin condition with. (K): NPCs grown on laminin + Y-27632 alone slightly promoted the oligodendrocyte differentiation of NPCs however these results were not significant (p = .204 in comparison to laminin only). N = 4 independent experiments. The data show the mean ± SEM. *, p < .05, one-way ANOVA. Abbreviations: ChABC, chondroitinase ABC, CSPG, chondroitin sulfate proteoglycan; DAPI, 4',6-diamidino-2-phenylindole;.
Additionally, our findings revealed that interruption of the Rho/ROCK signaling in NPCs augmented their capacity for oligodendrocyte differentiation in the presence of inhibitory CSPGs (Fig. 5J, 5K, Supporting Information Fig. 5E, 5F). We observed a significant increase in the percentage of GFAP positive astrocytes and instead a marked decrease in Olig2 positive oligodendrocytes in our laminin+CSPGs experimental group which was entirely reversed by inhibition of the Rho/ROCK pathway. These results were complimented with Western blotting where Y-27632 treatment restored GFAP and CNPase protein expression in NPCs grown on CSPGs to levels near that of laminin control group. Notably, Y-27632 treatment also promoted oligodendrocyte differentiation of NPCs on a laminin substrate compared to laminin only control condition, however, this increase was not statistically significant. Altogether, these results indicate that the Rho/ROCK pathway plays a central role in mediating the modulatory effects of CSPGs on several aspects of NPCs activities. All values are provided in Supporting Information Table 7.
Erk and Akt Phosphorylation Plays a Key Role in CSPGs Effects on Spinal Cord NPCs
Recent studies show that interactions between RhoA and Akt pathways regulate the inhibitory effects of CSPGs on neurite outgrowth [24]. Here, we examined the same possibility in NPCs. Western blot analyses showed that exposure to CSPGs significantly decreased Akt phosphorylation in NPCs. There was a 49% reduction in the ratio of phosphorylated Akt to total Akt in NPCs plated on laminin+CSPGs substrate compared to laminin only substrate (Fig. 6A), which was reversed by ChABC pretreatment. We further demonstrated that CSPGs induce Akt dephosphorylation through both LAR and PTPσ receptors (Fig. 6B). We observed a 45% decrease in Akt phosphorylation in NPCs treated with control NC1 DsiRNA when plated on CSPGs whereas NPCs with downregulated LAR and RPTPσ only showed 8% and 17% decrease in their pAKT/tAKT ratio, respectively. Furthermore, we investigated whether the Rho/ROCK pathway mediated the CSPGs inhibitory effects on Akt phosphorylation (Fig. 6C). As shown in previous experiments, there was a significant 40% reduction in pAkt/tAkt ratio when NPCs were grown onto laminin+CSPGs, which was entirely overcome in NPCs pretreated with ROCK blocker Y-27632. Interestingly, ROCK inhibition also increased Akt phosphorylation in NPCs plated on laminin (134.92%) to levels higher than that of laminin alone (100%) and laminin+CSPGs+Y-27632 (118.16%) conditions indicating that suppressing Rho/ROCK pathway generally increases Akt activation in NPCs.
CSPGs limit the phosphorylation of Akt and Erk1/2 by signaling through LAR and RPTPσ and activation of the Rho/ROCK pathway. NPCs were grown on laminin (10 μg/ml) or laminin (10 μg/ml) + CSPG (5 μg/ml) substrate and differentiated in 2% serum media for 7 days. (A): Western blotting analysis showed a significant decrease in the ratio of pAkt to tAKT in NPCs exposed to laminin+CSPGs compared to laminin condition. (B): In control NC1 DsiRNA condition, NPCs had a significant decrease in the ratio of pAkt/tAkt when grown on laminin+CSPGs substrate which was significantly overcome through downregulation of LAR and RPTPσ. (C): Inhibition of the Rho/ROCK (Y-27632, 10 μM) pathway attenuated CSPGs effects on the phosphorylation state of Akt in NPCs grown on laminin+CSPGs substrate. Interestingly, the ratio of pAkt/tAkt was also significantly higher in laminin + Y-27632 condition. (D): Western blotting analysis showed decrease in the phosphorylation of Erk1/2 protein in NPCs grown on CSPGs compared to all control condition, although it did not reach a significant level (p = .135). (E): Downregulation of LAR and RPTPσ partially removed the inhibitory effects of CSPGs on the phosphorylated state of Erk1/2; however, this attenuation was only statistically significant in RPTPσ DsiRNA NPCs in comparison to control condition, DsiRNA NC1 NPCs when grown on laminin+CSPGs substrate. (F): Inhibition of the Rho/ROCK pathway attenuated CSPGs effects on Erk1/2 phosphorylation in NPCs grown on laminin+CSPGs substrate. For all figures, N = 4 independent experiments. The data show the mean ± SEM. (A, C, D, F) *, p < .05, one-way ANOVA; (B, E) *, p <.05, Student's t test. Abbreviations: ChABC, chondroitinase ABC; CSPG, chondroitin sulfate proteoglycan; LAR, leukocyte common antigen-related phosphatase; p-Akt, phosphorylated Akt; p-Erk1/2, phosphorylated Erk1/2; RPTPσ, receptor protein tyrosinase phosphate sigma; siRNA, dicer substrate interference RNA; t-Akt, total Akt; t-Erk1/2, total Erk1/2.
We further explored CSPGs effects on the phosphorylated state of Erk1/2 (Fig. 6D–6F). Reduced phosphorylation of Erk1/2 has been previously shown to correlate strongly with a decrease in oligodendrocytes differentiation and myelin thickness [38-40]. Based on this evidence, we asked whether CSPGs inhibit oligodendrocyte differentiation in NPCs by suppressing Erk1/2 phosphorylation. We observed that the ratio of pErk1/2 to tErk1/2 was significantly decreased in NPCs grown on laminin+CSPGs substrate in comparison to laminin condition (Fig. 6D). Our DsiRNA studies showed that CSPGs cause Erk1/2 dephosphorylation primarily by signaling on RPTPσ (Fig. 6E). We observed a 35% decrease in the ratio of phosphorylated Erk1/2 to total Erk1/2 in control NC1-treated NPCs grown on laminin+CSPGs substrate. Downregulation of RPTPσ in NPCs attenuated Erk1/2 dephosphorylation to 13% compared to laminin substrate while LAR downregulation had no effects. These data indicate that signaling through RPTPσ but not LAR mediates the effects of CSPGs on the phosphorylation of Erk1/2 in NPCs. Lastly, we asked whether the Rho/ROCK pathways modulate the phosphorylation status of Erk1/2 in NPCs in response to CSPGs (Fig. 6F). To this end, our data demonstrated that Y-27632 treatment restored Erk1/2 phosphorylation in NPCs on CSPGs in laminin+CSPGs condition to levels comparable to NPCs plated on laminin or laminin+ Y-27632 suggesting that activation of Rho/ROCK may be linked to dephosphorylation of Erk1/2 in NPCs after exposure to CSPGs.
Discussion
In this study, we report that CSPGs directly and negatively regulate the behavior of spinal cord NPCs in a receptor-mediated manner. Using in vitro assays mimicking the ECM composition of injury, we demonstrate that exposure to CSPGs inhibits several cellular aspects of NPCs including their growth, integration, survival, proliferation, and oligodendrocyte differentiation. CSPGs exert these effects by signaling on both LAR and RPTPσ as well as activation of the Rho/ROCK pathway. At the intracellular level, activation of CSPGs signaling decreases the phosphorylated state of Akt and Erk1/2 in NPCs, which appear to be downstream effectors of CSPGs effects on spinal cord NPCs. Our findings, for the first time, introduce the cellular mechanisms by which CSPGs may modulate the regenerative response of NPCs within the microenvironment of injury and identifies potential targeted interventions for efficient optimization of NPCs-based repair strategies for CNS injuries.
Application of NPCs holds a great promise for SCI repair [3, 4, 41-46]. However, recent studies indicate that the survival and multipotential capacity of CNS-derived NPCs for cell replacement is restricted in the post-SCI milieu [3-6, 44]. Our previous work in chronic SCI demonstrated a strong correlation between the injury-induced upregulation of CSPGs and the suboptimal proliferation and oligodendrocyte differentiation of transplanted and resident NPCs in spinal cord lesion [4, 8]. Recently, CSPGs are also shown to inhibit the process outgrowth of oligodendrocyte precursor cells and myelination in vitro [36, 47] and in demyelinating lesions [47, 48]. Replacement of oligodendrocytes is an essential approach for remyelination and functional repair following SCI [3, 4]. Here, we used a direct in vitro system to model the ECM of SCI where laminin and CSPGs are highly elevated [27, 28]. When exposed to a matrix containing CSPGs, adult NPCs showed a significant decrease in their ability to integrate, grow, proliferate, and survive. Of note, recent evidence indicates that proper activation of the resident NPC population is required to limit the extent of lesion following SCI and release neurotrophic factors which improve neuronal survival and attenuate tissue degeneration [49].
We also found that CSPGs restricted the potential of NPCs to differentiate into oligodendrocytes and instead favored generation of new astrocytes. Notably, our data suggest that the inhibitory effect of CSPGs on oligodendrocyte differentiation is an active process and not due to initial selective attachment of astrocytes to the substrate resulting in a decrease in oligodendrocyte differentiation. This possibility was eliminated by plating undifferentiated NPCs and allowing them to attach to the substrate prior to serum differentiation. Our findings also explain our in vivo observation where we showed a strong correlation between upregulation of CSPGs and increased differentiation of resident or transplanted NPCs towards astrocytic fate following SCI [5-8].
The recent discovery of CSPGs receptors, RPTPσ [23, 32], LAR [24], as well as Nogo receptor (NgR) family members, NgR1 and NgR3 [50] allows new understandings of CSPGs mechanism in CNS regeneration. LAR and RPTPσ are widely expressed in adult CNS neurons and are shown to mediate the inhibitory effects of CSPGs on axonal regeneration [23, 24]. Lack of RPTPσ promoted axon growth on CSPGs substrates in vitro [23]. RPTPσ knockout mice showed improved regeneration of the corticospinal tract following SCI [23, 32] and enhanced axonal regeneration after optic nerve and peripheral nerve injuries [51-53]. Blocking LAR also resulted in increased axonal growth of serotonergic fibers around and beyond the SCI lesion accompanied by functional recovery [24]. Here, our new evidence extends these concepts by revealing that spinal cord NPCs express both LAR and RPTPσ in vitro and in vivo and thereby their activities can be influenced directly by upregulation of CSPGs. Using DsiRNA gene silencing as well as RPTPσ knockout studies, we demonstrate that downregulation of LAR or RPTPσ partially reversed CSPGs inhibitory effects on NPCs with additive effects after their coinhibition. Deletion of RPTPσ per se did not alter the potential of NPCs for self-renewal and differentiation which is in agreement with an earlier study [54]. Interestingly, embryonic derived neuronal restricted precursor cells (NRPs) are intrinsically insensitive to CSPGs, and this was correlated with reduced expression of both RPTPσ and LAR in growth cones of NRP derived neurons [55].
We found that knockdown of LAR and/or RPTPσ did not entirely reverse CSPGs effects on NPCs. This could be due to the transient DsiRNA-mediated knockdown of the LAR and RPTPσ, the compensatory action of LAR and RPTPσ and/or involvement of LAR and RPTPσ independent mechanisms of CSPGs in NPCs. CSPGs are additionally shown to inhibit axon growth by signaling on Nogo receptors, NgR1 and NgR3 [50] and by blocking laminin/integrin signaling [56, 57]. Importantly, aggrecan directly inhibits laminin-mediated axon growth by impairing integrin signaling through decreasing phosphorylated FAK and Src levels [58, 59]. Notably, aggrecan was included in our CSPGs mixture and our substrate also contained laminin to more closely represent the ECM composition of SCI [27]. Thereby, it is plausible that blockade of integrin/laminin signaling in NPCs partly underlined the inhibitory effects of CSPGs independent of LAR and/or RPTPσ. Our findings also indicate that knockdown of LAR and/or RPTPσ did not alter the behavior of spinal cord NPCs on laminin only substrate suggesting that these receptors are not involved in laminin signaling in NPCs. Collectively, our data demonstrates that the inhibitory role of LAR and RPTPσ in NPCs regulation is CSPGs specific.
At the intracellular level, we demonstrate a pronounced upregulation of RhoA protein expression in NPCs after exposure to CSPGs. Inhibition of the downstream ROCK was sufficient to reverse nearly the entire of CSPGs effects on NPCs. Notably, elevated levels of active RhoA are detected after axonal injury [60, 61] resulting in growth cone collapse [33-35, 62]. Moreover, Rho activation has been associated with p75-mediated apoptosis in neurons and glial cells following SCI [60], which may be the underlying mechanism of the restored survival of NPCs on CSPGs following blockade of Rho/ROCK pathway in our experiments. CSPGs-LAR interaction also triggers the activation of RhoA in neurons [24]. Interestingly, CSPGs induced inhibition of oligodendrocyte myelination was also overcome by blocking ROCK and downregulation of RPTPσ [36, 63]. Importantly, both CSPGs and myelin associated inhibitors converge on RhoA pathway and inhibit axonal growth [64]. Ba-210, a Rho inhibitor trademarked as Cethrin, is currently being assessed in clinical trials for SCI [65, 66], and our findings suggest that such strategy could be a potential combinatorial approach for promoting NPCs cell replacement activities following SCI.
Reduced phosphorylation state of both Akt and Erk1/2 in spinal cord NPCs through interaction with LAR and RPTPσ and the Rho/ROCK pathway (Fig. 7) may underlie the CSPG-mediated decrease in NPCs growth, adhesion, proliferation, and differentiation in our studies. Activation of PI3K/Akt and/or MAPK/Erk pathways has been previously linked to induced NPCs proliferation [67] and enhanced oligodendrocytes survival and differentiation in vitro and in vivo [68-70]. Downstream effector of the PI3K/Akt, mTOR, is critical for oligodendrocyte differentiation [71] and myelination [72, 73]. MAPK/Erk signaling also plays central role in oligodendrocyte process extension and myelination [38, 72, 74, 75] and our knockdown studies identified that CSPGs/RPTPσ signaling appears to be the key mediator of Erk1/2 dephosphorylation in NPCs (Fig. 6E). Altogether, our findings provide novel evidence that CSPGs signaling negatively regulates activation and oligodendrocyte differentiation of spinal cord NPCs likely by modulating their Akt and Erk1/2 activities.
Schematic illustration of proposed mechanisms of CSPGs in regulating spinal cord NPCs. CSPGs modulate their inhibitory effects on adult spinal cord NPCs through multiple mechanisms. CSPGs effects are directed through LAR and RPTPσ and mediated through the Rho/ROCK pathway and inhibition of the phosphorylation state of Akt and Erk1/2. Abbreviations: CSPG, chondroitin sulfate proteoglycan; LAR, leukocyte common antigen-related phosphatase; NPC, neural precursor cells; RPTPσ, receptor protein tyrosinase phosphate sigma.
Conclusion
In conclusion, this study provides new direct insights into the mechanisms by which CSPGs restrict the regenerative activities of adult spinal cord NPCs. Upregulation of CSPGs is an inevitable outcome of CNS injuries with major impact on repair and regeneration. Here, we demonstrate that CPSGs modulate several properties of adult spinal cord NPCs through direct receptor-mediated mechanisms. Our findings suggest that blockage of LAR and RPTPσ signaling and/or the Rho/ROCK pathway in NPCs are plausible interventions for overcoming the inhibitory effects of CSPGs on cell replacement strategies. Knowledge gained in this study will aid in optimizing current cellular therapies for SCI and other CNS injuries characterized by the upregulation of CSPGs.
Acknowledgments
This work was supported by grants from the Natural Sciences and Engineering Council of Canada (NSERC), the Manitoba Health Research Council (MHRC), the Health Sciences Center Foundation, and the Manitoba Medical Services Foundation to S.K.A. S.D. and A.A. were supported by studentships from MHRC. RPTPσ−/NRP mouse was obtained from Dr. Michel Tremblay (McGill University).
Author Contributions
S.M.D.: concept and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; A.A., T.K.S., E.H.P., and C.-L.W.: collection and/or assembly of data; S.K.-A.: conception and design, financial support, administrative support, provision of study material or patients, data analysis and/or assembly of data, manuscript writing, and final approval of manuscript.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
