Satellite glia not DRG neurons constitutively activate EGFR but EGFR inactivation is not correlated with axon regeneration

Neurobiology of Disease 39 (2010) 292–300

Contents lists available at ScienceDirect

Neurobiology of Disease j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y n b d i

Satellite glia not DRG neurons constitutively activate EGFR but EGFR inactivation is not correlated with axon regeneration Zubair Ahmed ⁎, Martin L. Read, Martin Berry, Ann Logan Molecular Neuroscience Group, Neuropharmacology and Neurobiology Section, School of Clinical and Experimental Medicine, University of Birmingham, Birmingham, B15 2TT, UK

a r t i c l e

i n f o

Article history: Received 10 December 2009 Revised 12 April 2010 Accepted 26 April 2010 Available online 6 May 2010 Keywords: EGFR DRGN Axon/neurite growth siRNA DRG satellite glia

a b s t r a c t To test the possibility that phosphorylated epidermal growth factor receptor (pEGFR) mediates axon growth inhibition, we determined if pEGFR levels were raised in dorsal root ganglia (DRG) after non-regenerating dorsal column (DC) lesions and suppressed in regenerating sciatic nerve (SN) and preconditioning (P) SN + DC lesioned DRG. Levels of EGFR mRNA and protein in DRG were unchanged between control and all injury models. Satellite glia and not DRG neurons (DRGN) constitutively contained pEGFR and, only in PSN + DC rats, were levels significantly reduced in these cells. In vitro, siRNA mediated knockdown of EGFR (siEGFR) mRNA and protein was associated with suppressed RhoA activation, but fibroblast growth factor-2 (FGF2) was a mandatory requirement for DRGN neuritogenesis after addition of inhibitory concentrations of CNS myelin. Thus, EGFR activation in satellite glia was not consistently correlated with DRGN axogenesis and siEGFR reduction of pEGFR with attenuated Rho-GTP signalling did not promote DRGN disinhibited neurite outgrowth without exogenous FGF2 stimulation. Together, these data argue against a direct intra-axonal involvement of pEGFR in axon regeneration. © 2010 Elsevier Inc. All rights reserved.

Introduction After dorsal column (DC) lesions, the growth of all centrally projecting dorsal root ganglion neuron (DRGN) axons is arrested at the injury site by both CNS myelin-/scar-derived axon growth inhibitory ligands (e.g. Nogo-A, myelin associated glycoprotein (MAG), chondroitin sulphate proteoglycan (CSPG), ephrins and semaphorins) and a limited availability of neurotrophic factors required to maintain DRGN survival and promote axon regeneration (Hunt et al., 2002; Sandvig et al., 2004; Fournier et al., 2002; McKerracher, 2001; McKerracher and Winton, 2002; Hou et al., 2008; Fabes et al., 2007; Du et al., 2007; Berry et al., 2008). However, a preconditioning (P) lesion to the sciatic nerve (SN), 1–2 w before DC transection, promotes DRGN axonal regeneration (Neumann and Woolf, 1999; Chong et al., 1999). The in vitro growth potential of DRGN neurites is also increased after PSN lesions, suggesting that conditioning enhances the intrinsic growth of axotomised DRGN and leads to limited axon regeneration through the non-permissive DC neuropil (Neumann and Woolf, 1999). The epidermal growth factor receptor (EGFR) has been implicated in mediating disinhibited CNS axon regeneration using the specific reversible and irreversible EGFR kinase antagonists AG1478 and ⁎ Corresponding author. Molecular Neuroscience Group, Neuropharmacology and Neurobiology Section, School of Clinical and Experimental Medicine, College of Medical and Dental Sciences, University of Birmingham, Room WX2.17, Institute of Biomedical Research (West), Edgbaston, Birmingham B15 2TT, UK. Fax: + 44 121 414 8867. E-mail address: [email protected] (Z. Ahmed). Available online on ScienceDirect (www.sciencedirect.com). 0969-9961/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.nbd.2010.04.013

PD168393, respectively (Koprivica et al., 2005; Erschbamer et al., 2007). For example, in the injured optic nerve, PD168393 promotes retinal ganglion cell (RGC) axon regeneration (Koprivica et al., 2005) and rescues RGC from death in a chronic glaucoma model (Liu et al., 2006). In the latter study, EGFR expression was localised to astrocytes, not to RGC or their axons, suggesting that the beneficial effects of EGFR inhibition may be mediated by glia and not neurons. EGFR activation triggers quiescent astrocytes into a reactive phenotype (Liu et al., 2006) that secrete CSPG (Smith and Strunz, 2005) and form a cribriform lattice with their processes that may contribute to the formation of glial scars formed after penetrant injury (Liu and Neufeld, 2004). EGFR is not expressed in mature astrocytes (Gomezpinilla et al., 1988) except after ischemia (Planas et al., 1998; Jin et al., 2002), spinal cord tractotomy (Lisovoski et al., 1997), and in the optic nerve of patients with glaucoma (Liu and Neufeld, 2003; Liu et al., 2006). After spinal cord injury, inhibition of EGFR improves motor and sensory function as well as bladder emptying (Erschbamer et al., 2007). In adult primary DRG and retinal cultures with added CNS myelin extracts (CME), DRGN and RGC neurite outgrowth is induced through the off-target release of neurotrophins from both glia and neurons and raised cAMP levels after AG1478 treatment (Ahmed et al., 2009a; Douglas et al., 2009). Thus, these latter studies assert that EGFR does not directly mediate the inhibition of CNS axon regeneration. In the present study, we investigated the in vivo role of EGFR in adult rat DRGN axon regeneration in a non-regenerating transected DC model, a regenerating SN lesion model, and in a paradigm in which DC axons were stimulated to regenerate after PSN lesions, predicting that neuronal pEGFR levels would be raised in the non-regenerating,

Z. Ahmed et al. / Neurobiology of Disease 39 (2010) 292–300

but reduced in the regenerating models if an intra-axonal role is valid. In primary DRG cultures, we used an siRNA to EGFR (siEGFR) to reduce EGFR titres, predicting that siEGFR knockdown would promote disinhibited DRGN neurite outgrowth in the presence of CME, similar to that induced by AG1478 and PD168393 (Koprivica et al., 2005). In agreement with our previous findings (Ahmed et al., 2009a; Douglas et al., 2009), the results of the current experiments provide no support for direct intra-axonal negative regulation of neurite/axon growth through the EGFR signalling axis. Materials and methods In vivo experimental design The experiments comprised 4 groups of animals each containing 10 adult male Sprague–Dawley rats (150–200 g) (Charles River, Margate, UK) for each analytical end-point (n = 10 DRG), designated as: (1), uninjured controls; (2) DC crush (non-regenerating DC lesions); (3), SN crush (regenerating SN lesions); and (4), PSN lesion 1 w before a DC crush (regenerating PSN + DC lesions). Surgical procedures All surgical procedures were licensed by the UK Home Office, approved by the University of Birmingham Ethical Committee and conducted under inhalation anaesthesia induced with 5% isofluorane with 1.5 L/min O2 and maintained with reducing levels of isofluorane. The DC was crushed bilaterally at the level of T8 using calibrated watchmaker's forceps inserted through the dorsal cord meninges to a depth of 1.5 mm (Lagord et al., 2002). The left SN was exposed at a mid-thigh level and crushed using suture forceps at the level of the sacrotuberous ligament. A preconditioning lesion of the SN was performed 1 w before DC crush, as described above. Animals were killed by CO2 exposure 10 d after surgery and the L4/L5 DRG was harvested for RNA and protein analysis by snap freezing in liquid nitrogen. The contralateral L4/L5 DRG served as uninjured controls. For immunohistochemistry, animals were intracardially perfused with 4% formaldehyde and processed as described later. Microarray analysis The rat genome AROS™ V3.0 set (Operon Biotechnologies GmbH, Cologne, Germany) contained 26,962 long mer probes representing 22,012 genes and 27,044 gene transcripts. Slide preparation was performed by the Functional Genomics Laboratory (University Of Birmingham, UK). Oligonucleotides were resuspended in Pronto Universal Slide Spotting Solution (Fisher Scientific, Loughborough, UK) and subsequently spotted onto UltraGAPS Coated Slides with Bar Code (Fisher Scientific) using a BioRobotics Microgrid II spotter (Genomic Solutions Ltd, Huntingdon, UK). Extracted total RNA (0.5–1 μg) was amplified using the Amino Allyl Message Amp II aRNA kit (Ambion, Austin, TX, USA) and then labelled with either Cy3, or Cy5 dyes (GE Healthcare, Little Chalfont, UK) according to the manufacturer's protocol. The frequency of label incorporation was calculated using the dye incorporation calculator (www.ambion.com/ tools/dye), and frequencies of 30–60 dye molecules/1000 nucleotides were considered a successful coupling reaction. Probes were resuspended in long oligo Pronto hybridisation buffer (Corning, Lowell, MA, USA) with mouse Hybloc solution (Insight Biotechnology, London, UK) before hybridization. Microarray hybridisation used a Pronto hybridisation kit (Corning) according to the manufacturer's protocol. The slides were scanned with an Axon GenePix® 4000B scanner (Molecular Devices Ltd, Berkshire, UK) set at 600 V and images were analysed using GenePix® V5.0 software (Molecular Devices). Each experiment was replicated at least ×4. Normalisation of the microarray data and subsequent analyses were performed using Gene-

293

Spring GX7 software (Agilent Technologies, Cheshire, UK). In this study, we selected neuropeptide Y (npy), galanin (gal), activating transcription factor 3 (atf3) and growth associated protein 43 (gap43) as our markers for a regenerative phenotype in DRGN (Costigan et al., 2002) and egfr mRNA for further analysis. Real-time RT-PCR Ipsilateral and contralateral to the crushed SN, L4/L5 DRG were dissected and total RNA was extracted using the RNeasy Mini Kit (Qiagen, Crawley, UK) according to the manufacturer's protocol. Purified RNA was reverse-transcribed into cDNA using the Reverse Transcription System (Promega, Southampton, UK) according to the manufacturer's protocol. PCR reactions were set up in ABI PRIZM™ 96 well optical reaction plates using 40 ng of the cDNA template, 2× Universal PCR Mastermix and either 18S endogenous control probe, or the relevant Taqman Gene Expression Assay (all from Applied Biosystems, Foster City, CA, USA), and amplified on an ABI PRIZM® 7700 (Applied Biosystems) set at 95 °C for 10 min, followed by 40 cycles at 95 °C for 15 s and then at 60 °C for 1 min. PCR data were collected and the relative expression of target genes was calculated using the 2ΔΔCt method (DataAssist™ Software, Applied Biosystems). Taqman Gene Expression Assays (Applied Biosystems) were used to measure relative mRNA levels of the axon regeneration related genes npy, gal, atf3, gap43 (Costigan et al., 2002) and compared with egfr expression. Antibodies Polyclonal rabbit anti-human EGFR and polyclonal goat anti-human pEGFR (Santa Cruz Biotechnology, CA, USA) were used to localise total EGFR and phosphorylated EGFR in immunohistochemistry/immunocytochemistry both at 1:200 dilution. Blocking peptides for EGFR and pEGFR (Santa Cruz Biotechnology, CA, USA) were used at 1:20 concentration to confirm antibody specificities. Monoclonal β-III tubulin antibody (Sigma, Poole, UK) labelled DRGN neurites at 1:200 by immunocytochemistry. For western blotting both EGFR and pEGFR were used at 1:500 dilution, while total Rho was detected using a monoclonal anti-Rho antibody diluted at 1:200 (Upstate Biotechnology, Milton Keynes, UK) and β-actin (Sigma) at 1:10,000 dilution as a loading control for western blots. Adult DRGN cultures L4-L7 DRG pairs were dissected from 6 to 8 w-old Sprague–Dawley rats and dissociated into single cells using a solution of Neurobasal-A (Invitrogen, Paisley, UK) containing 0.1% collagenase (Sigma) and 200 U/ml DNaseI (Worthington Biochem, New Jersey, USA) as previously described by us (Ahmed et al., 2009a). DRGN were cultured at 500 cells/chamber on sterile glass chamber slides (BD Biosciences, Oxford, UK) pre-coated with 100 µg/ml poly-D-lysine followed by 20 µg/ml Laminin-I (Sigma), in supplemented Neurobasal-A medium, in both the presence and absence of the same batch of rat CNS myelin known to contain Nogo-A, OMgp, MAG and CSPG (Ahmed et al., 2005), for 48 h at 37 °C in a humidified atmosphere containing 5% CO2. Cultures were also treated with 30 µM forskolin (Sigma) to raise cAMP levels and fibroblast growth factor-2 (FGF2) (Peprotech, London, UK) to promote DRGN neurite outgrowth (Ahmed et al., 2005) where appropriate. Immunohistochemistry Animals were killed 10 d post injury (dpi) and intracardially perfused with 4% formaldehyde (TAAB Laboratories, Berkshire, UK). L4/L5 DRG were then post-fixed in 4% formaldehyde, cryoprotected through a graded series of sucrose solutions and blocked up in OCT mounting compound (TAAB Laboratories). Sections of DRG were cut

294

Z. Ahmed et al. / Neurobiology of Disease 39 (2010) 292–300

10 µm thick using a cryostat and adhered onto charged glass slides, fixed in 100% ethanol for 1 min before washing ×3 in phosphate buffered saline (PBS), permeabilised in PBS containing 1% Triton X-100 (Sigma) for 10 min and blocked in PBS containing 0.5% bovine serum albumin (BSA) (Sigma) and 0.05% Tween 20 (PBS-T-BSA) for 30 min at room temperature. Sections were then incubated with the relevant primary antibody diluted appropriately in PBS-T-BSA in a humidified chamber overnight (16–18 h) at 4 °C. Sections were then washed in ×3 in PBS and incubated with the relevant secondary antibody diluted in PBS-T-BSA for 1 h at room temperature in a humidified chamber. For fluorescent detection of the antigen, secondary antibodies were either coupled to Alexa Fluor 488 (Green), or Texas Red (Red) (both from Molecular Probes, Oregon, USA), diluted at 1:400. Sections were also incubated with biotinylated secondary antibody (diluted 1:400) followed by incubation with DAB substrate (Vector Laboratories). Coverslips were mounted in Vectamount with and without DAPI (Vector Labs, Peterborough, UK) and viewed under a fluorescent/light microscope (Carl Zeiss, WelwynGarden City, UK). Control sections were incubated in each run for each antibody tested either after pre-incubation of antibodies with their relevant blocking peptides (blocking peptide:antigen, 20:1), or with no primary antibody (Douglas et al., 2009; Ahmed et al., 2009a).

cold lysis buffer containing 20 mM Tris–HCl (pH 7.4), 1 mM EDTA, 0.5 mM EGTA, 150 mM NaCl, 1% NP-40 (Sigma) and protease inhibitor cocktail (Sigma), incubated on ice for 30 min, and centrifuged at 13,000 rpm at 4 °C. Lysates were normalised for protein concentration using a colorimetric DC protein assay (Bio-Rad, Hercules, CA, USA). Homogenates and cell lysates were stored at −70 °C until used for western blot analysis. Each 40 µg total protein sample was incubated ×2 in Laemmli loading buffer at 90 °C for 4 min and separated on a 12% SDS-polyacrylamide gel (Invitrogen). Proteins were transferred to PVDF membranes (Millipore UK, Gloucestershire, UK), blocked for 1 h at room temperature in Tris-buffered saline containing 0.1% Tween 20 and 5% non-fat milk. Membranes were blotted overnight for the relevant antibody. For detection, an enhanced chemiluminescence (ECL) system (Amersham, Buckinghamshire, UK) and HRP-conjugated secondary antibody (1:1000, Amersham) were used. Each blot was stripped and re-probed with relevant antibodies thereafter, including β-actin as a protein loading control. To quantify detected bands by densitometry, blots were scanned into Photoshop (Adobe Systems, San Jose, CA, USA) keeping all scanning parameters the same and analysed using the built-in gel plotting macros in ScionImage (Scion Corporation, Maryland, USA).

sEGFR preparation and transfection

Rho activation assay

We used Lipofectamine 2000 reagent (Invitrogen) to transfect DRGN with 50 nM siEGFR (Dharmacon, Colorado, USA)) to knockdown EGFR mRNA: sense 5′-GAAGAGACCUGCAUUAUCAUU-3′, following the manufacturer's instructions. DRGN were also transfected with Lipofectamine alone, a scrambled version of the above siEGFR sequence (Scr-siEGFR, Dharmacon) and a non-specific siRNA to GFP (siGFP) (Dharmacon) as controls. After 5 h of transfection, supplemented Neurobasal-A was added to the transfection medium to bring the final volume to 500 µl/chamber and DRGN were incubated for a further 48 h either in the presence or absence of CME or FGF2 before cell lysis and western blotting.

GTP bound Rho was assayed using a commercially available kit according to the manufacturer's instructions (Millipore, Watford, UK).

Statistical analysis Unless otherwise stated, each variable was performed in triplicate and experiments were repeated on three separate occasions. Sample means were calculated and analysed for statistical significance using GraphPad Prism (GraphPad Software Inc., Version 4.0, San Diego, USA) by one-way analysis of variance (ANOVA) followed by post-hoc testing with Dunnett's method.

Immunocytochemistry DRGN cultures were fixed in 4% formaldehyde for 10 min before washing ×3 in PBS, blocking in PBS-T-BSA, and incubation with the relevant primary antibody diluted 1:200 in PBS containing PBS-T-BSA for 1 h at room temperature in a humidified chamber. Cells were then washed ×3 in PBS and incubated with either Alexa Fluor 488 (Green), or Texas Red (Red), diluted 1:100 in PBS-T-BSA for 1 h at room temperature. After ×3 washes in PBS, coverslips were mounted in FluorSave (Calbiochem, San Diego, USA) and viewed under the fluorescent microscope. Antibody specificity controls were included in each run for each antibody tested, which included no primary and pre-incubation with relevant blocking peptides (Douglas et al., 2009; Ahmed et al., 2009a). Measurement of neurite outgrowth Photomicrographs of βIII-tubulin+ immunostained DRGN neurites were captured using Axiovision Software (Carl Zeiss) from 60 randomly selected DRGN/chamber using Axiovision (Ahmed et al., 2006a) and, using the built-in measurement facilities, the lengths of the longest neurites and the numbers of βIII-tubulin+ DRGN with neurites were recorded and represented as means ± SEM. Protein extraction and western blotting To determine the levels of EGFR, pEGFR, and β-actin in vivo in treated DRG pairs and total Rho in treated DRG cultures, harvested DRG tissue or cells were washed ×2 with PBS and incubated in ice-

Results mRNA levels of axon regeneration related genes in DRG following DC, SN and PSN + DC lesions Microarray results showed that the relative levels of npy, gal, atf3 and gap43 mRNA in DRG all increased significantly after injury in both SN, and PSN + DC regenerating models (Table 1), compared to the non-regenerating DC paradigm. However, the levels of EGFR mRNA remained unchanged in the DRG of all experimental lesion groups suggesting that modulation of EGFR mRNA expression was unrelated to axon regeneration. These changes in individual mRNA levels were corroborated by RT-PCR (Fig. 1) and implied that axon regeneration was unrelated to EGFR transcription.

Table 1 Relative mRNA levels of described genes in rat DRG after either SN, DC or PSN + DC injury as indicated. mRNA npy gal atf3 gap43 egfr

Description Neuropeptide Y Galanin Activating transcription factor 3 Growth associated protein 43 Epidermal growth factor receptor

SNa 36.7 31.2 12.7 3.3 1.0c

DCa

PSN + DCb c

0.85 1.4c 1.1c 0.95c 0.93c

23.6 30.9 12.9 2.7 1.2c

Normalised mean values are shown from 4 samples (p-values b 0.05 unless otherwise indicated). Data normalised to levels of specific genes detected in untreated, intact DRG. a Non-regenerating. b Regenerating. c p-values N 0.05.

Z. Ahmed et al. / Neurobiology of Disease 39 (2010) 292–300

295

correlate with the regeneration of DRGN axons i.e., pEGFR levels were not changed in non-regenerating DC lesioned and regenerating SN lesioned DRGN but were reduced in the regenerating PSN + DC model. Localisation of pEGFR in DRG in uninjured and lesioned animals

Fig. 1. RT-PCR validation of microarray data confirmed up-regulation of npy, gal, atf3 and gap43 mRNA in DRG after non-regenerating DC and regenerating SN and PSN + DC lesions, relative to the DC model and control intact DRG. mRNA levels for egfr did not change in any of the experimental groups whereas all mRNA were unchanged in the DC model (n = 4 animals/group; 3 independent experiments).

pEGFR and RhoA levels are reduced in DRG after regenerating PSN + DC but not after regenerating SN lesions Western blotting (Fig. 2A) and immunohistochemistry (Fig. 2B) demonstrated that EGFR levels did not change from control levels in any of the treated groups and pEGFR titres did not consistently

Immunofluorescent staining of pEGFR in DRG sections from control uninjured (Fig. 3A), non-regenerating DC (Fig. 3B) and both regenerating SN (not shown) and PSN + DC lesioned animals (Fig. 3C) demonstrated that the localisation of pEGFR was primarily restricted to DRG satellite cells; very low levels of constitutive pEGFR immunofluorescence were observed in DRGN somata (Supplementary Fig. 1-low power). Pre-incubation of pEGFR with its blocking peptide completely abrogated any pEGFR+ immunostaining in DRG sections (Fig. 3D). High power magnification of DRG sections taken from non-regenerating DC (Figs. 3E and G) and regenerating PSN + DC lesioned rats (Figs. 3F and H) confirmed the observations that fewer satellite cells were pEGFR+ compared to those in sections of DRG from either unlesioned controls (Fig. 3A), or regenerating SN rats (Fig. 4). Using antibodies to S100 that specifically labels satellite cells, we were able to confirm that the majority of immunostaining for pEGFR was present in satellite cells (Supplementary Fig. 2). These results demonstrate that pEGFR is primarily modulated in satellite cells post-axotomy and that pEGFR levels are unrelated to the regenerative status of DRGN axons (Fig. 4). siEGFR knockdown of mRNA and protein reduced pEGFR, attenuated Rho-GTP levels but did not induce DRGN neurite outgrowth No changes in EGFR protein were detected in controls after treatment with either Lipofectamine alone (not shown), Scr-siEGFR (Fig. 4A), or siGFP (not shown). However, N90% knockdown of total EGFR protein occurred in DRG cultures after 48 h treatment with

Fig. 2. Changes in EGFR and pEGFR protein levels in the DRG of non-regenerating and regenerating paradigms. (A) Western blots showed reduced pEGFR levels only in DRG from the regenerating PSN + DC lesioned model, while total EGFR levels were unaffected. β-actin was used as a loading control. (B) Immuno-peroxidase staining (brown) demonstrated localisation of EGFR and pEGFR in DRG sections from uninjured controls, 10 dpl non-regenerating DC and regenerating SN and PSN + DC lesioned animals (n = 4 animals/group, 3 independent experiments; scale bars in B = 50 µm).

296

Z. Ahmed et al. / Neurobiology of Disease 39 (2010) 292–300

Fig. 3. Localisation of pEGFR in DRG from non-regenerating DC and regenerating PSN+ DC lesioned rats by immunofluorescence at 10 dpl. Low levels of pEGFR immunoreactivity were present in satellite glia in uninjured control DRG sections (A), while most satellite glia were immuno-positive in the DRG of non-regenerating DC lesioned rats (B) (green = pEGFR; blue = DAPI). Similar immunostaining was seen in DRG in the regenerating SN model (not shown). However, fewer satellite glia were immuno-positive in the DRG of regenerating PSN + DC lesioned rats (C). Pre-incubation of pEGFR with its blocking peptide completely abrogates pEGFR+ immunoreactivity (D). High power images confirmed satellite glial localisation of pEGFR in DRG of non-regenerating DC (E, G) and regenerating PSN + DC paradigms (F, H) (pEGFR (green) and DAPI (blue); scale bars in A–F = 10 µm; G and H = 5 µm).

siEGFR, together with N95% attenuation of pEGFR levels (Figs. 5A–C). Moreover, EGFR knockdown was associated with 80% reduction in RhoA-GTP compared to Scr-siEGFR (Figs. 5A and D), suggesting that knockdown of EGFR is associated with perturbation of Rho-mediated neurite growth inhibitory signalling, which may facilitate disinhibited neurite outgrowth.

FGF2 and cAMP promote DRGN neurite outgrowth in CME-inhibited DRG cultures after reducing pEGFR and RhoA-GTP protein levels with siEGFR knockdown In CME-inhibited DRG cultures, addition of either FGF2 (Figs. 6A, F–H), or siEGFR failed to stimulate disinhibited DRGN neurite outgrowth in DRG cultures with added CME (Figs. 6B, G and H) but, in CME inhibited DRG cultures treated with both siEGFR and FGF2 significant DRGN neurite outgrowth occurred (Fig. 6C), increasing both the proportion of DRGN with neurites (Fig. 6G) and mean neurite length (Fig. 6H) compared to cultures treated with CME+siEGFR.

Fig. 4. Diagrammatic representation of the localisation of pEGFR (green shading) in DRGN and satellite glia in the DRG in: (i), unlesioned controls; and after: (ii), nonregenerating DC; (iii), regenerating SN; and (iv), regenerating PSN + DC lesioned rats. Constitutive low levels of pEGFR were unchanged in DRGN of any of the lesion paradigms. Constitutive high levels of pEGFR were found in satellite glia of control, nonregenerating DC lesioned and regenerating SN lesioned DRGN, but levels in regenerating PSN + DC lesioned DRGN were reduced.

Raising the levels of cAMP by the addition of forskolin also overcame CME inhibition (Fig. 6D) and promoted similar DRGN neurite outgrowth as seen in DRG cultures treated with FGF2 without CME (Figs. 6G and H). In addition, EGFR knockdown with the addition of forskolin, promoted significantly higher disinhibited DRGN neurite outgrowth (Figs. 6E, G and H) when compared to forskolin + CME (Figs. 6D, G and H). The combination of CME + siEGFR + forskolin + FGF2 promoted longer DRGN neurite outgrowth than either CME + siEGFR + forskolin (P b 0.0001, ANOVA), or CME+ siEGFR+ FGF2 treatments (P b 0.0001, ANOVA) (Figs. 6F and G). Similarly, the combination of siEGFR + forskolin+ FGF2 with CME stimulated the growth of significantly more DRGN neurites than did siEGFR+ forskolin (P b 0.05, ANOVA) and siEGFR+ FGF2 treatments (P b 0.001, ANOVA) (Fig. 6G). Discussion In this study, the expression of regeneration related molecules were up-regulated in L4/5 DRG after both SN and PSN+ DC lesions, but the levels of EGFR mRNA and protein remained unchanged. By contrast, no

Z. Ahmed et al. / Neurobiology of Disease 39 (2010) 292–300

297

Fig. 5. Knockdown of EGFR and subsequent reduction of pEGFR and Rho-GTP in primary DRG cultures. Compared to Scr-siEGFR, siEGFR achieved a N 90% knockdown of EGFR protein (A, B), and reductions of N 95% in pEGFR (A, C) and N80% in Rho-GTP (A, D).

change in any gene was seen after DC lesion. We chose npy, gal, atf3 and gap43 as regeneration related markers since all of these molecules are known to be up-regulated in DRGN regenerating their axons. For example, Npy expression increases in large diameter DRGN and laminae III–V of the spinal cord after SN transection (Wakisaka et al., 1992; Wakisaka et al., 1991; Zhang et al., 1995) and Npy protein indirectly enhances DRGN neurite outgrowth in vitro (White and Mansfield, 1996; White, 1998). Gal is a neuropeptide that is expressed developmentally in DRG and rapidly up-regulated after SN transection, while targeted disruption of the Gal gene reduces both the number of DRGN in mice and their capacity to regenerate axons/neurites both in vivo and in vitro (Hokfelt et al., 1987; Holmes et al., 2000). Atf3 is a transcription factor which is activated in all DRGN after SN injury, and stimulates DRGN neurite outgrowth by potentiating the intrinsic growth capacity of neurons (Seijffers et al., 2007). Finally, GAP43 is targeted to the growing tips of developing and regenerating axons and up-regulated after SN injury, while the growth of neurites of neurons that over-express GAP43 is potentiated (Bomze et al., 2001; Woolf, 2001). Our microarray data corroborate previous findings that these molecules are up-regulated after injury and during axon regeneration demonstrating their usefulness as markers for the DRGN regenerative phenotype (Costigan et al., 2002). CNS axon regeneration requires suppression of myelin and non-myelin derived inhibitory ligand signalling and concomitant mobilisation of axon growth (Fischer et al., 2004a, b; Ahmed et al., 2005; Ahmed et al., 2006a, b). However, the discovery that AG1478/ PD168393-induced EGFR inactivation promotes RGC axon regeneration in vivo challenged this perception (Koprivica et al., 2005). In a recent study, we confirmed the findings of Koprivica et al. (2005) that AG1478/ PD168393 suppressed EGFR phosphorylation and promoted disinhibited DRGN neurite outgrowth, but demonstrated that EGFR antagonists acted off-target to up-regulate intracellular cAMP levels and stimulate

glia/neurons to secrete neurotrophins which were ultimately responsible for neuritogenesis (Ahmed et al., 2009a). We also demonstrated that EGFR knockdown, with the same siRNA molecule used in the current experiments, failed to promote DRGN neurite outgrowth unless AG1478 was also added to the cultures to stimulate off-target neurotrophin release (Ahmed et al., 2009a). In our current study, we predicted that, if EGFR mediated, or was required for the intra-axonal inhibition of axon regeneration, levels of the active isoform, pEGFR, would be suppressed in regenerating DRGN after SN and PSN + DC lesions and raised in the non-regenerating DC model. Our in vivo results demonstrated that the levels of EGFR mRNA and protein were not modulated in either paradigm, but that pEGFR levels were suppressed in satellite glia after a regenerating PSN+ DC lesion, but not after regenerating SN, or non-regenerating DC alone lesions. The low and constitutive pEGFR levels did not change in DRGN in any of the injury paradigms (Fig. 4). SN transection puts axotomised DRG into a growth activated state and drives SN axon regeneration and DRGN axon sprouting after DC lesion (Neumann and Woolf, 1999), and thus our paradoxical inconsistent finding does not support the hypothesis that EGFR activation must be suppressed in neurons in order to promote axon regeneration. The pattern of attenuated pEGFR levels in satellite glia after PSN + DC lesions was similar to that seen in retinal and optic nerve glia in a regenerating optic nerve model (Ahmed et al., 2006b) in which b3% RGC constitutively activate pEGFR, and no pEGFR is immuno-localised in RGC axons (Douglas et al., 2009). Thus, as in RGC, EGFR signalling may not be relevant to DRGN axon regeneration. AG1478-mediated suppression of pEGFR prevents the loss of RGC in a model of glaucomatous optic neuropathy (Liu et al., 2006), stimulates RGC axon regeneration in vivo (Koprivica et al., 2005) and, after DC injury, enhances functional recovery (Erschbamer et al., 2007).

298

Z. Ahmed et al. / Neurobiology of Disease 39 (2010) 292–300

Fig. 6. Knockdown of EGFR mRNA and protein did not disinhibit DRGN neurite outgrowth in the presence of CME without simultaneous FGF2 stimulation. The addition of either FGF2 (A), or siEGFR (B) failed to overcome the inhibition of neuritogenesis by CME. However, in siEGFR-treated cultures, addition of FGF2 overcame CME inhibition (C) and significantly enhanced mean neurite length (G) and the proportion of total DRGN with neurites (H). Raising cAMP levels using forskolin also overcame CME inhibition (D) but treatment of DRG cultures with siEGFR prior to raising cAMP levels significantly enhanced DRGN neurite outgrowth (E, G, H), an effect which was further potentiated by FGF2 inclusion (F, G, H). (A–F) DRGN immunostained with antibodies to βIII-tubulin. Scale bars in A–F = 50 µm; *P b 0.05; **P b 0.01; ***P b 0.0001, ANOVA).

Therefore, one might expect that EGFR gene silencing would substitute for EGFR inhibition and induce similar axogenic effects. However, direct specific blockade of EGFR expression by siRNA attenuates pEGFR, but has no effect on DRGN neurite outgrowth in DRG cultures unless AG1478 is added, suggesting that outgrowth is promoted by the offtarget AG1478 induction of neurotrophin release from glia/neurons, and also with elevation of cAMP levels (Ahmed et al., 2009a). Titres of endogenous neurotrophins were not raised in DRG cultures by siEGFR treatment, but simultaneous addition of exogenous FGF2 and/or raised cAMP levels promoted disinhibited neuritogenesis in the presence of

CME. DRGN and their axons projecting peripherally in the spinal nerves and centrally in the dorsal roots are part of the peripheral nervous system, ensheathed by Schwann cells and hence spontaneously regenerate after injury. There is no explanation of how peripheral axons are disinhibited during regeneration through the peripheral nerve environment is rich in inhibitory ligands including MAG and CSPG (Shim and Ming, 2010; Quarles, 2009; Bahr and Przyrembel, 1995). Spontaneous neurite outgrowth does occur in primary DRG cultures in which neurotrophin titres are low, but addition of CME completely inhibits this growth (Ahmed et al., 2009a). Thus, levels of neurotrophins

Z. Ahmed et al. / Neurobiology of Disease 39 (2010) 292–300

in our primary DRG cultures (presumably secreted by Schwann cells and satellite glia) are high enough to promote neurite outgrowth, but too low to induce regulated intramembraneous proteolysis (RIP) of p75NTR/ TROY and disinhibit DRGN neuritogenesis (Ahmed et al., 2006a, b; Douglas et al., 2009). In DRG cultures with added CME, silencing the EGFR gene was associated with inactivation of RhoA, but addition of FGF2 was required to provide the neurotrophic stimulus for growth cone advance. A possible explanation for the negative regulation of RhoGTP by pEGFR is that pEGFR knockdown facilitates Rho-GDP–Rho-GTP exchange through RhoA guanine nucleotide exchange factor (GEF) (Peng et al., 2010). In vitro pEGFR levels in DRGN were artefactually raised compared to the low constitutive titres normally seen in vivo. Heightened pEGFR levels have also been recorded in culture by Douglas et al. (2009) in RGC and by Ahmed et al. (2009c) in DRGN. Thus in vitro, although siEGFR knockdown in satellite glia is unlikely to disinhibit DRGN neurite outgrowth directly, it could depress RhoA activation in the latter cells indirectly allowing FGF2 to drive DRGN neurite outgrowth through a CME inhibitory culture environment. Unlike neurotrophins, FGF2 does not induce RIP of p75NTR/TROY (Ahmed et al., 2006b) but the attenuated levels of RhoA-GTP after siEGFR treatment of DRG cultures probably prevent growth cone collapse. Axon regeneration requires a neurotrophic stimulus and the induction of disinhibition, as exemplified in primary retinal cultures in which: (1), siEGFR knockdown does not induce RIP, but addition of AG1478 to siEGFRtreated RGC in the presence of CME does, and is probably induced by neurotrophins released from glia and RGC by off-target AG1478 effect (Douglas et al., 2009); and (2), ROCK antagonist blockade of inhibitory signalling in retinal cultures with added CME does not promote RGC neurite outgrowth unless CNTF is added (Ahmed et al., 2009b). In conclusion, our study demonstrates that observed changes in the transcription of growth promoting genes in DRG correlate with DRGN axon regeneration after SN and PSN + DC lesions but do not correlate with either the cellular localisation, or inactivation of EGFR in DRGN. Thus, pEGFR was localised within satellite cells and attenuated only in PSN + DC animals. In all other experimental groups, constitutive levels of pEGFR were maintained. Unlike in vivo, raised pEGFR levels were induced in DRGN in primary DRG cultures, in which DRGN neurite growth was inhibited by added CME. Knockdown of EGFR with siRNA suppressed both pEGFR and led to attenuated Rho-GTP levels, but failed to promote disinhibited DRGN neurite outgrowth unless axon growth was co-stimulated by a neurotrophic factor (FGF2), and cAMP levels were raised. In vivo, the absence of pEGFR in DRGN, and failure of pEGFR levels in satellite glia to correlate with the axon growth status of DRGN, excludes a direct intra-axonal EGFR mediated mechanism of axon growth inhibition in peripheral and central somatosensory projections. Acknowledgments This work was funded by the University of Birmingham Scientific Projects Committee and Biotechnology and Biological Sciences Research Council grant no. G181986. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.nbd.2010.04.013. References Ahmed, Z., Dent, R.G., Suggate, E.L., Barrett, L.B., Seabright, R.J., Berry, M., Logan, A., 2005. Disinhibition of neurotrophin-induced dorsal root ganglion cell neurite outgrowth on CNS myelin by siRNA-mediated knockdown of NgR, p75(NTR) and Rho-A. Mol. Cell. Neurosci. 28, 509–523. Ahmed, Z., Mazibrada, G., Seabright, R.J., Dent, R.G., Berry, M., Logan, A., 2006a. TACEinduced cleavage of NgR and p75NTR in dorsal root ganglion cultures disinhibits

299

outgrowth and promotes branching of neurites in the presence of inhibitory CNS myelin. FASEB J. 20, 1939–1941. Ahmed, Z., Suggate, E.L., Brown, E.R., Dent, R.G., Armstrong, S.J., Barrett, L.B., Berry, M., Logan, A., 2006b. Schwann cell-derived factor-induced modulation of the NgR/p75 (NTR)/EGFR axis disinhibits axon growth through CNS myelin in vivo and in vitro. Brain 129, 1517–1533. Ahmed, Z., Berry, M., Logan, A., 2009b. ROCK inhibition promotes adult retinal ganglion cell neurite outgrowth only in the presence of growth promoting factors. Mol. Cell. Neurosci. 42, 128–133. Ahmed, Z., Jacques, S.J., Berry, M., Logan, A., 2009. Epidermal growth factor receptor inhibitors promote CNS axon growth through off-target effects on glia. Neurobiol. Dis. 36, 142–150. Bahr, M., Przyrembel, C., 1995. Myelin from peripheral and central nervous system is a nonpermissive substrate for retinal ganglion cell axons. Exp. Neurol. 134, 87–93. Berry, M., Ahmed, Z., Lorber, B., Douglas, M., Logan, A., 2008. Regeneration of axons in the visual system. Restor. Neurol. Neurosci. 26, 147–174. Bomze, H.M., Bulsara, K.R., Iskandar, B.J., Caroni, P., Skene, J.H., 2001. Spinal axon regeneration evoked by replacing two growth cone proteins in adult neurons. Nat. Neurosci. 4, 38–43. Chong, M.S., Woolf, C.J., Haque, N.S.K., Anderson, P.N., 1999. Axonal regeneration from injured dorsal roots into the spinal cord of adult rats. J. Comp. Neurol. 410, 42–54. Costigan, M., Befort, K., Karchewski, L., Griffin, R.S., D'Urso, D., Allchorne, A., Sitarski, J., Mannion, J.W., Pratt, R.E., Woolf, C.J., 2002. Replicate high-density rat genome oligonucleotide microarrays reveal hundreds of regulated genes in the dorsal root ganglion after peripheral nerve injury. BMC Neurosci. 3, 16. Douglas, M.R., Morrison, K.M., Jacques, S.J., Leadbeater, W.E., Gonzalez, A.M., Berry, M., Logan, A., Ahmed, Z., 2009. Off-target effects of epidermal growth factor receptor antagonists mediate retinal ganglion cell disinhibited axon regeneration. Brain 132, 3102–3121. Du, J., Fu, C., Sretavan, D.W., 2007. Eph/ephrin signaling as a potential therapeutic target after central nervous system injury. Curr. Pharm. Des. 13, 2507–2518. Erschbamer, M., Pernold, K., Olson, L., 2007. Inhibiting epidermal growth factor receptor improves structural, locomotor, sensory, and bladder recovery from experimental spinal cord injury. J. Neurosci. 27, 6428–6435. Fabes, J., Anderson, P., Brennan, C., Bolsover, S., 2007. Regeneration-enhancing effects of EphA4 blocking peptide following corticospinal tract injury in adult rat spinal cord. Eur. J. Neurosci. 26, 2496–2505. Fischer, D., Petkova, V., Thanos, S., Benowitz, L.I., 2004. Switching mature retinal ganglion cells to a robust growth state in vivo: gene expression and synergy with RhoA inactivation. J. Neurosci. 24, 8726–8740. Fournier, A.E., GrandPre, T., Gould, G., Wang, X., Strittmatter, S.M., 2002. Nogo and the Nogo-66 receptor. Prog. Brain Res. 137, 361–369. Gomezpinilla, F., Knauer, D.J., Nietosampedro, M., 1988. Epidermal growth-factor receptor immunoreactivity in rat-brain — development and cellular-localization. Brain Res. 438, 385–390. Hokfelt, T., Wiesenfeld-Hallin, Z., Villar, M., Melander, T., 1987. Increase of galanin-like immunoreactivity in rat dorsal root ganglion cells after peripheral axotomy. Neurosci. Lett. 83, 217–220. Holmes, F.E., Mahoney, S., King, V.R., Bacon, A., Kerr, N.C., Pachnis, V., Curtis, R., Priestley, J.V., Wynick, D., 2000. Targeted disruption of the galanin gene reduces the number of sensory neurons and their regenerative capacity. Proc. Natl. Acad. Sci. U. S. A. 97, 11563–11568. Hou, S.T., Jiang, S.X., Smith, R.A., 2008. Permissive and repulsive cues and signalling pathways of axonal outgrowth and regeneration. Int. Rev. Cell Mol. Biol. 267 (267), 125–181. Hunt, D., Coffin, R.S., Anderson, P.N., 2002. The Nogo receptor, its ligands and axonal regeneration in the spinal cord; a review. J. Neurocytol. 31, 93–120. Jin, K.L., Mao, X.O., Sun, Y.J., Xie, L., Jin, L., Nishi, E., Klagsburn, M., Greenberg, D.A., 2002. Heparin-binding epidermal growth factor-like growth factor: hypoxia-inducible expression in vitro and stimulation of neurogenesis in vitro and in vivo. J. Neurosci. 22, 5365–5373. Koprivica, V., Cho, K.S., Park, J.B., Yiu, G., Atwal, J., Gore, B., Kim, J.A., Lin, E., Tessier-Lavigne, M., Chen, D.F., He, Z.G., 2005. EGFR activation mediates inhibition of axon regeneration by myelin and chondroitin sulfate proteoglycans. Science 310, 106–110. Lagord, C., Berry, M., Logan, A., 2002. Expression of TGFbeta2 but not TGFbeta1 correlates with the deposition of scar tissue in the lesioned spinal cord. Mol. Cell. Neurosci. 20, 69–92. Lisovoski, F., Blot, S., Lacombe, C., Bellier, J.P., Dreyfus, P.A., Junier, M.P., 1997. Transforming growth factor alpha expression as a response of murine motor neurons to axonal injury and mutation-induced degeneration. J. Neuropathol. Exp. Neurol. 56, 459–471. Liu, B., Neufeld, A.H., 2003. Activation of epidermal growth factor receptor signals induction of nitric oxide synthase-2 in human optic nerve head astrocytes in glaucomatous optic neuropathy. Neurobiol. Dis. 13, 109–123. Liu, B., Neufeld, A.H., 2004. Activation of epidermal growth factor receptors directs astrocytes to organize in a network surrounding axons in the developing rat optic nerve. Dev. Biol. 273, 297–307. Liu, B., Chen, H.Y., Johns, T.G., Neufeld, A.H., 2006. Epidermal growth factor receptor activation: an upstream signal for transition of quiescent astrocytes into reactive astrocytes after neural injury. J. Neurosci. 26, 7532–7540. McKerracher, L., 2001. Spinal cord repair: strategies to promote axon regeneration. Neurobiol. Dis. 8, 11–18. McKerracher, L., Winton, M.J., 2002. Nogo on the go. Neuron 36, 345–348. Neumann, S., Woolf, C.J., 1999. Regeneration of dorsal column fibers into and beyond the lesion site following adult spinal cord injury. Neuron 23, 83–91.

300

Z. Ahmed et al. / Neurobiology of Disease 39 (2010) 292–300

Peng, F., Zhang, B., Ingram, A.J., Gao, B., Zhang, Y., Krepinsky, J.C., 2010. Mechanical stretch-induced RhoA activation is mediated by the RhoGEF Vav2 in mesangial cells. Cell Signal. 22, 34–40. Planas, A.M., Justicia, C., Soriano, M.A., Ferrer, I., 1998. Epidermal growth factor receptor in proliferating reactive glia following transient focal ischemia in the rat brain. Glia 23, 120–129. Quarles, R.H., 2009. A hypothesis about the relationship of myelin-associated glycoprotein's function in myelinated axons to its capacity to inhibit neurite outgrowth. Neurochem. Res. 34, 79–86. Sandvig, A., Berry, M., Barrett, L.B., Butt, A., Logan, A., 2004. Myelin-, reactive glia-, and scar-derived CNS axon growth inhibitors: expression, receptor signaling, and correlation with axon regeneration. Glia 46, 225–251. Seijffers, R., Mills, C.D., Woolf, C.J., 2007. ATF3 increases the intrinsic growth state of DRG neurons to enhance peripheral nerve regeneration. J. Neurosci. 27, 7911–7920. Shim, S., Ming, G.L., 2010. Roles of channels and receptors in the growth cone during PNS axonal regeneration. Exp. Neurol. 223, 38–44. Smith, G.M., Strunz, C., 2005. Growth factor and cytokine regulation of chondroitin sulfate proteoglycans by astrocytes. Glia 52, 209–218.

Wakisaka, S., Kajander, K.C., Bennett, G.J., 1991. Increased neuropeptide Y (NPY)-like immunoreactivity in rat sensory neurons following peripheral axotomy. Neurosci. Lett. 124, 200–203. Wakisaka, S., Kajander, K.C., Bennett, G.J., 1992. Effects of peripheral nerve injuries and tissue inflammation on the levels of neuropeptide Y-like immunoreactivity in rat primary afferent neurons. Brain Res. 598, 349–352. White, D.M., 1998. Contribution of neurotrophin-3 to the neuropeptide Y-induced increase in neurite outgrowth of rat dorsal root ganglion cells. Neuroscience 86, 257–263. White, D.M., Mansfield, K., 1996. Vasoactive intestinal polypeptide and neuropeptide Y act indirectly to increase neurite outgrowth of dissociated dorsal root ganglion cells. Neuroscience 73, 881–887. Woolf, C.J., 2001. Turbocharging neurons for growth: accelerating regeneration in the adult CNS. Nat. Neurosci. 4, 7–9. Zhang, X., Bean, A.J., Wiesenfeld-Hallin, Z., Hokfelt, T., 1995. Ultrastructural studies on peptides in the dorsal horn of the rat spinal cord—IV. Effects of peripheral axotomy with special reference to neuropeptide Y and vasoactive intestinal polypeptide/ peptide histidine isoleucine. Neuroscience 64, 917–941.