(Received for publication, June 13, 1996, and in revised form, August 30, 1996)
From the a Lipid and Lipoprotein Research Group and the Departments of e Biochemistry, h Anatomy and Cell Biology, and c Medicine, University of Alberta, Edmonton, Alberta, T6G 2S2, Canada
Sphingolipids are abundant constituents of neuronal membranes and have been implicated in intracellular signaling. We show that two analogs of glycosphingolipid biosynthetic intermediates, fumonisin B1 (which inhibits dihydroceramide synthesis) and DL-1-phenyl-2-palmitoylamino-3-morpholino-1-propanol (PPMP) (which inhibits glucosylceramide synthesis) decrease glycosphingolipid synthesis in rat sympathetic neurons. Although both fumonisin and PPMP inhibit glycosphingolipid synthesis, these inhibitors have differential effects on ceramide metabolism in axons. threo-PPMP, but not erythro-PPMP or fumonisin, induces an accumulation of [3H]palmitate-labeled ceramide and impairs axonal growth. Moreover, exogenously added, cell-permeable C6-ceramide, but not C6-dihydroceramide, mimicks the effect of PPMP. Our studies suggest that the lipid second messenger ceramide acts in distal axons, but not cell bodies, as a negative regulator of neurite growth.
Glycosphingolipids (GSLs)1 are major components of eukaryotic cell membranes and are particularly enriched in neuronal membranes. Lipids of this class contain one or more sugar residues attached to a sphingoid base backbone. GSLs are present in the outer leaflet of the plasma membrane (1, 2) where they have been postulated to play a role in a number of important cellular processes including cell-cell and cell-substratum recognition, adhesion, differentiation, proliferation, and oncogenic transformation. The pattern of GSLs differs among cell types and changes during development, cellular differentiation, and oncogenic transformation suggesting an important role for GSLs in cell growth and proliferation (3-5). In addition, intermediates in the biosynthesis and catabolism of sphingolipids and GSLs may function as lipid second messengers mediating the effects of extracellular agents and agonists (6-8). One approach used extensively to examine the function of GSLs is the exogenous addition of GSLs to cells, since the amphipathic nature of GSLs permits their incorporation into cellular membranes. The enrichment of gangliosides (sialic acid containing GSLs) in neuronal membranes induces neuritogenesis (3, 9-12), modulates growth factor receptor activity (reviewed in Refs. 3, 13, 14), potentiates responses to neurotrophic factors (13, 15), and protects against apoptotic death caused by withdrawal of trophic support (16). Moreover, antibodies raised against gangliosides inhibit neurite outgrowth from neural cells and tissues slices in vitro (17, 18). Tettamanti and Riboni (19) have summarized the role of gangliosides in neurodifferentiation, neuritogenesis, and synaptogenesis.
Only recently has the effect of the reduction of cellular GSL levels on
neurons, through the inhibition of endogenous GSL synthesis, been
examined. Inhibition of dihydroceramide synthesis, an early step in the
synthesis of all GSLs (see Fig. 1), in cultured hippocampal neurons
disrupts axonal growth (20) and the formation and maintenance of axonal
branches (21). Similar studies have found that sphingolipid
biosynthesis is essential for dendrite growth and survival of
cerebellar Purkinje cells (22).
In the present study, we have investigated the role of GSLs in neurite growth using two inhibitors of endogenous GSL synthesis. One inhibitor, fumonisin B1 (FB1), has been demonstrated to inhibit sphinganine N-acyltransferases (Fig. 1) in a variety of cell types such as rat hepatocytes (23), LLC-PK1 pig kidney cells (24), and hippocampal neurons (20). The second agent, PPMP (DL-1-phenyl-2-palmitoylamino-3-morpholino-1-propanol), is a potent inhibitor of glucosylceramide synthesis (Fig. 1). A less active analogue of PPMP, PDMP (DL-1-phenyl-2-decanoylamino-3-morpholinopropanol), blocks GSL synthesis in cultured 3T3 cells (25), Madin-Darby canine kidney cells (26), and hippocampal neurons (21). Although PPMP and PDMP inhibit the synthesis of glucosylceramide in cellular homogenates to a similar extent, PPMP has 10 times the potency of PDMP in intact cells (26).
We present evidence that both FB1 and PPMP inhibit GSL synthesis in cultured primary rat sympathetic neurons. Contrary to our initial hypothesis, however, our results suggest that newly synthesized GSLs are not essential for neurite growth but that the lipid second messenger ceramide negatively regulates neurite growth.
[9,10-3H]Palmitate (specific activity 54 Ci/mmol) was purchased from Amersham Canada, Oakville, Ontario, Canada. FB1 was generously provided by Dr. A. Merrill, Emory University, Atlanta, or purchased from Sigma. threo-PPMP, erythro-PPMP, C6-ceramide, and C6-dihydroceramide were purchased from Matreya Inc., Pleasant Gap, PA. NBD-C6-ceramide was purchased from Molecular Probes, Inc. Eugene, OR. Thin layer chromatography plates and high performance thin layer chromatography plates (silica gel G) were obtained from BDH Chemicals, Edmonton, AB, Canada. Standard phospholipids were either isolated from rat liver or purchased from Avanti Polar Lipids, Birmingham, AL. Ganglioside standards were kindly supplied by Dr. R. Yu, Medical College of Virginia, Richmond, VA. L15 medium without antibiotics was purchased from Life Technologies, Inc. Mouse 2.5S NGF was obtained from Cedarlane Laboratories Ltd. (Hornby, ON, Canada). Rat serum was provided by the University of Alberta Laboratory Animal Services. All other reagents were from Sigma or Fisher.
Preparation of Neuronal CulturesGeneral procedures for culture of rat sympathetic neurons have been previously described (27). Briefly, neurons from the superior cervical ganglia of newborn rats (Sprague-Dawley, supplied by the University of Alberta Farms) were enzymatically and mechanically dissociated and plated at a density of two ganglia/well in collagen-coated 24-well plates. Alternatively, the neurons were plated in the center compartment of compartmented dishes at a density of 0.6 to 0.8 ganglia/dish. The compartmented cultures were constructed as described previously (28). Collagen-coated 35-mm Falcon tissue culture dishes were scratched so that 20 parallel tracks were formed on the dish surface. A Teflon divider was sealed to the floor of the dish with silicon grease thus partitioning the dish into three compartments. The dissociated neurons were plated in the center compartment. Within 1-2 days neurites elongated along the tracks, penetrated the silicon grease barriers beneath the divider and entered the left and right compartments. L15 medium without antibiotics, but supplemented with the additives prescribed by Hawrot and Patterson (29), including bicarbonate and methylcellulose, was used as basal culture medium. Medium supplied to the center compartment containing the cell bodies was supplemented with 2.5% rat serum, 1 mg/ml ascorbic acid, 10 µM cytosine arabinoside (to prevent growth of non-neuronal cells), and 10 ng/ml NGF. In many experiments delipidated rat serum (30) was used. Medium supplied to the side compartments contained 10 or 100 ng/ml NGF, as indicated. After 6 days, cytosine arabinoside treatment was discontinued, and NGF was confined to the side compartments (31, 32). Each compartment maintained a separate fluid environment with virtually no flow between compartments (28, 32). L15 medium supplied to neurons cultured in 24-well dishes was supplemented with 2.5% rat serum, 1 mg/ml ascorbic acid, and 100 ng/ml NGF. Culture medium was changed every 3-6 days. Typically, cells were cultured for 9-13 days prior to the start of experiments.
Treatment of Neuronal CulturesFB1, erythro-PPMP, and threo-PPMP were dissolved in water to make 10 mM stock solutions. C6-Ceramide and C6-dihydroceramide were dissolved in dimethyl sulfoxide to make 10 mM stock solutions which were used to prepare medium containing the desired final concentration of inhibitor or ceramide. The final concentration of dimethyl sulfoxide in the culture medium of cells treated with C6-ceramide or C6-dihydroceramide never exceeded 0.25%. Control cultures were given a similar aliquot of dimethyl sulfoxide.
Measurement of de Novo GSL BiosynthesisThe effect of FB1 and PPMP on GSL synthesis was determined by preincubation of neurons cultured in 24-well dishes in the presence or absence of the indicated concentration of inhibitor for 4 days. [3H]Palmitate (10 µCi/ml) was present during the last 24 h. Radiolabeled medium was aspirated, and neurons were washed twice with ice-cold phosphate-buffered saline (pH 7.4). Cellular material was collected in methanol/water (8:3, v/v) and sonicated for 10 s using a probe sonicator. Chloroform was added to give a final ratio of chloroform/methanol/water of 4:8:3 (v/v). The solvents were evaporated, and gangliosides and phospholipids were extracted with diisopropyl ether/butanol/50 mM NaCl (6:4:5, v/v) as described by Ladisch and Gillard (33). Under these conditions gangliosides are recovered in the lower aqueous phase. The samples were desalted by passage through a Sephadex G-50 column. Gangliosides were separated, along with authentic standards, by high performance thin layer chromatography in the solvent system chloroform/methanol/15 mM CaCl2 (55:45:10, v/v). The plates were sprayed with resorcinol-HCl-Cu2+ reagent (34) and heated at 180 °C for 20 min. Bands corresponding to authentic standards of GM1, GM3, and GT1b were scraped, and radioactive incorporation was measured. The phospholipids PC and SM were separated by thin layer chromatography in the solvent system chloroform/methanol/acetic acid/formic acid/water (35:15:6:2:1, v/v). Bands corresponding to authentic PC and SM were scraped, and radioactivity was measured. Ceramide was separated by thin layer chromatography in two consecutive solvent systems, chloroform/methanol/acetic acid (9:1:1, v/v) followed by petroleum ether/diethyl ether/acetic acid (60:40:1, v/v). The band corresponding to authentic ceramide was scraped, and radioactivity incorporated was measured. In all cases, radioactive incorporation was normalized to total phospholipid mass.
Measurement of Neurite ExtensionNeurite extension was measured in compartmented cultures. Neuritotomy was performed by mechanical removal of distal neurites from left and right compartments with a jet of sterile distilled water delivered with a syringe through a 22-gauge needle. The water was aspirated and the wash repeated twice followed by addition of fresh culture medium. This procedure effectively removes all visible traces of neurites from the side compartments (35). Measurements of neurite growth were made as described previously (35) using a Nikon Diaphot inverted microscope with phase contrast optics outfitted with a MD2 microscope digitizer (Minnesota Datametrics Corp., Minneapolis, MN) which tracks stage movements to an accuracy of ±5 µm. An on-line computer using custom software (Minnesota Datametrics Corp.) calculated the distance from the edge of the silicone grease to the farthest extending neurite on each track and combined these measurements to obtain means and standard errors. In each culture, neurites in 16 tracks were measured in left and right compartments.
Metabolism of Fluorescent Ceramide13-Day-old neurons cultured in compartmented dishes were given 10 µM N-{6-[(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino]caproyl}-D-erythro-sphingosine (NBD-C6-ceramide) in the distal neurite-containing compartment. After 24 h, the medium was aspirated, and neurites were washed three times with ice-cold phosphate-buffered saline (pH 7.4). Cellular material was collected and lipids extracted (36) and separated by thin layer chromatography using a three-solvent system (37). Spots corresponding to authentic standards of ceramide, ceramide-1-phosphate, and SM were scraped, extracted from the silica, and quantified using a Hitachi F-2000 fluorimeter.
Other MethodsThe phospholipid content of cells was measured by lipid phosphorous determination (38).
We examined the effects of FB1 and PPMP on
sphingolipid synthesis in rat sympathetic neurons. 11-Day-old neurons,
cultured in 24-well dishes, were incubated with various concentrations of FB1, threo-PPMP, or erythro-PPMP
for 4 days. [3H]Palmitate (10 µCi/ml) was added for the
last 1 day of incubation, and incorporation of radiolabel into
gangliosides was measured. There was no advantage to using
compartmented cultures for this study since when
[3H]palmitate is supplied to distal axons the label
rapidly equilibrates throughout the neurons. Fig. 2
shows that both FB1 and threo-PPMP inhibited the
incorporation of [3H]palmitate into
GM1 in a dose-dependent manner. As
expected, erythro-PPMP did not reduce the incorporation of
[3H]palmitate into GM1 since the
erythro isomer does not inhibit sphinganine
N-acyltransferase (25). Table I summarizes
the effect of FB1 (25 µM) and PPMP (5 µM) on the incorporation of [3H]palmitate
into several gangliosides, namely GM1,
GM3, and GT1b. The incorporation of
[3H]palmitate into all three gangliosides was greatly
reduced by treatment of cells with 25 µM FB1
or 5 µM threo-PPMP, but not erythro-PPMP, indicating that FB1 and
threo-PPMP are effective inhibitors of GSL synthesis in rat
sympathetic neurons.
|
Since FB1
and PPMP inhibited GSL synthesis in rat sympathetic neurons, the
response of neurite growth to these inhibitors was investigated.
Neurons were plated in the center compartment of compartmented culture
dishes and allowed to grow for 11 days. At that time, medium containing
5 µM PPMP was given to either the center (cell
body-containing) compartment alone, or to the side (distal
axon-containing) compartments. Control cultures were given medium
without PPMP. After 3 days distal neurites were removed from the left
and right compartments (neuritotomy/axotomy), and cells were incubated
as before. Neurite extension was measured every day for the following 4 days (Fig. 3). In cultures that had been given
threo-PPMP in the distal axon-containing compartments alone,
3.7 days after neuritotomy neurite extension was 61% less than that of
cultures given medium lacking PPMP (Fig. 3A). In contrast,
growth of neurites of cultures given threo-PPMP to the cell
body-containing compartment alone was unaffected, and neurites elongated at the same rate as did untreated cells. The effect of
threo-PPMP on neurite growth was most likely specifically
related to inhibition of glucosylceramide synthase since the
erythro isomer of PPMP did not affect axonal elongation
regardless of whether it was added to the cell body- or distal
axon-containing compartments (Fig. 3B). Only in those
cultures to which threo-PPMP had been added to the distal
axons was neurite growth impaired. The same effect (i.e.
inhibition of distal neurite growth) was observed when cultures were
treated with threo-PPMP in both center and side compartments
at the same time (data not shown).
To exclude a general toxic effect of threo-PPMP on sympathetic neurons, some cultures that had been incubated with threo-PPMP in the distal axon-containing compartments for the first 3 days were given medium lacking the inhibitor after axotomy. Neurite growth resumed at approximately the normal rate (0.7 mm/day for PPMP-treated cultures versus 0.8 mm/day for untreated neurons) indicating that the effect of threo-PPMP on neurite growth was reversible. Moreover, threo-PPMP did not exert a direct cytotoxic effect since, when threo-PPMP was added to the cell body-containing compartment alone, neurites extended at the same rate as in untreated cells (Fig. 3A).
The effect of FB1 on axonal elongation was also
investigated. Sympathetic neurons, cultured for 11 days in
compartmented dishes, were given 25 µM FB1
either in the center, cell body-containing compartment alone, or in the
distal axon-containing compartments alone. After 3 days, cultures were
axotomized and given FB1 in either the side or center
compartments, as before. Control cultures were given medium without
FB1 in all compartments. Neurite extension was subsequently
measured every 24 h. Axon growth in cultures given FB1
to the cell body-containing compartment alone (Fig. 4),
to the distal axon-containing compartments alone (Fig. 4), or to both
compartments (data not shown) was identical to that in control
cultures. This result was unexpected since in these neurons 25 µM FB1 inhibited GSL synthesis to a similar
extent to that of threo-PPMP (Fig. 2 and Table I). When
cells were treated with a higher concentration of FB1 (50 µM) in the cell body- or distal axon-containing
compartments, neurite extension was 28 and 21% less than in control
cells, respectively. However, this effect was probably due to
cytotoxicity since cells incubated for 3 days in the presence of 50 µM FB1, and transferred after axotomy to
medium containing no FB1, did not recover normal axonal growth (data not shown).
Culture medium supplemented with 2.5% rat serum, which may be a source of some sphingolipids (39), is routinely given to the cell body-containing compartment. Thus, possible explanations for the observed lack of effect of the inhibitors when supplied to the cell body-containing compartment were either that sphingolipids were present in serum or that the inhibitors bound albumin, reducing the effective concentration of FB1 or PPMP in the medium. Medium containing 1% delipidated rat serum, or medium completely lacking serum, was used to minimize the possible interference of exogenously added lipids. Results obtained under these conditions were identical to those obtained using medium containing 2.5% rat serum, i.e. neurite growth was impaired only in cultures treated with threo-PPMP in distal axons and FB1 did not affect axonal elongation.
Since threo-PPMP impaired axonal growth, whereas FB1 did not, we concluded that threo-PPMP is acting on neurite extension via a mechanism other than inhibition of GSL synthesis. In addition, our results suggest that newly synthesized GSLs are not required for axonal growth.
FB1 Decreases, Whereas PPMP Increases, Labeled CeramideThe experiments presented in Fig. 2 and Table I
demonstrate that both FB1 and PPMP inhibit GSL synthesis in
rat sympathetic neurons. However, because FB1 and PPMP
inhibit different enzymes (Fig. 1), we would anticipate that the two
inhibitors would differentially affect ceramide levels: FB1
would be expected to decrease the level of ceramide, whereas PPMP
treatment would result in ceramide accumulation. In rabbit skin
fibroblasts threo-PDMP has been reported to increase the
intracellular mass of ceramide (40). Alternatively, we might expect
that PPMP would increase SM levels since threo-PDMP enhances
the synthesis of labeled SM and ceramide from
[3H]palmitate in 3T3 cells (41) and causes SM
accumulation in hippocampal neurons (21). We therefore investigated the
effect of FB1 and PPMP on the incorporation of
[3H]palmitate into SM, ceramide, and PC in rat
sympathetic neurons. Neurons were treated for 4 days according to the
protocol used in the experiment depicted in Fig. 2.
[3H]Palmitate was added during the last 24 h, and
radioactivity incorporated into SM, ceramide, and PC was measured.
FB1 treatment decreased the incorporation of
[3H]palmitate into SM by ~50% (Fig.
5C). However, neither isomer of PPMP affected the
incorporation of radioactivity into SM (Fig. 5D). Treatment
of neurons with FB1 inhibited the incorporation of
[3H]palmitate into ceramide (Fig. 5A), which
might explain the reduced incorporation of radioactivity into SM (Fig.
5C). On the other hand, treatment of neurons with
threo-PPMP, but not erythro-PPMP, increased the
incorporation of [3H]palmitate into ceramide (Fig.
5B). The effects of FB1 and PPMP on the
incorporation of [3H]palmitate into SM and ceramide were
specific since the incorporation of [3H]palmitate into PC
was not affected by either FB1 or PPMP (Fig. 5,
E and F).
Exogenously Added Ceramide Inhibits Neurite Extension
Since threo-PPMP, but not erythro-PPMP or FB1, inhibited neurite extension in rat sympathetic neurons (Figs. 3 and 4), and threo-PPMP elevated labeled ceramide levels within the neurons (Fig. 5), we considered the possibility that ceramide might be responsible for the growth inhibitory effect of threo-PPMP. If this were the case, pretreatment of neurons with FB1 might be expected to prevent the accumulation of ceramide that occurred when cells were incubated with threo-PPMP (see Fig. 1) and might thereby eliminate the inhibitory effect of threo-PPMP on neurite growth. We tested this hypothesis by incubation of distal axons of 11-day-old neurons with 25 µM FB1 for 4 days, after which the medium was changed, and fresh medium containing both FB1 (25 µM) and threo-PPMP (5 µM) was added. After 3 days, distal neurites were axotomized and both inhibitors were supplied to the side compartments for an additional 3 days, after which axon extension was measured. Some cells were given 5 µM threo-PPMP without FB1 both before and after axotomy, and other cells were incubated without FB1 and threo-PPMP throughout the experiment. In the latter cultures, neurite extension was 2.65 ± 0.04 mm. In cultures provided with threo-PPMP neurite growth was inhibited (extension was 0.93 ± 0.03 mm) as occurred in the experiment depicted in Fig. 3. However, neurite extension of neurons given both FB1 and threo-PPMP was very similar to that of neurons incubated without either inhibitor (2.48 ± 0.06 mm). The results of this experiment support the idea that ceramide inhibits axon growth.
As additional evidence that ceramide inhibits neurite extension, we
next determined the effect of cell-permeable, short-acyl chain (6 carbons) ceramide (C6-ceramide) on neurite extension. Sympathetic neurons were plated in three-compartment dishes and allowed
to grow for 14 days, after which 10 µM
C6-ceramide or 10 µM
erythro-C6-dihydroceramide was supplied to
either the cell body-containing compartment alone or the distal
axon-containing compartments alone. After 2 days the cells were
neuritotomized and re-incubated with C6-ceramide or
erythro-C6-dihydroceramide, then neurite
extension was measured every day for the following 5 days. Fig.
6A shows that neurons treated with
C6-ceramide in the cell body-containing compartment alone
elongated at the same rate as did untreated cells. On the other hand,
treatment of neurons with C6-ceramide in the distal
axon-containing compartments alone inhibited neurite elongation by
~60%. Neurite growth was also inhibited when cultures were treated
with C6-ceramide simultaneously in all compartments (data
not shown). These results correlate with the effect of
threo-PPMP on neurite growth (Fig. 3A),
indicating that ceramide might mediate the inhibitory effect of
threo-PPMP on axonal growth.
A possible explanation for the observed lack of an effect of ceramide when supplied to the cell body-containing compartment alone was that ceramide might have bound components of serum present in the medium and thereby reduce the effective concentration of available ceramide. To eliminate this possibility we repeated the experiment depicted in Fig. 6 with medium lacking rat serum. Neurite extension 2.5 days after neuritotomy was 2.04 ± 0.04 mm for cells incubated without ceramide and 2.15 ± 0.04 mm for neurons given 10 µM ceramide in the cell body-containing compartment, demonstrating that when C6-ceramide is given to cell bodies/proximal axons extension of distal axons is unaffected.
Bielawska et al. (42) have previously shown that C2-ceramide inhibits the growth of HL-60 human myelocytic leukemia cells and induces apoptosis, whereas erythro-C2-dihydroceramide is inactive. We therefore examined the effect of erythro-C6-dihydroceramide on neurite growth. Erythro-C6-dihydroceramide (10 µM) did not inhibit axonal elongation when given to either the cell body-containing compartment or the distal axon-containing compartments (Fig. 6B) confirming the specificity of the inhibitory effect of C6-ceramide on axon elongation.
Dioctanoylglycerol (diC8-glycerol) is a cell-permeable analog of diacylglycerol, a known antagonist of the effects of ceramide on viability/apoptosis (7). We therefore determined if diC8-glycerol reversed the ceramide-mediated inhibition of neurite growth. Rat sympathetic neurons were plated in compartmented dishes for 13 days. Distal neurites were removed and allowed to regenerate for 1 day, after which the right side compartments were supplied with medium containing either 10 µM C6-ceramide or 30 µM diC8-glycerol, or both compounds simultaneously. In all cultures, the left compartment contained neither C6-ceramide nor diC8-glycerol and acted as a control. Neurite extension was measured immediately before treatment and for 2 days following treatment. Neurites in right side compartments that were given C6-ceramide retracted slightly after 24 h and did not resume normal growth (neurite elongation before treatment was 0.83 ± 0.03 mm and 1.8 days after treatment was 0.55 ± 0.03 mm). Similar results were obtained in cultures given diC8-glycerol together with C6-ceramide (neurite elongation before treatment was 0.90 ± 0.03 mm and 1.8 days after treatment was 0.55 ± 0.03 mm). Neurites in the left (control) compartment elongated at a rate of 1.06 ± 0.07 mm/day. Neurites in right side compartments treated with diC8-glycerol alone elongated at the same rate as control neurites. We were unable to assess whether or not higher concentrations of diC8-glycerol reversed the inhibitory effect of ceramide on neurite growth because 100 µM diC8-glycerol inhibited neurite elongation. These observations suggest that mechanisms involved in ceramide-mediated inhibition of neurite growth differ from those involved in ceramide-mediated induction of apoptosis. The experiments also show that ceramide acts rapidly on axonal extension and suggest that the delay in the effect of PPMP is due to the time required for accumulation of ceramide within the neurites.
Fig. 7 shows photomicrographs of cultures treated with
C6-ceramide as in the previous experiment. The bottom
panel shows the right side compartment 24 h after treatment
with 10 µM C6-ceramide. The
C6-ceramide-treated neurites were significantly shorter
than were untreated neurons (top panel) and were abnormal in
appearance. Cell bodies and proximal neurites in the center compartment
appeared normal (center panel). Since cell bodies looked
morphologically normal when directly exposed to 10 µM
C6-ceramide (data not shown), we conclude that under the
conditions used C6-ceramide did not induce apoptosis.
Ceramide Metabolism
To determine whether inhibition of axonal growth by PPMP and C6-ceramide was due to ceramide itself or to a metabolite of ceramide, we studied the metabolism of a fluorescent, short chain analogue of ceramide, C6-NBD-ceramide. First, we demonstrated that C6-NBD-ceramide exhibited the same growth inhibitory properties as C6-ceramide. 14-Day-old sympathetic neurons cultured in compartmented dishes were axotomized and given 10 µM C6-NBD-ceramide in the right side compartment. The left compartment was given medium lacking ceramide as a control. After 24 h, untreated neurites had elongated by 1.03 ± 0.08 mm, whereas neurites treated with C6-NBD-ceramide extended by only 0.30 ± 0.04 mm, indicating that this analogue behaved as did C6-ceramide in terms of growth inhibition.
To analyze the metabolism of C6-NBD-ceramide in rat sympathetic neurons, 14-day-old neurons were incubated with 10 µM C6-NBD-ceramide in the distal neurite-containing compartments for 24 h, after which neurites were harvested and lipids extracted and isolated. Of the total cell-associated fluorescence 90.2% was recovered in ceramide, 2.3% was in ceramide 1-phosphate, and 2.0% was in SM, showing that in axons of sympathetic neurons, as in other cell types (21, 43, 44), short chain ceramide is inefficiently metabolized. It is likely that the small unrecovered percentage of fluorescence is present in the form of glucosylceramide or hexanoate derived from the hydrolysis of ceramide to sphingosine. Most likely, therefore, ceramide itself, rather than one of its metabolites, is responsible for inhibition of neurite growth.
Our initial goal was to investigate the role of GSLs in neurite growth. A comparison of the effects of two inhibitors of the endogenous synthesis of GSLs (FB1 and PPMP) suggests that newly synthesized GSLs are not required for normal neurite elongation, rather the lipid second messenger ceramide negatively regulates neurite growth.
Newly Synthesized GSLs Are Not Essential for Neurite GrowthBoth FB1 and threo-PPMP inhibited the incorporation of [3H]palmitate into several gangliosides in rat sympathetic neurons indicating that these agents inhibited GSL synthesis. Consequently, since only threo-PPMP, but not FB1, impaired neurite growth newly synthesized GSLs are not essential for axonal growth. This finding is in accordance with a study of Schwarz et al. (21) in which incubation of hippocampal neurons with FB1 or PDMP (an analogue of PPMP) dramatically decreased the amount of cell surface gangliosides as early as 24 h after the initiation of treatment. However, no effect on axonal growth was observed at this time or even 48 h after the addition of inhibitor; inhibition of GSL synthesis did not affect the formation of the parent axon during its emergence from the cell body. In hippocampal neurons, the most significant effect of inhibition of GSL synthesis was impairment of formation or stabilization of collateral axonal branches (21). Furuya et al. (22) observed an aberrant growth of dendrites of Purkinje cells depleted of membrane GSLs by FB1 treatment, and FB1 treatment resulted in less complex patterns of dendritic arborization, suggesting a role of GSLs in neurite branching. However, no differences in axonal outgrowth were observed between treated and untreated cultures. Unfortunately, our growth assay does not monitor branching. Nor do our experiments exclude the possibility that chronic inhibition of GSL biosynthesis, leading to depletion of cellular GSLs, would eventually affect axonal growth.
Ceramide Negatively Regulates Neurite GrowthSeveral lines of evidence from our study imply that the distinct effects of FB1 and PPMP on neurite growth relate to their differential effects on ceramide metabolism. Treatment of rat sympathetic neurons with FB1 reduced the incorporation of [3H]palmitate into ceramide (Fig. 5A), whereas the opposite effect (increased incorporation of [3H]palmitate into ceramide) was observed in threo-PPMP-treated cells (Fig. 5B). Furthermore, pretreatment of neurons with FB1 prevented the inhibitory effect of threo-PPMP on neurite growth. In addition, exogenously added cell-permeable ceramide mimicked the effect of threo-PPMP on neurite growth (Fig. 6). In light of the potent biological effects of ceramide on cells (reviewed in Refs. 7, 8), these results imply that ceramide is involved in regulation of neurite growth.
A signal transduction pathway initiated by SM hydrolysis, and leading
to generation of the second messenger ceramide, has been proposed (7,
45). Candidate cellular targets for ceramide include a
ceramide-activated protein kinase (46), a ceramide-activated protein
phosphatase (47, 48), and protein kinase C (49). A
number of extracellular agents, including tumor necrosis factor-
(50-52), interferon-
(50), and interleukin-1 (53, 54), activate SM
hydrolysis and generate ceramide upon receptor binding. The involvement
of this pathway in neuronal signaling is also suggested by the finding
that NGF (55), as well as other neurotrophins (56), rapidly activate SM
hydrolysis in cell lines expressing the low affinity neurotrophin
receptor, p75NTR. In actively dividing cells, a growing
body of data suggests a role for ceramide in growth suppression and
apoptosis (7, 8). The role of growth factor-induced SM catabolism in
post-mitotic neurons, however, has not been elucidated. Our study
suggests that ceramide negatively regulates neurite growth. It is clear from studies with compartmented rat sympathetic neurons that regulation of neurite growth in response to NGF involves mechanisms localized to
the site of NGF application and does not directly involve mechanisms within the cell body (reviewed in Ref. 57). It would be logical, therefore, that factors involved in regulation of neurite growth would
act at the site of NGF application. Ceramide is a candidate factor
since it inhibits neurite growth of compartmented sympathetic neurons
only when applied to distal neurites but not when applied to cell
bodies (Fig. 6). The specificity of the effect of ceramide on neurite
growth is demonstrated by the lack of any effect of dihydroceramide,
which differs from ceramide only in the absence of a
4,5-trans-double bond, on neurite growth (Fig. 6).
Ito and Horigome (58) have recently demonstrated that cell-permeable ceramide delays the onset of programmed cell death induced by NGF deprivation in cultured rat sympathetic neurons. The authors suggest that NGF may suppress programmed cell death by activation of SM hydrolysis via p75NTR. Thus, ceramide might be involved in both survival and growth of neurites of sympathetic neurons. A role for the p75NTR receptor in neuronal survival and axon growth has already been suggested by studies using p75-deficient mice (59-61). We speculate that NGF-responsive neurons regulate neurite growth via the cross-talk between signaling pathways induced by trkA and p75NTR. A mechanism could be envisioned in which, under conditions favoring growth, trkA signaling would be active and would suppress the ability of p75NTR to mediate SM hydrolysis to ceramide, thus promoting neurite growth. Under conditions where growth arrest and retraction of neurites are required, the activity of trkA would be inhibited leading to induction of SM hydrolysis via p75NTR and rapid arrest of neurite growth. One model of neuronal path finding suggests growth-promoting and growth-inhibitory signals are interpreted in the growth cone and translated into a directed growth response (62, 63). The local generation of ceramide may be one inhibitory signal regulating neuronal path finding.
The mechanism by which ceramide suppresses neurite growth remains to be
determined. Possibly, downstream targets of ceramide, such as
ceramide-activated protein kinase (46), ceramide-activated protein
phosphatase (47, 48), or protein kinase C (49) are
activated. The inability of dioctanoylglycerol to reverse the
inhibition of axonal growth caused by ceramide highlights important
differences between mechanisms involved in regulation of axonal growth
and those governing cell viability and induction of apoptosis since the
induction of apoptosis by ceramide is reversed by simultaneous addition
of either diacylglycerol or phorbol ester (64).
Ceramide can be converted into a number of putative bioactive lipids such as sphingosine and ceramide 1-phosphate. It is, therefore, possible that a metabolite of ceramide might be responsible for the effects of ceramide on neurite growth. Indeed, sphingosine suppresses NGF-induced neurite sprouting of PC12 cells (65). Similarly, treatment of distal neurites, but not cell bodies, of compartmented sympathetic neurons with sphingosine inhibits neurite growth (27). No effect of ceramide on the neurite growth of sympathetic neurons was observed in the latter study since long chain, non-cell-permeable ceramides were used. Addition of exogenous sphingosine rapidly elevates cellular ceramide levels in A431 cells and Neuro2a cells (66, 67) suggesting that some effects of sphingosine may be mediated by its conversion to ceramide. The reverse reaction, the catabolism of exogenous short chain ceramide to sphingosine, has also been demonstrated in Chinese hamster ovary and Madin-Darby canine kidney cells (43, 44). The time course and quantity, of sphingosine generation from exogenous short chain ceramide, however, suggest that sphingosine is unlikely to mediate the rapid effects of ceramide on cells (42). Ceramide 1-phosphate is generated via phosphorylation of ceramide by ceramide kinase (68), and the activity of a calcium-dependent ceramide kinase has been detected in rat brain synaptic vesicles (69). It is, however, unlikely that ceramide 1-phosphate mediates the effects of ceramide on neurite growth since apparently only ceramide generated from SM hydrolysis, not de novo synthesized ceramide, serves as a substrate for ceramide kinase (70). Moreover, exogenous C2-ceramide is poorly metabolized into C2-ceramide phosphate (42). We have also shown that in axons of rat sympathetic neurons only ~2% of cell-associated NBD-C6-ceramide was converted into ceramide phosphate. We, therefore, conclude that ceramide itself mediates the inhibition of the growth of neurites of sympathetic neurons.
In summary, we have demonstrated that newly synthesized GSLs are not essential for neurite growth. More importantly, these studies support and strengthen the idea that neurite growth is locally regulated within the neurite itself and strongly suggest that the lipid second messenger, ceramide, negatively regulates neurite growth.
We thank Russ Watts and Grace Martin for their excellent technical assistance, Lori O'Brien for helpful technical discussions, Dr. Robert Yu (Medical College of Virginia) for providing ganglioside standards, and Dr. Al Merrill (Emory University) for generously providing FB1.