Changes in cell behavior, whether differentiation or apoptosis, can be divided into three phases: induction, activation, and acquisition of signature biochemical and morphological features. Ceramide has emerged as a controversial second messenger inducing varied signals leading variously to differentiation, growth, and cell death. Since many distinct pathways can generate ceramide, the origin and perhaps intracellular location of ceramide is likely to determine its downstream signaling activity. Here we investigated what happens to ceramide as rat pheochromocytoma (PC12) and TrkA (Trk 6–24 PC12) cells differentiate or die as controlled by nerve growth factor.
Naive PC12 and TrkA cells respond differently to NGF stimulation. We confirm here that NGF stimulation activates sphingomyelin hydrolysis as well as ceramide synthesis in both cell lines. Small amounts of ceramide are generated by sphingomyelin hydrolysis in TrkA cells, suggesting that TrkA receptor-mediated inhibition of sphingomyelin hydrolysis does not occur in this system; that TrkA-mediated inhibition of sphingomyelin hydrolysis is leaky; or that we have not resolved a potential difference between acid sphingomyelinase, the presumptive target of desipramine, and neutral sphingomyelinase. Although TrkA cells differentiate in response to NGF, they are unable to generate as much ceramide as PC12 cells. PC12 cells exhibit a 2-fold increase in ceramide prior to initiating differentiation in response to NGF stimulation, and they complete differentiation only after an additional 5-fold increase in cellular ceramide. In contrast, naïve TrkA cells have already initiated differentiation even before administration of NGF. Glassman et al (24) reported that overexpression of TrkA led to receptor dimerization and potentiation of the NGF signal. We report here that TrkA cells complete differentiation with only a 2-fold increase in cellular ceramide. We conclude that minimal amounts of ceramide are required for progression of differentiation, whereas larger amounts are required for the initiation of differentiation. This is important as high amounts of ceramide activate p38 and JNK/c-Jun actin to induce apoptosis in neurons and other cell types (30, 31).
NGF-induced initiation of differentiation in PC12 cells is regulated by sphingomyelin hydrolysis, but not ceramide synthesis, as desipramine blocks both ceramide elevation and differentiation. This is in agreement with Herget et al. (32) that retinoic acid-induced differentiation is not regulated by ceramide synthesis. However, we report that NGF-induced progression of differentiation in TrkA cells is regulated by both sphingomyelin hydrolysis and ceramide synthesis, since both desipramine and FB1 block both ceramide elevation and differentiation (Fig. 7). These results suggest that during differentiation most ceramide is derived from plasma membrane sphingomyelin, with smaller quantities synthesized in cytoplasmic organelles. Furthermore, while both forms of ceramide mediate progression of differentiation, only plasma membrane-derived ceramide initiates NGF-induced differentiation. We suggest that the large increase in sphingomyelin might play a role in neurite outgrowth since sphingomyelin is an important structural component of membranes (33). During the first two days of differentiation in control PC12 cells, sphingomyelin increases substantially until it is nearly 300–350% of original content. However, as ceramide levels increase sphingomyelin levels drop by a third. Sphingomyelin content in NGF-induced TrkA cells also increases as differentiation progresses, reaching nearly 300–350% of original content by day 4 but there is no comparable drop in sphingomyelin as ceramide levels increase. The increased accumulation of SM in both PC12 and TrkA cells is most likely fueled by ceramide synthesis since the presence of FB1 in either cell blocks SM accumulation. Nevertheless, FB1 inhibits progression of differentiation only in TrkA cells. Thus, sphingomyelin accumulation in itself appears unnecessary for progression of differentiation and furthermore the capacity to generate ceramide in the absence of ceramide synthesis and thus to complete differentiation, is present only in PC12 cells.
In contrast to stimulation of growth and differentiation, apoptosis induced in PC12 cells deprived of NGF was mediated by de novo ceramide synthesis alone and not sphingomyelin hydrolysis, which suggests that these two pathways are specific to downstream events in the PC12 cell line. This finding is in general agreement with a reported finding that ceramide, generated by distinctly different mechanisms, mediates induction of either apoptosis or differentiation by retinoic acid (30). Nevertheless, TrkA cells respond to NGF withdrawal in the same way they respond to NGF stimulation, namely by activation of both ceramide pathways, which again suggests that these two pathways may substitute for one another (32) although ultimately TrkA cells do not die when NGF is withdrawn.
Although control differentiated TrkA cells have roughly 20% as much ceramide as differentiated PC12 cells, their resistance to apoptosis apparently is not due to an inability to synthesize ceramide. Differentiated TrkA cells can generate ceramide as their ceramide content doubles when NGF is removed in the presence of desipramine. Furthermore, both ceramide pathways appear to be functional in TrkA cells as both are activated by NGF deprivation. Increased ceramide levels are observed during apoptosis following by growth factor withdrawal in PC12 cells (34) and after lethal ischemia in the gerbil hippocampus (35). Furthermore, apoptosis can be induced by exogenous ceramide treatment of PC12 cells (36), mesencephalic (37), hippocampal (38) or cortical neurons (39). All of this research suggests that apoptosis is dependent on stimulation by an abundance of ceramide.
It is possible that the loss of sphingomyelin that produces the apoptotic phenotype profoundly alters the fluidity of the plasma membrane (40). However, we have not found that apoptosis is caused by loss of sphingomyelin in either of our cell lines as little sphingomyelin is lost following NGF deprivation. Although sphingomyelin hydrolysis is activated when PC12 cells are deprived of NGF, blocking only ceramide synthesis prevents apoptosis.
In neonatal sympathetic neurons, TrkA and p75NTR function antagonistically for cell survival (16). TrkA is active in promoting cell survival signals through PI3K/ Akt signaling (4, 5) and suppression of c-Jun phosphorylation and Bax activation (40, 41) while p75NTR contains an intracellular death domain that stimulates both ceramide generation and downstream signaling of JNK (16). p38, c-Jun phosphorylation by JNK, and MAPK are key regulators of the apoptotic program in post-mitotic neurons after growth factor withdrawal (42). Since ceramide is generated by sphingomyelin hydrolysis in the plasma membrane and on lysosomal membranes, we speculate that in PC12 cells, ceramide in the plasma membrane signals apoptosis and ceramide in the lysosomal membrane signals differentiation. Ceramide generation via SMase been implicated in cell death caused by anti Fas/CD95, TNFa, IL-1, IFN-g, vitamin D3, ionizing radiation, heat shock and oxidative stress (34, 43).