Molecular mechanisms involved in insulin-like growth factor I- and growth hormone-induced oligodendrocyte development

Insulin-like growth factor I (IGF-I) is a potent inducer of oligodendrocyte development and myelination in perinatal stages of development. The growth hormone (GH), the main regulator of IGF-I levels in postnatal life, has also been suggested to have a role in myelination during development, although this is still controversial. We report a stimulatory effect of both GH and IGF-I on olygodendrocyte, and the Correspondence/Reprint request: Dr. Lucinda Cacicedo, Servicio de Endocrinologia, Hospital Ramón y Cajal Carretera de Colmenar Km 9, 28034 Madrid, Spain. E-mail:lcacicedo@ hotmail.com Nuria Palacios et al. 150 intracellular mechanisms that mediate this effect are identified. By using specific inhibitors of intracellular signaling pathways, we demonstrate that the mitogenactivated protein kinase (MAPK) and the phosphatidylinositol 3-kinase (PI3K) signaling pathways are required for the full effect of IGF-I and GH on oligodendrocyte development in primary mixed rat cerebrocortical cell cultures. Both GH and IGF-I also activate the transcription factor cAMP response element binding (CREB), that is known to regulate the process of myelination. The activation of CREB in response to IGF-I involves the MAPK but not the PI3K signaling pathway, and is necessary for IGF-I to induce oligodendrocyte development. Although the somatic growth promoting effect and the metabolic actions of the GH-IGF-I system have been well established for many years, the central nervous system (CNS) was not initially considered as a GH/IGF-I target. More recently, however, GH and IGF-I effects on adult and fetal brain, both in vivo and in vitro, have also been demonstrated and CNS is now included as a GH/IGF-I target tissue. The presence of the IGF-I system components in the fetal brain has been known for more than a decade [1-3], and the manipulation of IGF-I gene or IGF-I binding protein genes is accompanied by changes in brain growth that are parallel to IGF-I levels or biodisponibility [4-6]. Immunoassayable GH is also present in the fetal whole brain in rodents, and can be found there as early as day 10 of gestation, that is before its detection in the fetal pituitary [7]; moreover brain GH immunoreactivity in the fetus is not affected by hypophysectomy [8]. GH mRNA is present in some regions of the brain in which GH immunoreactivity has been reported [9] and specific receptors for GH (GHR) have been localized in discrete areas of the rat and human CNS during development [10,11]. An increase in GHR/binding protein immunoreactivity is evident in 12 to 18-day-old embryos with localization in the nervous system beginning in the neuroepithelium [10]. The development of oligodendrocytes (OD), the myelin-forming cells in the CNS, and myelination itself are processes cleary affected by IGF-I and probably also by GH. Myelination in the rodent takes place during the first weeks of postnatal life. It is preceded by differentiation of proliferating glioblasts into mature OD, which express structural myelin proteins (such as myelin basic protein [MBP]) and they are the only cells of the lineage able to initiate myelination. In vitro IGF-I increases proliferation and differentiation of OD precursors [12-15] and enhances survival of OD and their precursors [16,17]. In vivo, IGF-I overexpression increases OD numbers, the rate of myelin synthesis, the expression of myelin-related genes and the thickness of myelin sheath [5,18]. In contrast, IGF-I-deficient mice exhibit microcephalia and hypomyelination [19]. Furthermore, IGF-I mitigates experimentally induced hypomyelination Mechanisms of IGF-Iand GH-induced oligodendrogenesis 151 [20, 21], promotes remyelination [20-22], and protects the brain from the hypomyelination that results from neonatal undernutrition [23,24], a condition associated with low IGF-I levels. With regard to GH, it has been shown that this hormone is able to increase the differentiation and myelination of brain cell agregates in vitro [25]. In vivo, immunologically-induced GH deficiency decreases OD differentiation and migration and induces brain hypomyelination [26]. In addition, Snell and Little dwarfs, both deficient in GH, show alterations related to myelination such as a diminished glial proliferation, a diminished 2’,3’-cyclic nucleotide 3’phosphodiesterase activity (a marker of myelinogenesis) and a reduced number of myelinated fibers in corpus callosum, many of which reverse after GH administration [27-30]. However, a recent study has demonstrated that the Little dwarf brain is not hypomyelinated [31] and thus the role of GH in myelination at present is not as well established as that of IGF-I. The initial event in the pathway transducing the IGF-I stimulus is binding to the type 1 IGF-I tyrosine kinase receptor. The activated receptor phosphorylates members of the insulin receptor substrate (IRS) family [32,33], mainly IRS-1, which then serves as a docking protein by binding to numerous SH2 domain-containing proteins. These include the p85 regulatory subunit of phosphatidylinositol 3-kinase (PI3K) and the guanine-nucleotide exchange factor Grb2/Sos [34-37]. Grb2/Sos binding results in sequential activation of Ras, the protein kinase Raf and the mitogen-activated protein kinase kinase (MEK1), which then activates MAPK [also called extracellular signalregulated protein kinases (ERK)] [38-40]. GH binding to the GHR, a transmembrane protein receptor belonging to the large family of cytokine receptors [41], induces receptor dimerization and activation of the tyrosine kinase JAK2. Tyrosil phosphorylation of GHR and JAK2 itself recruits and activates signaling molecules such as STAT transcription factors, SHC and insulin receptor substrates (IRS) 1 and 2 that lead to the release of second messengers such as diacylglycerol, calcium and nitric oxide, and activation of MAPK, protein kinase C (PKC), phospholipase A2, and PI3K. These pathways are known to regulate cellular functions including gene transcription, metabolite transport, and enzymatic activity that results in the ability of GH to control body growth and metabolism [42]. Demyelinating disorders are one of the most prevalent neurological diseases in humans, for which there is no decisive treatment. A potential therapeutic approach for these diseases is the stimulation of remyelination by modifying intracellular signaling mediators. As IGF-I and probably GH stimulate myelination, it was of particular interest to determine the signaling pathways involved in IGFI and GH-induced OD development. Thus, using O4 (a lipidic antigen expressed by immature OD) and MBP as markers of immature and mature OD respectively, we have investigated the signal transduction pathways involved in IGF-Iand Nuria Palacios et al. 152 GH-induced OD differentiation in a cerebrocortical cell culture system. We have particularly focused on MAPK and PI3K as both these signaling pathways are involved in both GH and IGF-I actions in other cell-types. IGF-I and GH are inducers of oligodendrocyte development To verify the effect of IGF-I and GH on OD development in our primary mixed culture system, cerebrocortical cells from postnatal and prenatal donors were exposed to IGF-I (100 ng/ml) or GH (50 ng/ml) starting at P1 (for IGF-I) and P7 (for GH), a time when very few immature OD can be seen in the culture and mature OD are not yet present. The effect of IGF-I on OD development was analyzed at different times after IGF-I exposure by measuring MBP expression using immunocytochemistry and Western blot. As seen in Fig. 1, IGF-I markedly increased the expression of MBP, confirming the known IGF-I effect on oligodendrogenesis in our culture system. As with IGF-I, GH treatment was also capable of increasing the number of MBP (+) cells as measured by immunocytochemistry (Fig. 2). These data support a GH promoting effect on myelination and OD-development in agreement with previous studies. Figure 1. IGF-I effect on MBP expression. A. Cerebrocortical cells from 1-day-old rats were incubated with or without 100 ng/ml IGF-I for 4 days (from P7 to P11) and MBP levels measured by Western immunoblot. Upper panel: representative Western blot. Lower panel: quantification of the MBP levels by scanning densitometry of the bands. The results represent the mean ± SEM from three independent experiments. ***p<0.001. B. Cerebrocortical cells from 17-day-old rat embryos were treated as in A, fixed for immunocytochemistry and stained with a monoclonal antibody against MBP. The experiment was repeated at least four times with similar results. Scale bar: 100 μm. (Reprinted from Palacios et al., J. Neurochem. 2005 with permission from Blackwell Publishing). Mechanisms of IGF-Iand GH-induced oligodendrogenesis 153 Figure 2. GH effect on MBP (+) cell number. After 4 days in vitro cerebrocortical cells from 17-day old embryos were treated with 50 ng/ml GH for 8 days. The number of mature OD was analyzed by immunocytochemistry against MBP and subsequent counting. The results represent the mean ± SEM from five independent experiments.**, p< 0.005. To determine whether the increase in MBP elicited by GH was the result of an increase in the rate of acquisition of this protein by immature OD or was the consequence of an increase in the number of immature OD that subsequently differentiate into mature OD, the effect of GH on the number of O4 immunoreactive cells was investigated. As shown in Fig. 3, GH increased the number of O4 (+) cells, reflecting a stimulatory effect on intermediate stages of OD development. However the effect of GH on O4 (+) cells was of Figure 3. GH effect on O4 (+) cell number. After 4 days in vitro cerebrocortical cells from 17-day old embryos were treated with GH for 7 days. The number of immature OD was analyzed by immunocytochemistry against O4 and subsequent counting. The results represent the mean ± SEM from five independent experiments. **, p< 0.005. Nuria Palacios et al. 154 less magnitude than the effect on MBP positive cells, maybe reflecting an effect of GH also on the latest stages of OD development i.e, an accelerated rate of differentiation from immature into mature cells. The number of O4 (+) cells seen in the cultures would then be the result of the balance between an increased appearance as a result of increased proliferation and an increased disappearance as a result of their maturation into terminally differentiated OD. Time-course of MAPKand PI3K-dependent signaling pathways activation in mixed cultures in response to IGF-I and GH In vivo, full activation of the serine/threonine protein kinase Akt results from the PI3K-dependent phosphorylation on Thr-308 and Ser-473 residues [43]. The activation of MAPK/ERK p44/p42 (ERK1/2) depends on Thr-202 and Tyr-204 residues phosphorylation by the MAPK kinase [44]. Therefore the phosphorylated levels of Akt (phospho-Akt) and ERK1/2 (phospho-ERK1/2) were analyzed by Western blotting to determine PI3K and MAPK activities respectively. Figure 4. IGF-I activates MAPKand PI3K-dependent signaling pathways in primary mixed cultures. Cerebrocortical cells from 1-day-old rats were serum starved for 24 h and then treated with 100 ng/ml IGF-I for the indicated times. Phospho-ERK1/2 (pERK) and phospho-Akt (pAkt) were measured by Western immunoblot using phopho-specific antibodies, and after stripping membranes, total (phosphorylated and unphosphorylated) levels of both kinases were measured using phosphorylation state-independent antibodies. 0 represents the levels before treatment. Upper panels: representative Western blots. Lower panels: quantification of pERK levels versus total ERK1/2 (left) and pAkt levels versus total Akt (right). The results represent the mean ± SEM from three independent experiments. *, p<0.05; **, p<0.01; ***, p<0.001 vs time 0. (Reprinted from Palacios et al. J. Neurochem. 2005 with permission from Blackwell Publishing). Mechanisms of IGF-Iand GH-induced oligodendrogenesis 155 Cerebrocortical cells, serum starved for 24 h, were treated with IGF-I for times ranging from 2 min to 24 h. As shown in Fig. 4, IGF-I treatment resulted in a marked stimulation of ERK phosphorylation that began at 2 min, reached a maximum between 5 and 20 min and returned to baseline at 24 hours. Levels of total ERK expression did not change suggesting that the increased phosphorylation was not due to a change in the total amount of these kinases. Phospho-Akt levels were also increased in response to IGF-I within 2 min, but, unlike phospho-ERK, peaked at 60 min and were still higher than the control at 24 hours. Expression of total Akt was unaltered by IGF-I treatment. Treatment with GH also resulted in a rapid stimulation of the phosphorylation of ERK1 and ERK2 that reached a maximum between 5 and 10 min. Phosphorylation declined to basal levels after 10 min even under the presence of the GH. Phospho-Akt levels were also increased in response to GH but, unlike ERK, peaked at 30 min and remained elevated at 60 min. Total ERK and Akt levels remained unchanged by GH treatment (Fig. 5). These results show the capacity of IGF-I and GH to activate the MAPK and PI3K signaling pathways albeit with a different time course. Figure 5. GH activates MAPK and PI3K dependent signaling pathways in primary mixed cultures. Cerebrocortical cells from 17-day old embryos were treated with GH for the indicated times. Phosphorylated and total ERK (A) and Akt (B) levels were measured as in Fig.4. 0 represents the levels before treatment. Activation of MAPKand PI3K-dependent signaling pathways are required for IGF-I and GH to induce OD development To elucidate whether the PI3K or MAPK signaling pathways were involved in GH and IGF-I-induced OD development, we evaluated the effect of GH and IGF-I on O4 and MBP expression in the presence or the absence of LY294002 or PD098059, specific inhibitors of PI3K and MAPK respectively. As shown in Figure 6, both LY294002 and PD098059 markedly decreased the Nuria Palacios et al. 156 IGF-I-induced MBP expression measured by Western immunoblot and immuno cytochemistry. The decrease was about 40% greater in the presence of PD098059 than in the presence of LY294002 (MBP increment in response to IGF-I in the absence of inhibitors: 7.5 ± 0.61 fold; in the presence of PD098059: 2.46 ± 0.58; in the presence of LY294002: 4,26 ± 1,49) but the difference was not statistically significant. These results indicate that both PI3K and MAPK signaling pathways are involved on IGF-I-induced OD differentiation. Figure 6. Activation of MAPK and PI3K dependent signaling pathways are required for IGFI to induce MBP expression. A. Cerebrocortical cells from 1-dayold donors were treated with 100 ng/ml IGF-I for 4 days (from P7 to P11) in the presence or absence of 20 μM LY294002 (LY), 30 μM PD098059 (PD) or both simultaneously and then MBP expression was measured by Western immunoblot. Upper panels: representative Western blots. Lower panels: quantification of the respective bands by scanning densitometry. The results are expressed as percentage of the control values and represent the mean ± SEM from three independent experiments.**, p<0.01; ***, p<0.001. B. Cerebrocortical cells from 17 day-old rat embryos were treated as in A and MBP expression measured by immunocytochemistry. The experiment was repeated twice with similar results. Scale bar: 100 μm. (Reprinted from Palacios et al., J. Neurochem. 2005 with permission from Blackwell Publishing). Mechanisms of IGF-Iand GH-induced oligodendrogenesis 157 The combined inhibition of PI3K and MAPK signaling pathways on IGF-I-induced MBP expression by simultaneous addition of LY294002 and PD098059 to the cultures resulted in IGF-I being unable to induce a noticeable increase in MBP expression (Fig. 6). This suggests that the effect of IGF-I on OD development is mediated exclusively by these two signaling pathways. As with IGF-I, the effect of GH on MBP expression measured by immunocytochemistry was markedly, but not completely inhibited, by the blockade of either MAPK or PI3K (Fig. 7). As GH effect on myelination is at least partially mediated by an increase in immature OD proliferation we also investigated the involvement of PI3K and MAPK in the GH-induced O4 (+) cell number increase. Again the presence of PD098059 and LY294002 inhibited the effect of GH on O4 (+) cell number (Fig. 7). The magnitude of the inhibition was greater in the presence of PD098059 than in that of LY294002. Treatment with the inhibitors decreased the number of mature and immature OD under basal conditions indicating that the activation of the MAPK and PI3K pathways by GH is involved in the basal oligodendrogenesis process. Figure 7. Activation of MAPKand PI3K-dependent signaling pathways are required for GH to induce oligodendrocyte development. After 4 days in vitro cerebrocortical cells from 17-day old embryos were treated with GH (50 ng/ml) in the presence or absence of 2.5 μM LY294002 (LY), 15 μM PD098059 (PD) or both simultaneously. The number of MBP(+) (left) and O4(+) (right) cells was analyzed by immunocytochemistry after 8 or 7 days respectively. The experiment was repeated five times with similar results. Scale bar: 100 μm. Nuria Palacios et al. 158 The simultaneous presence of both inhibitors also abolished the GH effects on the induction of MBP and O4 (Fig. 7) indicating that the MAPK and PI3K pathways exclusively convey the actions of GH on the differentiation of OD. The transcription factor CREB is activated in response to IGF-I via MAPK Several lines of evidence indicate that the transcription factor CREB plays an important role in oligodendrogenesis. CREB is highly expressed in oligodendroglial cells and its expression increases markedly around day 14 of postnatal life, immediately preceding the myelination peak [45-47]. As CREB was known to be activated by IGF-I in several cell types [48-51], we investigated whether CREB may be involved in IGF-I-induced OD development. First, we assayed, in a time-course experiment, the effect of IGF-I treatment on CREB activation in our culture system. As a parameter for CREB activation, we measured levels of CREB phosphorylated in the Ser 133 residue as these levels are correlated with the ability of CREB to activate transcription. Treatment with IGF-I induced an increase in Ser133 phospho-CREB levels that peaked at 10 min and returned to near basal levels after 60 min. The levels of total CREB did not change upon IGF-I addition (Fig. 8) Figure 8. IGF-I activates CREB. Cerebrocortical cells, after serum starvation for 24 h, were treated with 100 ng/ml IGF-I for the indicated times. pCREB levels were measured by Western immunoblot, and after stripping membranes, total (phosphorylated and unphosphorylated) CREB levels were measured. 0 represents the levels before treatment. Upper panels: representative Western blots. Lower panel: quantification of pCREB versus total CREB levels. The results represent the mean ± SEM from three independent experiments.**, p<0.01; ***, p<0.001. (Reprinted from Palacios et al., J. Neurochem. 2005 with permission from Blackwell Publishing). Mechanisms of IGF-Iand GH-induced oligodendrogenesis 159 To investigate the signal transduction pathways by which IGF-I activates CREB, we studied the effect of IGF-I on phospho-CREB levels in the presence of LY294002 or PD098059 inhibitors. As can be seen in Fig.9, the IGF-Iinduced increase in phospho-CREB was completely abolished by preincubating the cells with PD098059. In contrast, the PI3K inhibitor LY294002 had no effect. These results indicate that IGF-I activates CREB via a MAPK-dependent signaling pathway, without the involvement of the PI3Kdependent pathway. Figure 9. The MAPK signaling pathway mediates the IGF-I effect on CREB activation. Cerebrocortical cells, after serum starvation for 24 h, were treated with 100 ng/ml IGF-I for 10 min with or without preincubation with 20 μM LY294002 or 30 μM PD098059 for 30 min. Then pCREB levels were measured by Western immunoblot. Upper panel: representative Western blot. Lower panel: quantification of pCREB levels bands by scanning densitometry. The results represent the mean ± SEM from three independent experiments. ***, p<0.001; ns, non significant. (Reprinted from Palacios et al., J. Neurochem. 2005 with permission from Blackwell Publishing). CREB is required for IGF-I-induced OD development To determine whether the transcription factor CREB is required for IGF-I to induce OD development, CREB was blocked by preincubating the cultures with increasing amount of a rabbit-raised specific antibody against total CREB, and then the effect of IGF-I on MBP expression under these conditions was tested. As a control for the specificity of the CREB antibody, a group of Nuria Palacios et al. 160 Figure 10. CREB blockade inhibits IGF-I-induced MBP expression. A. Cerebrocortical cells from 1-day-old rats were treated with 100 ng/ml IGF-I for 4 days (from P7 to P11) in the absence or the presence of CREB antibody at the indicated concentrations, added to cultures 15 min before IGF-I. MBP expression was measured by Western immunoblot. A separate group of dishes received 10 μg of purified IgG from normal rabbit serum (NRS) as control of CREB antibody specificity. Upper panel: representative Western blot. Lower panel: quantification of the respective bands by scanning densitometry. The results are expressed as percentage of the control values and represent the mean ± percentage of the control values and represent the mean ± SEM from three independent experiments. ***, p<0.001; ns, non significant. B. Cerebrocortical cells from 17-day-old rat embryos were treated as in A (with the CREB antibody amount corrected for the volume of culture media in the wells) and MBP expression measured by immunocytochemistry. The experiment was repeated two times with similar results. Scale bar: 100 μm. (Reprinted from Palacios et al., J. Neurochem. 2005 with permission from Blackwell Publishing). Mechanisms of IGF-Iand GH-induced oligodendrogenesis 161 dishes from the same culture were preincubated with increasing amounts of purified immunoglobulins from normal rabbit serum. Although the mechanisms by which antibodies cross the plasma membrane are not well understood, their ability to penetrate into different cell types has been reported [52-56], thus supporting the validity of this approach. When cells were incubated in the presence of the CREB antibody, a reduced MBP response to IGF-I was observed (Fig. 10). This effect was concentration-dependent as indicated by the progressive decrease in IGF-I-induced MBP expression with increasing antibody concentrations. In contrast, increasing concentrations of immunoglobulins from normal rabbit serum did not modify MBP response to IGF-I (only the highest concentration is shown), confirming the specificity of the blockade induced by the CREB antibody. These findings suggest that the transcription factor CREB is needed for IGF-induced OD development. CREB is activated by GH Phospho-CREB levels were also measured in cerebrocortical cultures in response to GH treatment. As shown in Fig. 11, the exposure of cells to 50 ng/ml GH induced an increase in phospho-CREB levels that peaked 5 min after GH addition and remained elevated after 60 min, without any change in the levels of total CREB. This suggested the involvement also of CREB in GH-induced oligodendrogenesis. Figure 11. Time course of CREB activation in response to GH. Cerebrocortical cells, after serum starvation for 24 h, were incubated in the presence of GH (50 ng/ml) for the indicated times. Phosphorylated and total CREB levels were measured by