Bone morphogenetic protein-6 promotes cerebellar granule neurons survival by activation of the MEK/ERK/CREB pathway

Abstract

Bone morphogenetic proteins (BMPs) have been implicated in the generation and postnatal differentiation of cerebellar granule cells (CGCs). Here, we examined the eventual role of BMPs on the survival of these neurons. Lack of depolarization causes CGC death by apoptosis in vivo, a phenomenon that is mimicked in vitro by deprivation of high potassium in cultured CGCs. We have found that BMP-6, but not BMP-7, is able to block low potassium-mediated apoptosis in CGCs. The neuroprotective effect of BMP-6 is not accompanied by an increase of Smad translocation to the nucleus, suggesting that the canonical pathway is not involved. By contrast, activation of the MEK/ERK/CREB pathway by BMP-6 is necessary for its neuroprotective effect, which involves inhibition of caspase activity and an increase in Bcl-2 protein levels. Other pathways involved in the regulation of CGC survival, such as the c-Jun terminal kinase and the phosphatidylinositol 3-kinase (PI3K)-Akt/PKB, were not affected by BMP-6. Moreover, failure of BMP-7 to activate the MEK/ERK/CREB pathway could explain its inability to protect CGCs from low potassium-mediated apoptosis. Thus, this study demonstrates that BMP-6 acting through the noncanonical MEK/ERK/CREB pathway plays a crucial role on CGC survival.

Figure 1.

BMP-6 protects against low potassium–induced apoptosis in CGCs. Cerebellar granule cells cultures (6 DIV) were placed in serum-free medium containing 25 mM (K25) or 5 mM (K5) KCl and when indicated, BMP-6 (100 ng/ml) or BMP-7 (30 ng/ml). Chromatin condensation was assayed 24 h later by staining with Hoechst 33258. (A) Condensed or fragmented nuclei were counted and represented as percentage versus total nuclei. Results are the mean ± SEM from three independent experiments performed in triplicate. **p < 0.01 and *p < 0.05 versus K5. (B) Fluorescence photomicrographs of representative fields from 7 DIV cultures treated as indicated. Yellow arrows show condensed nuclei.

Figure 2.

BMP-6 inhibits K5-induced caspase-3 activation. Cerebellar granule neurons (6 DIV) were placed in serum-free medium containing 25 mM (K25) or 5 mM (K5) KCl and when indicated, with BMP-6 (100 ng/ml) or BMP-7 (30 ng/ml). Activation of caspase-3 was assessed 10 h later by immunocytochemistry against its active fragment (A and B) and by an activity assay (C). Both approaches were performed in three independent experiments by triplicate. **p < 0.01 and *p < 0.05 versus K5.

Figure 3.

BMP associated Smads are localized in the cytoplasm as well as in the nucleus of cultured CGCs. Cerebellar granule cells cultures (6 DIV) were placed in serum-free medium containing 25 mM (K25) or 5 mM (K5) KCl, and BMP-6 (100 ng/ml) or BMP-7 (30 ng/ml). (A) After 1 h, cells were subjected to immunocytochemistry for Smad 1. Representative microphotographs of confocal microscope images are shown for each treatment. Hoechst staining was used to visualize cell nuclei (merge). Yellow and red arrows, cytoplasmic and nuclear localization, respectively. (B) Representative Western blots of phosphorylated and total Smad 1 in cytosolic (C) and nuclear (N) fractions. Histone H1 and GAPDH were used to monitor the purity of nuclear and cytosolic fractions, respectively. (C) Ratio between phosphorylated and total Smad 1 in the nuclear fraction. Results are the mean ± SEM of three independent experiments. No significant differences were observed.

Figure 4.

BMP-6 treatment induces ERK phosphorylation. Cerebellar granule neurons (6 DIV) were switched to a serum free medium with low potassium (K5) and/or BMP-6 (100 ng/ml). Cell lysates were obtained at the indicated times after treatment and subjected to Western blot analysis with (A) phospho-Akt and Akt, (B) phospho-JNK and JNK, and (C) phospho-ERK and ERK antibodies. Results are the mean ± SEM of four independent experiments. Significant increase in ERK phosphorylation induced by BMP-6 treatment was observed at short times (5′ and 15′) and at 8 h. **p < 0. 01; ⁺p < 0.05; ^##p < 0.01 versus K5 at 5 and 15 min and 8 h, respectively.

Figure 5.

ERK pathway activation has a key role in BMP-6 neuroprotection. At 6 DIV, CGC cultures were placed in a serum-free medium and treated with high (K25) or low (K5) potassium with or without BMP-6. The MEK inhibitor PD98059 (50 μM) was added 1 h before the treatment. (A and B) Chromatin condensation was assayed 24 h later by staining with Hoechst 33258. Condensed nuclei were counted and represented as percentage versus total nuclei. Results are the mean ± SEM from three independent experiments performed in triplicate. (C and D) The presence of the active form of caspase-3 was monitored 10 h after treatment by immunocytochemistry (see Materials and Methods). Quantification of positive immunostained cells was performed by triplicate in four experiments. Results are the mean ± SEM. (E) Caspase activity was determined by a fluorometric method after 10 h (see Materials and Methods). Results are the mean ± SEM from three independent experiments performed in duplicate. (F) Cell lysates of indicated treatments were subjected to a pulldown assay with GST-Ral-GDS. Rap1 was detected by Western blotting. Two additional assays gave similar results. **p < 0.01 and *p < 0.05 versus K5; ^××p < 0.01 and ^×p < 0.05 versus BMP-6 + PD.

Figure 6.

CREB mediates BMP-6 neuroprotective effect. Mature CGC cultures (6 DIV) were placed in a serum-free medium and treated with high or low potassium (K25 or K5) with or without BMP-6 and the MEK inhibitor, PD98059 (50 μM). (A) Phospho-CREB and total CREB levels were assayed by Western blot at 5 min. Results are the mean ± SEM of four independent experiments. (B) CGCs were treated for 1 h as indicated, and nuclear extracts were prepared for EMSA mobility assay with a CRE probe. A representative assay is shown. Two additional assays gave similar results. (C–E) Cells were infected at 4 DIV with HSV-GFP or HSV-MCREBGFP and treated as described at 6 DIV. A representative Western blot showing the decrease in CREB activation in HSV-MCREBGFP infected CGC cultures is shown in C. Data represent the mean and SEM (between brackets) from four independent Western blots. (D) Representative microphotographs of non-treated and infected cells. Condensed or fragmented nuclei of GFP-expressing cells were determined with Hoechst 33258 24 h after treatment. Data represent the percentage of condensed or fragmented nuclei versus total nuclei in GFP-expressing cells (E). Results are the mean ± SEM from three to five independent experiments performed in duplicate. **p < 0.01 and *p < 0.05 versus K5; ⁺p < 0.05 versus K5 + BMP-6; ^##p < 0.01 versus K5 + BMP-6 + HSV-MCREBGFP; ^$$ K25 + HSV-MCREBGFP.

Figure 7.

Bcl-2 acts downstream of the MEK/ERK/CREB pathway in the BMP-6 neuroprotective effect. Cerebellar granule neurons were infected at 4 DIV with HSV-GFP or HSV-MCREBGFP or plated in the presence of lentivirus of Bcl-2 shRNA or scramble shRNA as described in Materials and Methods. At 6 DIV cells were switched to K5 and treated with BMP-6 or/and PD98059 as indicated. Cleaved caspase-3 was assessed by Western blot (A and E) or immunocytochemistry (D) 10 h after the treatment. (B and C) Bcl-2 levels were assessed by Western blot 6 h after the treatment. Condensed or fragmented nuclei of treated cells were determined (F) with Hoechst 33258 24 h after treatment. Data represent the percentage of condensed or fragmented nuclei versus total nuclei. Data in K5 were considered as 100%. Actin was used as loading control in Western blot determinations. Results are the mean ± SEM from three to four independent experiments, and representative blots are shown in A, B, C, and E. ***p < 0.001 and **p < 0.01 versus K5; ⁺⁺p < 0.01 and ⁺p < 0.05 versus K5 + BMP-6; ^$$p < 0.01 versus K5 + BMP-6 HSV-GFP. ^##p < 0.01 vs. K5 + BMP-6 + shRNA scramble; ^&&p < 0.01 versus K5 + shRNA Bcl-2; n.s., not significant.

Figure 8.

BMPs role during postnatal cerebellum development. BMPs and their receptors are expressed in embryonic and postnatal cerebellum and have different roles in the generation, migration, and differentiation of CGCs. (A) BMP-6 and -7 are necessary for the generation of CGC precursors at embryonic days 14 and 15 in the proliferation zone of the rhombic limb (Alder et al., 1999; Qin et al., 2006). In the external granule cell layer (EGL), BMP-2 allows CGC precursors to enter their differentiation program by antagonizing sonic hedgehog-mediated signaling (Gao et al., 2006). CGCs complete their differentiation during the migration from the EGL toward the internal granule cell layer (IGL). It has been reported that BMP-4, which is expressed in the EGL and in migrating CGCs, promotes the differentiation of CGCs in culture. Lack of afferent stimulation during CGC migration toward the IGL causes apoptotic death. Our results show that BMP-6 promotes CGC survival in culture, suggesting that it could be an important factor regulating CGC survival during their migration toward IGL. (B) Schematic representation of the mechanisms involved in BMP-6–mediated neuroprotection of K5-mediated CGC apoptosis. BMP-6 activates the MEK/ERK/CREB/Bcl-2 pathway leading to inhibition of caspase activation induced by K5 resulting in protection (reduced) of K5-mediated apoptosis.