Schwann cells are a major source of neurotrophins and cytokines, in particular during nerve regeneration. Expression of factors such as nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-4 (NT-4), members of the glial cell-line-derived neurotrophic factor (GDNF) family and leukemia inhibitory factor (LIF), which play crucial roles during development, are able to respond to nerve injury in adults, and to be up-regulated in Schwann cells of injured nerve, serve as neuronal survival and/or neurite elongation factors. In fact, exogenous administration of these factors succeeded in rescuing certain groups of neurons from programmed cell death or injury induced cell death (Lindsay 1995). In contrast, ciliary neurotrophic factor (CNTF) is unique in terms of expression during neuronal development and regeneration. Ciliary neurotrophic factor was proven to possess strong survival activity in programmed and injury induced neuronal cell death (Arakawa et al. 1990; Sendtner et al. 1990; Oppenheim et al. 1991). However, CNTF is not abundant both in the target organ and Schwann cells during development (Stockli et al. 1989). In addition, CNTF expression level is relatively higher among growth factors in intact nerves, whereas in injured and regenerating nerves, CNTF expression level becomes markedly lower (Friedman et al. 1992; Sendtner et al. 1992; Seniuk et al. 1992). This marked down-regulation of CNTF expression after nerve injury, despite its potent rescuing activity in injured neurons (Sendtner et al. 1997), has been an intriguing puzzle. In the present study, we therefore addressed why CNTF expression in Schwann cells is suppressed in response to nerve injury, using a Schwann cell line, IMS32 (Watabe et al. 1995), and mouse sciatic nerve. Here, we demonstrated that a signal via Ras extracellular-signal-regulated kinase (ERK) critically influences CNTF expression in injured Schwann cells in vivo and cultured Schwann cells in vitro.
Ciliary neurotrophic factor (CNTF) can prevent injury-induced motor neuron death. However, it is also evident that expression of CNTF in Schwann cells is suppressed during nerve regeneration. In this report, we have addressed the mechanism underlying the down-regulation of CNTF expression in injured nerves using a mouse Schwann cell line IMS32 and mouse sciatic nerve. In IMS32 cells, activation of the Ras extracellular-signal-regulated kinase (ERK) pathway by adenoviral vector-mediated expression of dominant active MEK1 did not alter a basal level of CNTF expression, whereas inhibition of the Ras-ERK pathway by using adenoviral vectors resulted in a marked increase in CNTF expression. This inverse relation between before and after axotomy was also observed in mouse sciatic nerve. In the axotomized sciatic nerve, the phosphorylated ERK was markedly increased; in contrast, the expression of CNTF was markedly decreased. These findings suggest that an inactive state of ERK is crucial for the CNTF expression in Schwann cells, and that activation of ERK following nerve injury critically influences the expression of CNTF. This might well explain why CNTF is highly expressed in quiescent Schwann cells in the peripheral nervous system, and also why CNTF is not abundant in axotomized nerves or cultured Schwann cells in which the proliferation signal is obviously active.
brain-derived neurotrophic factor
ciliary neurotrophic factor
Dulbecco's modified Eagle's medium
GTPase activated protein
glial cell line-derived neurotrophic factor
leukemia inhibitory factor
mitogen-activated protein kinase
nerve growth factor
protein kinase A
Materials and methods
A spontaneously immortalized Schwann cell line from mouse, designated IMS32 (Watabe et al. 1995), was grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, penicillin (50 units/mL), and 100 µg/mL streptomycin, at 37°C in humidified 95% air and 5% CO2 atmosphere.
Construction of adenoviral vectors
The cDNA of rat Gap1m (Maekawa et al. 1994) was donated by Dr S. Hattori (National Institute for Neuroscience, Tokyo, Japan). The cDNA of a dominant negative mutant of ERK (dnERK1 and dnERK2) were provided by Dr MH. Cobb (University of Texas South-western Medical Center, Dallas, Texas; Robbins et al. 1993). These molecules representing recombinant adenoviruses, which contain the expression unit under the control of cytomegalovirus enhancer plus chicken β-actin promoter, and β-globin polyadenylation signal (CAG) promoter (Niwa et al. 1991), were prepared using the same procedure as described in our previous work (Namikawa et al. 2000). High-titer recombinant viral stocks were generated in HEK293 cells and purified by cesium gradient centrifugation (Kanegae et al. 1994) and stored at − 80°C until use. The viral titers were determined by plaque-forming assay in HEK293 cells. AxCAdaMEK, representing constitutively active human MEK1 (designated as dominant active MEK1, daMEK1), was constructed as described elsewhere (Namikawa et al. 2000). AxCANLacZ, which contained recombinant beta-galactosidase, was provided by Drs Saito and Kanegae (University of Tokyo, Tokyo, Japan).
IMS32 cells with adenoviral vector infection
Confluent cultures of IMS32 cells were infected with various recombinant adenoviral vectors to obtain a multiplicity of infection (m.o.i) of 50, according to the procedures described elsewhere (Kanegae et al. 1994). No cytotoxic effect of the infection on morphology and viability was observed. The cells were harvested three days after infection as described below.
Mouse sciatic nerve preparation
Sciatic nerves were obtained from adult mice. Male 6-week-old BALB/c mice (n = 12) were anesthetized with sodium pentobarbital (50 mg/kg i.p.), then the right sciatic nerve was cut. Four days after transection, mice were re-anesthetized, and the bilateral sciatic nerves were dissected out and snap frozen in liquid nitrogen. Mice were killed with an overdose of anesthetic. The collected nerves were stored at − 80°C until use.
Protein was extracted with protein lysis buffer containing 20 m m HEPES, pH 7.5, 25 m m beta-glycerophosphate, 150 m m NaCl, 10% glycerol, 0.5% Triton X-100, 1.5 m m MgCl2, 2 m m EGTA, 50 m m NaF, 1 m m PMSF, 1 m m sodium vanadate, 0.5 U/mL aprotinin (A6279), and 200 m m DTT. The nerves were homogenized with a ground glass homogenizer on ice. IMS32 cells were collected from dishes with a scraper and suspended in the lysis buffer. Suspended Schwann cells were sonicated on ice-cold water for 2 min intermittently using an Astrason XL2020 sonicator. Homogenates and extracts were then centrifuged for 15 min at 14000 r.p.m. and 4°C using a microcentrifuge, and supernatant fractions were collected. The protein concentration of these samples was measured using BCA protein assay kit according to the manufacturer's instructions. These samples were stored at − 80°C until use.
Western blot analysis
First, 15 µg protein of each sample was denatured by boiling for 5 min and then applied to 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels, running at 20 mA/gel. Proteins were then transferred to polyvinylidene difluoride membrane at a constant voltage of 14 V using a semi-dry transfer apparatus. Blocking of the membranes was performed with Tris-buffered saline (TBS-T with 0.1% Tween-20) containing 5% skimmed milk overnight at 4°C, and the membranes were washed once with TBS for 5 min. For detection of CNTF, the blots were incubated with polyclonal anti-CNTF antibody diluted 1 : 500 in TBS-T/5% bovine serum albumin overnight at 4°C. The blots were washed three times, and incubated with the second antibody (anti-goat IgG biotinylated antibody) diluted 1 : 6000 in TBS-T/10% skimmed milk for an hour at room temperature (20°C). The membranes were washed three times and incubated with third antibody (antibiotin horseradish peroxidase linked antibody) (New England Biolabs, Inc., Beverly, MA, USA) diluted 1 : 1000 with TBS-T/5% skimmed milk for an hour at room temperature (20°C). For detection of ERK1 and ERK2, or phosphorylated ERK1 and ERK2, the membranes were blocked as described above and incubated with the first antibodies, mouse monoclonal anti-ERK1 antibody and rabbit polyclonal anti-phospho-p42/p44 MAP kinase antibody (NEB), respectively, diluted 1 : 2000 with TBS-T/5% BSA overnight at 4°C. The blots were washed as described above, incubated with the second antibodies, anti-mouse IgG horseradish peroxidase conjugated antibody and anti-rabbit IgG horseradish peroxidase conjugated antibody, respectively, diluted 1 : 2000 with TBS-T/5% skimmed milk for an hour at room temperature. When the final incubation was completed, each blot was washed three times again with TBS-T. The bands were visualized using an enhanced-chemiluminescence (ECL) system. These experiments were repeated at least five times, and the representative data is given.
Northern blot analysis
Total RNA was isolated from IMS32 cells according to the acid guanidine isothiocyanate/phenol/chloroform extraction method (Chomczynski and Sacchi 1987). Equal samples (20 µg denatured total RNA) were run on a 1% agarose gel containing formaldehyde, and vacuum-blotted onto nylon membrane Hybond-N+. Blotted RNA was crosslinked with a UV-crosslinker. The conditions for prehybridization, hybridization (0.5–1.0 × 106 cpm/mL) and washing were determined according to the manufacturer's instructions. The following cDNAs were utilized as a probe: 0.7 kb fragment containing the entire coding sequence of rat CNTF (Genbank Acc.; X17457) and full-length cDNA of rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH). After digestion with adequate restriction enzymes, the fragments were electrophoresed on agarose gel and purified with GENECLEAN kit II. 32P-Labeled cDNA probes were prepared with an Oligolabelling kit. After washing, the blot was exposed to Kodak X-OMAT AR film for 3 days at − 80°C with intensifying screens. This northern blot analysis was repeated three times. Because we obtained similar results in each blotting, the representative data is given.
To clarify the consequence of signaling via the Ras-ERK pathway on CNTF expression in IMS32 cells, we conducted gene transfer using adenoviral vectors that encoded genes for Gap1m, dnERK1 and 2, daMEK1 or LacZ as a control. Three days after adenoviral vector infection, the amount of CNTF mRNA in IMS32 cells was evaluated by northern blot analysis (Fig.1 a). In IMS32 cells, the basal level of CNTF transcript was very low, almost below the detection level, although other growth factors such as GDNF and NGF were highly expressed (data not shown). The overexpression of daMEK1 resulted in the entire loss of the CNTF mRNA signal; in contrast, Gap1m or simultaneous infection of dnERK1 and dnERK2 induced a marked increase of CNTF mRNA. LacZ expression, which is another control to evaluate the effect of adenoviral infection, induced no expression of CNTF mRNA. These results indicate that inhibition of the Ras-ERK pathway at the signaling point of Ras or ERK caused substantial up-regulation of CNTF mRNA expression in IMS32 cells.
We examined the correlation between ERK activity and the expression level of CNTF protein by western blotting after infection of the adenoviral vectors in IMS32 cells (Fig.1 b). The ERK activity was demonstrated using an antibody that recognizes the phosphorylated form of ERK1 and ERK2 specifically. In cultured IMS32 cells, almost full phosphorylation of ERKs occured, and the overexpression of daMEK1 could not alter the phosphorylation level further. The phosphorylation level of ERKs among the control, LacZ-expressing and daMEK1-expressing cells appeared similar, and CNTF expression level in those cells was low. In contrast, in the dominant negative ERKs-expressing IMS32 cells, phosphorylated dnERKs were found in a slightly higher position of the blot (dnERK have additional tag sequence), and the phosphorylated ERKs at the normal positions became very weak, which indicated the ERK activity was inhibited by dnERK expression. In this case, significant increase of CNTF expression was detected. Furthermore, Gap1m-expressing cells demonstrated a substantial decrease of ERK phosphorylation and a marked increase of CNTF expression. These results indicate that ERK activity and CNTF expression are inversely regulated. To further clarify the Ras-ERK signal cascade in IMS32 cells, we inhibited the activity of the signaling pathway by inducing the expression of Gap1m upstream of the pathway (at Ras level), and additionally we co-infected either daMEK1 or LacZ as a control (Fig.1 c). While co-expression of Gap1m and LacZ resulted in a lower phosphorylation level of ERKs and increased expression of CNTF in IMS32 cells; Gap1m and daMEK1 expression resulted in a higher level of ERK phosphorylation and lower expression of CNTF. This result indicates that activation of MEK1, which is located downstream of the pathway, is able to cancel the induction of CNTF expression by Gap1m.
In non-injured sciatic nerve, substantial expression of CNTF was observed, whereas the expression was exclusively inhibited in the distal part of the injured nerve (Fig. 1d). We examined the changes in ERK expression and phosphorylation level of ERKs. Nerve injury caused a loss of ERK expression in the distal part of injured sciatic nerve (4 days after injury); however, the phosphorylation of ERKs was increased after nerve injury. These results indicate that the expression of CNTF is also inversely correlated with the activity of ERKs in Schwann cells of the sciatic nerve.
In the present study, a possible molecular mechanism underlying CNTF expression regulation in the Schwann cell after nerve injury was demonstrated. The major finding was that CNTF expression is negatively regulated by the activity of ERK. Although It has been documented that no exogenous factors or cytokines can induce CNTF expression in cultured Schwann cells (Carroll et al. 1993), and that axon–Schwann cell interaction is crucial for CNTF expression (Lee et al. 1995), the present study revealed that even without axonal contact, CNTF can be induced in cultured Schwann cells simply by inhibiting Ras-ERK signaling.
The Ras-ERK signal pathway can be affected at various levels of signaling. For instance, GTPase activated protein (GAP) family members such as Gap1m can attenuate downstream ERK activity dramatically at the level of Ras (Maekawa et al. 1994). Raf-1 activity would be inhibited by activation of protein kinase A (Cook and McCormick 1993; Wu et al. 1993). Furthermore, ERK itself can be inactivated by tyrosine phosphatase or dual specificity phosphatases, which are specific phosphatases for MAP kinase/ERK (for review, see Camps et al. 2000). All these molecules are potentially able to inhibit Ras-ERK activity in myelinated or axon-attached Schwann cells. We have preliminarily examined the activation of PKA by forskolin; however, it did not influence the expression of CNTF in IMS32. Change of expression of Gap1m in the sciatic nerve, for instance, was also examined by reverse transcriptase–polymerase chain reaction strategy; however, the level of expression was not changed after nerve injury (data not shown). Thus, we have not obtained any evidence indicating the inhibitory points along the Ras-ERK signaling pathway. Alternatively, it is also possible that a distinct signal pathway that is normally active for promoting CNTF expression is somehow inhibited by ERK activity. Further investigation would be needed to reveal whether ERK inhibition directly promotes the expression of CNTF in vivo.
Although the mechanism underlying CNTF release from Schwann cells is not clear yet, CNTF is assumed to be released when Schwann cells are injured. In fact, Sendtner et al. (1992) and Curtis et al. (1993) demonstrated that a significant amount of CNTF is released after nerve injury, and that the released CNTF is biologically active. Enriched CNTF in intact nerves is an atypical characteristic among neurotrophic factors. Expression of other growth factors such as LIF, GDNF, BDNF and NGF is suppressed in the Schwann cells of intact nerves, and axonal injury promotes their expression dramatically. This suggests that CNTF, which is released at the injured site of the nerve, could be an initial or triggering factor for the synthesis and release of other factors such as GDNF, BDNF and LIF. Once the synthesis of these growth factors are initiated and they are released in an autocrine manner, these would play a pivotal role in neuronal survival and neurite elongation. These growth factors are also likely to enhance the proliferation signal (ERK signal) in Schwann cells, which would further decrease CNTF expression.
In conclusion, CNTF expression is negatively regulated by mitogenic factors that stimulate Schwann cell proliferation via the ERK pathway, which may be inhibited by axon-Schwann cell contact. Although the interaction between axon-Schwann cell contact and Ras-ERK signaling inhibition is yet unknown, the present results explain well why CNTF is abundant in intact nerves and not expressed in proliferative cultured Schwann cells and the Schwann cells of injured nerve.
This work was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture, and the Ministry of Health and Welfare. We thank Drs I. Saito and Y. Kanegae (University of Tokyo, Tokyo, Japan) for pAxCAwt and AxCANLacZ; Dr J. Miyazaki (Osaka University, Osaka, Japan) for CAG promoter in adenoviral vectors; Dr S. Hattori (National Institute for Neuroscience, Tokyo, Japan) for plasmid containing rat Gap1m; Dr M. H. Cobb (University of Texas Southwestern Medical Center, Dallas, Texas) for plasmids of dnERK1 and dnERK2. We are grateful to Mr T. Sasaki and Mr K. Hazawa for technical assistance.
1Present address: Department of Medicine, Division of Gastroenterology, Vanderbilt University Medical Center, 1161 21st Avenue South, Medical Center North, C2104, Nashville, Tennessee, USA.