Neurotrophins induce sphingomyelin hydrolysis. Modulation by co-expression of p75NTR with Trk receptors.

We examined neurotrophin-induced sphingomyelin hydrolysis in cells expressing solely the low affinity neurotrophin receptor, p75NTR, and in PC12 cells that co-express p75NTR and Trk receptors. Nerve growth factor (NGF), brain-derived neurotrophic factor, neurotrophin-3 (NT-3), and NT-5 stimulated sphingomyelin hydrolysis with similar kinetics in p75NTR-NIH-3T3 cells. Although brain-derived neurotrophic factor (10 ng/ml) was slightly more potent than NGF at inducing sphingomyelin hydrolysis, NT-3 and NT-5 induced significant hydrolysis (30-35%) at 0.1 to 1 ng/ml in p75NTR-NIH-3T3 cells. NT-3 did not induce sphingomyelin hydrolysis in Trk C-NIH-3T3 cells nor in cells expressing a mutated p75NTR containing a 57-amino acid cytoplasmic deletion, thus demonstrating the role of p75NTR in this signal transduction pathway. In p75NTR-NIH-3T3 cells, neurotrophin-induced sphingomyelin hydrolysis 1) localized to an internal pool of sphingomyelin, 2) was not a consequence of receptor internalization, and 3) showed no specificity with respect to the molecular species of sphingomyelin hydrolyzed. In contrast to cells expressing solely p75NTR, NGF (100 ng/ml) did not induce sphingomyelin hydrolysis in PC12 cells. Interestingly, NT-3 (10 ng/ml) induced the same extent of sphingomyelin hydrolysis in PC12 cells as was apparent in p75NTR-NIH-3T3 cells. However, in the presence of NGF, NT-3 was unable to induce sphingomyelin hydrolysis, raising the possibility that Trk was modulating p75NTR-dependent sphingomyelin hydrolysis. Inhibition of Trk tyrosine kinase activity with 200 nM K252a enabled both NGF and NT-3 in the presence of NGF to induce sphingomyelin hydrolysis. These data support that p75NTR serves as a common signaling receptor for neurotrophins through induction of sphingomyelin hydrolysis and that cross-talk pathways exist between Trk and p75NTR-dependent signaling pathways.

The neurotrophins are a family of growth factors critical for the survival and development of specific populations of neurons within the central and peripheral nervous systems (1). Nerve growth factor (NGF), 1 the prototypic neurotrophin, is the best characterized member of this family of growth factors that also includes brain-derived neurotrophic factor (BDNF), and neurotrophin-3 (NT-3), -4, and -5 (2)(3)(4).
Neurotrophins interact with two classes of cellular receptors possessing both high and low affinity binding characteristics (5)(6)(7). Recent studies suggest that formation of high affinity binding sites requires the expression of a member of the trk gene family (8,9). The trk gene family encodes several receptorlinked tyrosine kinases, Trk A, Trk B, and Trk C, which preferentially interact with NGF, BDNF/NT-4/NT-5, and NT-3, respectively (10 -16). Similar to other receptor-linked tyrosine kinases (17,18), neurotrophins induce Trk receptor dimerization and autophosphorylation (19). Activation of Trk tyrosine kinase subsequently initiates multiple phosphorylation events regulating the activity of the MAP kinase cascade (20), phospholipase C␥ (21), and phosphatidylinositol-3-kinase (22). Activation of these downstream signaling components by Trk is critical for NGF-induced neurite outgrowth in PC12 cells (23).
In contrast to the rather restricted binding of neurotrophins to their respective Trk receptor, all of these molecules bind with lower affinity to a transmembrane protein known as the low affinity neurotrophin receptor, p75 NTR (24 -26). p75 NTR lacks kinase activity and possesses no structural motifs recognized to couple to established signal transduction pathways (27)(28)(29). Although p75 NTR possesses a consensus sequence for the potential binding of G-proteins (30), there is little evidence supporting a p75 NTR -dependent G-protein-mediated signal.
Recent evidence suggests that p75 NTR may also play a role in regulating cellular responses to neurotrophins. For example, p75 NTR has been demonstrated to participate in the formation of high affinity neurotrophin binding sites (21,31). Furthermore, specific domains in p75 NTR may be involved in modulating the effects of Trk tyrosine kinase activity on cell growth and differentiation (32). Moreover, we have recently demonstrated that p75 NTR may signal through a Trk-independent pathway. In this respect, NGF was found to activate a novel lipid second messenger pathway, known as the sphingomyelin cycle, specifically through p75 NTR (33). NGF-stimulated sphingomyelin metabolism resulted in the production of the bioactive sphingolipid metabolite ceramide. Ceramide has been implicated as a mediator in antimitogenic pathways leading to cell growth inhibition, cell differentiation, and apoptosis (34). Indeed, in rat T9 glioma cells, exogenous ceramide mimicked the effect of NGF on cell growth inhibition and differentiation (33). Moreover, T9 glioma cells express p75 NTR but not Trk A, suggesting that activation of the sphingomyelin cycle may mediate some of the growth suppressing and differentiative effects of neurotrophins via p75 NTR (33).
The ability of all neurotrophins to interact with p75 NTR suggested that neurotrophin-induced sphingomyelin hydrolysis may be a general signaling mechanism inherent to this family of growth factors. Therefore, in this study we examined the ability of other neurotrophins to induce sphingomyelin hydrolysis in fibroblast cells expressing wild-type or mutated forms of p75 NTR . In addition, we examined the effects of neurotrophins on sphingomyelin hydrolysis in cells that co-express both p75 NTR and Trk A. Co-expression of Trk A with p75 NTR abolished p75 NTR -dependent sphingomyelin hydrolysis. Interestingly, activation of other receptor-linked tyrosine kinases had no effect on p75 NTR -dependent sphingomyelin hydrolysis. Taken together, our results indicate that the neurotrophins induce sphingomyelin hydrolysis through p75 NTR and that coexpression of Trk and p75 NTR modulate neurotrophin-induced sphingomyelin hydrolysis.

EXPERIMENTAL PROCEDURES
Materials-Recombinant human neurotrophins were generously supplied by Genentech (BDNF, NT-3) or by Dr. Moses Chao (NT-3, NT-5). Mouse 2.5 S NGF and EGF were obtained from Harlan Bioproducts for Science. Neurotrophins were diluted in phosphate-buffered saline (PBS) and stored at 4°C. [ 3 H]Choline chloride (86 Ci/mmol) was purchased from DuPont NEN. Bacterial sphingomyelinases, bovine serum albumin, geneticin, and K252a were products of Sigma. Hygromycin B was from Calbiochem. Horseradish peroxidase-conjugated antiphosphotyrosine antibody was obtained from Signal Transduction Labs. Rabbit anti-pan Trk antibody 203 was generously supplied by Dr. David Kaplan. Protein A-Sepharose was from Pharmacia-Biotech Inc. All tissue culture media and supplements were products of Life Technologies, Inc.
Cell Lines-The production of p75 NTR -NIH-3T3 cells has been described previously (9). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum, 0.2 mg/ml hygromycin B, and 50 g/ml each of streptomycin and penicillin. Wildtype NIH-3T3 cells were maintained in the same medium in the absence of hygromycin B. PC12 cells were cultured directly on plastic tissue cultureware in DMEM supplemented with 10% fetal calf serum, 5% horse serum, and antibiotics. NIH-3T3 fibroblasts expressing Trk C were a kind gift from Dr. David Kaplan and were generated by Craig Dionne of Cephalon in collaboration with Dr. Luis Parada. PS cells, NIH-3T3 cells that express p75 NTR containing a 57-amino acid deletion in the cytoplasmic domain, were supplied by Dr. Moses Chao (35). Both cell lines were maintained in DMEM supplemented with 10% fetal calf serum and 0.2 mg/ml geneticin. All cells were maintained at 37°C in a humidified atmosphere containing 5% CO 2 .
Metabolic Labeling and Sphingomyelin Measurements-Cellular sphingomyelin pools were labeled and sphingomyelin was quantitated essentially as described previously (36). Briefly, sphingomyelin pools were labeled to metabolic equilibrium by incubation for 3 days in medium containing 0.5 Ci/ml [ 3 H]choline chloride (37). The medium was removed, and the cells were washed with PBS prior to placement in either serum-free DMEM containing 50 mM HEPES, pH 7.4, (fibroblast cell lines) or DMEM supplemented with 2% fetal calf serum and 1% horse serum (PC12 cells). The cells were incubated for 4 h in serum-free or low serum medium prior to treatment with growth factors. After transfer to a 37°C water bath, the cells were treated with growth factors for the indicated times. Subsequently, the medium was removed, and the cells were fixed with 2 ml of methanol. The cells were scraped into methanol, and lipids were extracted as described previously (33,38). Aliquots of the organic lipid extract were used for determination of total phospholipid phosphate (39) or for sphingomyelin measurements. To assay for sphingomyelin, the solvent was evaporated, and the lipid residue was resuspended in 50 l of 200 mM Tris-HCl, pH 7.5, 20 mM MgCl 2 containing 1% Triton X-100 (36). The tubes were placed in a 37°C water bath, and the reaction was initiated by the addition of 100 milliunits of Streptomyces sp. sphingomyelinase (50 l). The reaction proceeded for 2 h, and the released [ 3 H]choline phosphate was recovered in the aqueous phase of a Folch extraction (40). To ensure quantitative hydrolysis of the sphingomyelin by the bacterial sphingomyelinase, samples contained no more than 10 -15 nmol total phospholipid phosphate. In some experiments, aliquots of the lower phase of the Folch extracts were taken to assess the amount of [ 3 H]phosphatidylcholine present in the sample. Under these reaction conditions, the bacterial sphingomyelinase does not hydrolyze phosphatidylcholine (36). The amounts of sphingomyelin hydrolyzed, counts/min, were normalized to either nmol of total phospholipid phosphate or in some instances to total phosphatidylcholine counts/min (37). Sphingomyelin mass measurements were performed as described previously (36,37).
Treatment of Cells with Bacterial Sphingomyelinase-Cells were seeded at 2 ϫ 10 5 /10-cm dish and grown for 72 h in medium containing 0.5 Ci/ml [ 3 H]choline chloride. The cells were then washed with PBS and placed in serum-free medium containing 50 mM HEPES for 4 h prior to treatment. Bacterial sphingomyelinase from Staphylococcus aureus was added to a final concentration of 100 milliunits/ml, and the cells were incubated for 30 min at 37°C. This procedure has previously been shown to cleave sphingomyelin present in the outer leaflet of the plasma membrane (37). Control treatments received 50% glycerol in 0.25 M phosphate buffer as vehicle. Cells were then transferred to a 37°C water bath and treated with either PBS or growth factors for 12 min. The lipids were extracted, and sphingomyelin and phosphatidylcholine levels were quantitated as above.
Reverse Phase HPLC of Sphingomyelin Molecular Species-To determine the molecular species composition of sphingomyelin, lipids derived from [ 3 H]choline-labeled cells were subjected to RP-HPLC. Cells were labeled with [ 3 H]choline as above and treated with either PBS or 10 ng/ml NT-3 for 12 min in a 37°C water bath. Following recovery of the labeled lipids (38), samples were matched for total counts. Glycerophospholipids were removed by mild saponification at 37°C in 0.2 N methanolic KOH. The extracts were neutralized, the base-resistant lipids were recovered, and the solvent was removed by evaporation under nitrogen. The residue was dissolved in 50 l of chloroform, and an aliquot was used for RP-HPLC. Samples were fractionated on a Beckman 4.6 ϫ 250-mm octadecylsilyl column eluted isocratically with methanol, 5 mM potassium phosphate pH, 7.0 (98:2) at a flow rate of 1 ml/min (41). 70 fractions of 1 ml each were collected, and the amount of radioactivity was determined by scintillation spectrometry.
Trk C Autophosphorylation-Autophosphorylation of Trk C was performed essentially as described previously (12,42). Trk C-NIH-3T3 or wild type NIH-3T3 cells were grown to 60 -70% confluency and placed in serum-free medium for 4 h prior to treatment. Cells were treated with either PBS or 0.1-50 ng/ml NT-3 for 10 min in a 37°C water bath. The medium was rapidly aspirated, and the cells were washed with ice-cold PBS containing 0.5 mM Na 3 VO 4 . Cell lysates were prepared by the addition of 0.5 ml of ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM Na 3 VO 4 , 0.1 mM sodium molybdate, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 1 g/ml each of chymostatin, antipain, and pepstatin A). Lysates were incubated on ice for 15 min, and cell debris was sedimented at 12,000 rpm for 5 min at 4°C. Aliquots were removed for protein determination using bovine serum albumin as the standard (43). Equal amounts of protein were immunoprecipitated with 4 l of rabbit anti-Trk antibody 203 for 2 h at 4°C. A 10% slurry of protein A-Sepharose in lysis buffer was added, and the tubes were incubated for an additional 2 h at 4°C. The immune complexes were sedimented by centrifugation, washed 3 times with lysis buffer, once with distilled water and the proteins solubilized by boiling in 50 l of sample buffer (44). Proteins were resolved by SDS-polyacrylamide gel electrophoresis on 7.5% acrylamide gels. Proteins were transferred to nitrocellulose, and Trk C autophosphorylation was assessed using horseradish peroxidase-conjugated antiphosphotyrosine antibody according to the manufacturer's instructions. Bands were visualized using the ECL reagents from Amersham Corp.

Neurotrophin-induced Sphingomyelin
Hydrolysis-To examine the ability of neurotrophins to stimulate sphingomyelin hydrolysis, we utilized NIH-3T3 cells stably expressing p75 NTR (9). It is recognized that NIH-3T3 cells do not express endogenous neurotrophin receptors (10,12). Therefore, stable expression of p75 NTR in these cells allows a direct examination of the effect of p75 NTR on neurotrophin-induced sphingomyelin hydrolysis with minimal concern for an interaction with Trk receptors. We previously reported that NGF induced sphingomyelin hydrolysis in p75 NTR -NIH-3T3 cells at concentrations of 50 -150 ng/ml (33). Similarly, treatment of p75 NTR -NIH-3T3 cells with 1-100 ng/ml of BDNF found significant sphingomyelin hydrolysis occurring at 10 -100 ng/ml of BDNF (Fig. 1A). Although BDNF induced significant sphingomyelin hydrolysis at 10 ng/ml, sphingomyelin levels decreased a maximum of 30% at 100 ng/ml after 12 min of treatment. Thus, BDNF was more potent than NGF in inducing sphingomyelin hydrolysis. Additionally, both NT-3 and NT-5 showed a significantly increased potency for inducing sphingomyelin hydrolysis such that concentrations as low as 0.1 ng/ml induced a 15-30% decrease in sphingomyelin levels (Fig. 1A). Both NT-3 and NT-5 induced 30 -40% hydrolysis of sphingomyelin at 1-10 ng/ml and showed significantly greater hydrolysis than BDNF at similar concentrations. However, no significant differences were determined between the kinetics of NT-3 and BDNF-induced sphingomyelin hydrolysis (Fig. 1B). Similar to results obtained with NGF (33), maximal hydrolysis of sphingomyelin occurred within 10 -15 min of treatment with either BDNF or NT-3.
To ensure that the percent decrease in radiolabeled sphingomyelin corresponded to an actual mass decrease, sphingomyelin mass levels were determined following neurotrophin treatment. Total sphingomyelin mass in p75 NTR -NIH-3T3 cells was 40.2 Ϯ 3 pmol/nmol of phospholipid phosphate. Thus, based upon the percent of sphingomyelin hydrolyzed in the labeling studies (30 -35%), NT-3 decreased sphingomyelin levels by about 12-14 pmol/nmol of phospholipid phosphate. Correspondingly, sphingomyelin mass decreased by 16.5 pmol/nmol of phospholipid phosphate following treatment of p75 NTR -NIH-3T3 cells with 10 ng/ml NT-3 for 15 min. Collectively, these data demonstrate that p75 NTR mediates the effects of neurotrophins on sphingomyelin hydrolysis.
Activation of Trk C Does Not Couple to Sphingomyelin Hydrolysis-Since NT-3 was activating sphingomyelin hydrolysis at doses consistent with an interaction with Trk C receptors, the possibility arose that these receptors may be expressed in the p75 NTR -NIH-3T3 cells (possibly as a consequence of p75 NTR gene expression). Therefore, to directly establish the role of Trk C receptors in NT-3-induced sphingomyelin hydrolysis, we employed NIH-3T3 cells expressing Trk C. These cells express Trk C at levels similar to those measured for Trk A receptors in PC12 cells. Treatment of either wild-type NIH-3T3 cells or Trk C-NIH-3T3 cells with 10 ng/ml NT-3 did not induce any significant sphingomyelin hydrolysis ( Fig. 2A). It is possible that the lack of sphingomyelin hydrolysis in Trk C-NIH-3T3 cells may be due to Trk inducing increased sphingomyelin resynthesis. However, this is unlikely; neurotrophin treatment did not affect sphingomyelin mass levels in NIH-3T3 cells nor Trk C-NIH-3T3 cells (data not shown).
To confirm that the lack of NT-3-induced sphingomyelin hydrolysis was not due to a functional defect in activation of Trk C receptors, the effect of NT-3 on Trk C autophosphorylation was examined. As determined by anti-phosphotyrosine immunoblotting, incubation of Trk C-NIH-3T3 cells for 15 min water bath. Lipids were extracted, and sphingomyelin levels were assessed as described under "Experimental Procedures." Sphingomyelin levels were normalized to nmol of phospholipid phosphate, and the results expressed as the percent of control levels of sphingomyelin. Data are mean Ϯ S.E. and are derived from (n ϭ 9) NT-3; (n ϭ 6) BDNF and NGF; and (n ϭ 3) NT-5. Data were analyzed by analysis of variance with comparison of multiple means. Asterisks indicates p Ͻ 0.05 from control. Arrows indicate p Ͻ 0.05 from BDNF at same concentration. B, time course of NT-3 and BDNF-induced sphingomyelin hydrolysis. Cells were labeled as above and treated with 10 ng/ml NT-3 or 100 ng/ml BDNF for the indicated times in a 37°C water bath. Sphingomyelin levels were normalized to nmol of phospholipid phosphate, and the results were expressed as the percent of time-matched control levels of sphingomyelin. Results presented are the mean Ϯ S.E. of data from two experiments (n ϭ 6).

FIG. 2. Trk C is not coupled to activation of sphingomyelin hydrolysis.
A, Trk C-NIH-3T3 or wild-type NIH-3T3 cells were labeled for 3 days with [ 3 H]choline chloride, and the cells were placed in serum-free medium containing 50 mM HEPES, pH 7.4, for 4 h. Cells were treated with 10 ng/ml NT-3 or PBS for the indicated times in a 37°C water bath. Lipids were extracted and sphingomyelin levels assessed as described under "Experimental Procedures." Sphingomyelin levels were normalized to nmol of phospholipid phosphate, and the results were expressed as the percent of time matched control levels of sphingomyelin. Results presented are mean Ϯ S.E. of data from two experiments (n ϭ 6). B, NT-3 induces autophosphorylation of Trk C. Sub-confluent Trk C-NIH-3T3 cells were placed in serum-free medium containing 50 mM HEPES, pH 7.4 for 4 h prior to treatment with 0 -50 ng/ml NT-3 for 10 min in a 37°C water bath. Cells were lysed, and 0.5 mg of protein was immunoprecipitated with anti-pan Trk antibody 203 for 2 h at 4°C. Immune complexes were formed by the addition of protein A-Sepharose, and proteins were resolved by SDS-polyacrylamide gel electrophoresis. Tyrosine phosphorylation of Trk C (M r ϭ 145,000) was determined by immunoblotting as described under "Experimental Procedures." N.S. equals nonspecific binding. with 0.1-50 ng/ml of NT-3 increased Trk C autophosphorylation (Fig. 2B). Significant autophosphorylation of Trk C was seen at 10 ng/ml NT-3; a concentration that was ineffective at stimulating sphingomyelin hydrolysis in these cells. Taken together, these results support that Trk C does not couple to sphingomyelin hydrolysis and that the lack of NT-3-induced sphingomyelin hydrolysis in these cells was not due to a functional defect in Trk C.
NT-3-induced Sphingomyelin Signaling Is Sensitive to Cytoplasmic Deletions in p75 NTR -To examine potential structural requirements of neurotrophin-induced sphingomyelin hydrolysis, we utilized PS cells, which express a form of p75 NTR containing an in frame deletion of amino acids 249 -305 of the cytoplasmic domain (35). Interestingly, treatment of PS cells with 10 ng/ml NT-3 did not induce sphingomyelin hydrolysis (Fig. 3). Moreover, no changes in sphingomyelin mass were observed in response to NT-3. Although PS cells express a mutated form of p75 NTR , they bind NGF with a similar affinity as wild-type receptors (35). Since p75 NTR binds all neurotrophins with equal affinity, the lack of NT-3-induced sphingomyelin hydrolysis in PS cells is unlikely to be due to the inability of NT-3 to bind to the mutated receptor. These data suggest that amino acids within this domain may be important in coupling to either sphingomyelinase directly or to components that may serve as transducing intermediaries.
NT-3 Hydrolyzes an Internal Pool of Sphingomyelin-Recent studies on the membrane localization of sphingomyelin has identified two distinct pools that are defined by their sensitivity to hydrolysis by bacterial sphingomyelinase (37). The first pool is sensitive to hydrolysis by bacterial sphingomyelinase and resides in the outer leaflet of the plasma membrane. The second pool is internal and is resistant to bacterial sphingomyelinase treatment. To ascertain which pool of sphingomyelin was hydrolyzed by NT-3, metabolically labeled p75 NTR -NIH-3T3 cells were pretreated with 100 milliunits/ml of bacterial sphingomyelinase and then incubated with either PBS or 10 ng/ml NT-3. Pretreatment of cells with bacterial sphingomyelinase hydrolyzes the sphingomyelin in the outer leaflet of the plasma membrane (37). Therefore, following pretreatment with bacterial sphingomyelinase, any additional neurotrophin-induced loss of sphingomyelin defines a distinct neurotrophinsensitive pool.
Treatment of p75 NTR -NIH-3T3 cells with NT-3 alone induced a 26% decrease in sphingomyelin, defining the NT-3 sensitive pool (Fig. 4). Pretreatment of cells with bacterial sphingomyelinase decreased cellular sphingomyelin levels 52%, defining an outer leaflet bacterial sphingomyelinase-sensitive pool. If NT-3 is hydrolyzing a distinct pool of sphingomyelin, then treatment with bacterial sphingomyelinase plus NT-3 should decrease sphingomyelin levels by about 75-80% (52 ϩ 26%). Indeed, treatment of p75 NTR -NIH-3T3 cells with both bacterial sphingomyelinase and NT-3 decreased sphingomyelin levels by 70%; indicating that the effects of NT-3 on sphingomyelin are additive to the effects of bacterial sphingomyelinase. To determine the specificity of this response, p75 NTR -NIH-3T3 cells were also treated with EGF Ϯ bacterial sphingomyelinase pretreatment. Similar to results obtained in rat T9 glioma cells (33), EGF did not induce sphingomyelin hydrolysis. Subsequent to bacterial sphingomyelinase treatment, EGF also had no effect on the remaining pool of sphingomyelin (Fig. 4). Taken together, these results demonstrate that NT-3 specifically decreases a distinct pool of sphingomyelin that is not accessible to bacterial sphingomyelinase and that may reside on the internal leaflet of the plasma membrane (37).
Neurotrophin-induced Sphingomyelin Hydrolysis Does Not Require Receptor Internalization-Previous work has suggested that in cells expressing solely p75 NTR , the receptorligand complex may become internalized (45). Indeed, Kahle and Hertel (1992) speculated that receptor internalization may The cells were then transferred to a 37°C water bath and treated with PBS, 10 ng/ml NT-3, or 10 ng/ml EGF for 12 min. Lipids were extracted, and sphingomyelin was quantitated as described under "Experimental Procedures." Sphingomyelin levels were normalized to total phosphatidylcholine counts/min, and the results were expressed as the percent of control. Results are the mean Ϯ S.E. (n ϭ 3) of data from one representative experiment performed twice.
initiate signal transduction through this receptor. Since the NT-3-sensitive pool of sphingomyelin resides internally, we examined if internalization of p75 NTR was coupled to activation of sphingomyelin hydrolysis. Previous studies have indicated that treatment of cells in hyperosmolar medium blocks endocytosis and internalization of the tumor necrosis factor-␣, transferrin, and interleukin-1␤ receptors (46 -48). Therefore, we blocked receptor internalization by treatment in hypertonic medium and examined the effect on neurotrophin-induced sphingomyelin hydrolysis. Preincubation of p75 NTR -NIH-3T3 cells in 0.4 M sucrose for 30 min had no effect on the extent of NT-3-induced sphingomyelin hydrolysis (Table I). These data indicate that neurotrophin-induced sphingomyelin hydrolysis does not require receptor internalization and further support that the NT-3-sensitive pool of sphingomyelin resides at the plasma membrane.
Molecular Species Composition of the Neurotrophin-Sensitive Sphingomyelin Pool-Next, we determined the molecular species composition of the neurotrophin-sensitive sphingomyelin pool. Based upon RP-HPLC, p75 NTR -NIH-3T3 cells contain primarily one species of sphingomyelin (Fig. 5A). This species was tentatively identified as sphingomyelin composed of an 18:1 sphingoid base and an 18:0 fatty acid (stearoyl). Several minor peaks are also evident and identify sphingomyelin species differing in fatty acyl chain length and extent of unsaturation (41). NT-3-induced sphingomyelin hydrolysis was primarily confined to the major molecular species of sphingomyelin, although some hydrolysis of the minor species also occurred. These results suggest that no specific molecular species of sphingomyelin resides solely in the neurotrophin-sensitive pool. Indeed, PC12 cells display a more diverse molecular species composition of sphingomyelin than the fibroblast cell line (Fig. 5B). However, neurotrophin-induced sphingomyelin hydrolysis in PC12 cells (see below) was also not confined to any single molecular species of sphingomyelin.
Effect of Trk Co-expression on p75NTR-dependent Neurotrophin-induced Sphingomyelin Hydrolysis-Our results demonstrate that neurotrophins induce sphingomyelin hydrolysis in cells expressing solely p75 NTR . However, since many neurotrophin-responsive cells co-express Trk receptors and p75 NTR , it became important to determine the effect of receptor co-expression on neurotrophin-induced sphingomyelin hydrolysis. To examine this interaction, we utilized PC12 cells that co-express Trk A and p75 NTR (7,26). Intriguingly, incubation of PC12 cells with 100 ng/ml NGF did not induce a significant decrease in sphingomyelin levels (Fig. 6). However, NGF (100 ng/ml) did induce a significant increase in cellular tyrosine phosphorylation; demonstrating the functional signaling of Trk in the PC12 cells (data not shown).
The absence of NGF-induced sphingomyelin hydrolysis in PC12 cells was unexpected since this concentration of NGF effectively induced sphingomyelin hydrolysis in both T9 glioma cells and p75 NTR -NIH-3T3 cells. However, both of these cell lines express solely p75 NTR (33), while PC12 cells express p75 NTR at approximately 36-fold the level of Trk A. To determine if the presence of Trk A was affecting NGF-induced p75 NTR -dependent sphingomyelin hydrolysis, PC12 cells were incubated with NT-3. PC12 cells lack Trk C receptors and are unresponsive to NT-3 in terms of tyrosine phosphorylation and the ability to develop neurites (49). However, treatment of PC12 cells with 10 ng/ml NT-3 induced significant sphingomyelin hydrolysis over a time course similar to that observed in p75 NTR -NIH-3T3 cells (Figs. 6 and Fig. 5B). Although NT-3 can bind to Trk A (6), NT-3-induced sphingomyelin hydrolysis was not due to activation of Trk. NT-3 (10 ng/ml) had no effect on cellular tyrosine phosphorylation in PC12 cells (data not shown).  5. NT-3 does not hydrolyze a specific molecular species of sphingomyelin. p75 NTR -NIH-3T3 cells (A) or PC12 cells (B) were labeled with [ 3 H]choline chloride for 3 days and placed in serum-free medium containing 50 mM HEPES, pH 7.4, or DMEM containing 3% serum for 4 h, respectively. Cells were then treated with PBS or 10 ng/ml NT-3 in a 37°C water bath, and the lipids were extracted. Aliquots were used for determination of total radioactivity, and the samples were matched for total counts/min. Glycerophospholipids were removed by mild base hydrolysis, and the base-resistant lipids were extracted. Aliquots of the organic phase were quantitatively transferred and evaporated under nitrogen, and the lipid residue was dissolved in 50 l of chloroform. The molecular species of sphingomyelin were analyzed by RP-HPLC and eluted isocratically with methanol, 5 mM potassium phosphate, pH 7.0 (98:2) at a flow rate of 1 ml/min. One-ml fractions were collected, and the radioactivity was quantitated.
Trk Tyrosine Kinase Activity Specifically Modulates p75 NTRdependent Sphingomyelin Hydrolysis-The above results suggest that Trk expression may modulate p75 NTR -dependent sphingomyelin hydrolysis. To examine this possibility, PC12 cells were co-incubated with NGF (100 ng/ml) plus NT-3 (10 ng/ml), and sphingomyelin levels were measured. Co-incubation of NGF plus NT-3 abolished NT-3-induced sphingomyelin hydrolysis (Fig. 7). However, since NGF was at a 10-fold excess relative to NT-3, it was possible that the lack of NT-3-induced sphingomyelin hydrolysis may be due to simple ligand competition and not to cross-talk between Trk tyrosine kinase and p75 NTR -dependent signaling pathways. To further examine this hypothesis, PC12 cells were preincubated with the tyrosine kinase inhibitor K252a. K252a selectively inhibits Trk tyrosine kinase activity without affecting other receptor-linked tyrosine kinases (50). If NGF-induced activation of Trk is modulating p75 NTR -dependent sphingomyelin hydrolysis, then inhibition of Trk tyrosine kinase activity by K252a should enable NGF or NGF plus NT-3 to stimulate sphingomyelin breakdown. Indeed, incubation of PC12 cells with 200 nM or 2 M K252a rendered NGF or NGF ϩ NT-3 capable of stimulating sphingomyelin hydrolysis. In both instances, NGF and NGF ϩ NT-3 stimulated hydrolysis almost to the level induced by NT-3 alone (Fig. 7). The presence of K252a had no effect on NT-3induced sphingomyelin hydrolysis indicating that the drug itself does not effect sphingomyelinase activity.
Next, to determine if the inhibition of p75 NTR -dependent sphingomyelin hydrolysis was due to a general inhibitory signal generated by receptor-linked tyrosine kinases, PC12 cells were incubated with either EGF or platelet-derived growth factor in the presence or absence of NT-3. If activation of cellular tyrosine kinase activity is sufficient to inhibit p75 NTRdependent sphingomyelin hydrolysis, then EGF and plateletderived growth factor should inhibit NT-3-induced sphingomyelin hydrolysis. However, both ligands had no effect on NT-3induced sphingomyelin hydrolysis (Fig. 8). These results suggest that Trk tyrosine kinase activity was specifically modulating p75 NTR -dependent sphingomyelin hydrolysis.
Taken together, these data support that specific cross-talk pathways exist between Trk and p75 NTR -dependent signaling pathways.

DISCUSSION
Our data establish that neurotrophin-induced sphingomyelin hydrolysis is mediated solely through p75 NTR . Although NGF, BDNF, NT-3, and NT-5 bind to p75 NTR with rather equivalent affinities (25,26), the neurotrophins displayed a differential ability to induce sphingomyelin hydrolysis in p75 NTR -NIH-3T3 cells. In this respect, both NT-3 and NT-5 were about 100-fold more potent at inducing sphingomyelin hydrolysis than BDNF and NGF. However, BDNF, NT-3, and NGF (33) showed similar kinetics for inducing sphingomyelin hydrolysis in p75 NTR -NIH-3T3 cells.
That p75 NTR can recognize the various neurotrophins as similar, but not identical, molecules has been previously noted (25,26,51,52). For example, NGF, BDNF, and NT-3 show distinct rates of dissociation from p75 NTR (25,26). Therefore, it is possible that differences in the dissociation rates of the neurotrophins for p75 NTR may affect their ability to induce sphingomyelin hydrolysis.
An additional distinction between the neurotrophins is that NT-3 and BDNF exhibit positive cooperativity in their binding to p75 NTR at low ligand concentrations (25,26). As such, at low NT-3 concentrations (0.1-1 ng/ml), receptor occupancy of p75 NTR would be predicted to be low. However, significant sphingomyelin hydrolysis occurred at these concentrations of NT-3, suggesting that only a small percentage of p75 NTR receptors may need to be occupied to activate sphingomyelin hydrolysis. Such a mechanism may be analogous to neurotrophin-de- pendent neuronal survival where viability is enhanced well below full occupancy of high affinity receptors (6,49). Alternatively, a highly localized pool of p75 NTR may be involved in signal transduction (see below).
Although NT-3 induced significant sphingomyelin hydrolysis at ligand concentrations near the K d for Trk C (1.8 ϫ 10 Ϫ11 M; ϳ0.5 ng/ml) (15,26), NT-3 did not induce sphingomyelin hydrolysis in fibroblasts expressing Trk C. These data strongly suggest that Trk C does not directly activate neurotrophininduced sphingomyelin hydrolysis.
It is intriguing to speculate that differences in the potencies of the neurotrophins to activate sphingomyelin hydrolysis may reside in differences in the amino acids that influence neurotrophin binding to p75 NTR . Lysines 32, 34, and 95 are critical for the recognition of NGF by p75 NTR (24,51,52). Although BDNF lacks Lys-32 and Lys-34, three positively charged residues at positions 95, 96, and 97 sufficiently compensate for this absence (52), enabling BDNF to compete equally with NGF for binding to p75 NTR (25). However, BDNF was slightly more potent than NGF at inducing sphingomyelin hydrolysis. In contrast to BDNF, NT-3 and NT-4 have a conserved substitution of one or two Arg residues for Lys-32 and Lys-34, respectively (52). It is intriguing that both of these neurotrophins were 100-fold more potent at inducing sphingomyelin hydrolysis than NGF and BDNF.
In terms of structural requirements of p75 NTR , which may affect sphingomyelin breakdown, deletion of a 57-amino acid region within the cytoplasmic domain of p75 NTR abolished NT-3-induced sphingomyelin hydrolysis. This deletion removed amino acids 249 -305 within the cytoplasmic domain and partially disrupted the highly conserved juxtamembrane region of p75 NTR (35). The lack of NT-3-induced sphingomyelin hydrolysis in cells expressing this p75 NTR mutant suggests that this region of the cytoplasmic domain is necessary for stimulation of sphingomyelinase activity. Whether this region regulates direct coupling to sphingomyelinase or to intermediary effector proteins remains to be determined.
An additional consequence of this cytoplasmic deletion is the removal of Cys-279, which is a probable site of palmitoylation in p75 NTR (53). The functional role of p75 NTR palmitoylation is uncertain. However, receptor palmitoylation is associated with targeting of certain proteins to caveolae (54), small plasmallemal invaginations particularly enriched in sphingolipids (55,56). In this respect, we have localized a pool of p75 NTR to caveolae and have found that NT-3 induced significant sphingomyelin hydrolysis in this membrane fraction. 2 Therefore palmitoylation may help target p75 NTR to caveolae which contain an NT-3-sensitive pool of sphingomyelin. However, it remains to be determined if receptor palmitoylation is necessary for coupling of p75 NTR to sphingomyelin hydrolysis. As previously mentioned, p75 NTR contains an 11-amino acid sequence (amino acids 370 -381) homologous to the 14-amino acid wasp venom peptide mastoparan (30). However, the cytoplasmic deletion of the p75 NTR mutant does not encompass the mastoparan-like sequence, suggesting that the presence of this sequence is not sufficient to induce sphingomyelinase activation.
The cellular location of the neurotrophin-sensitive pool of sphingomyelin was addressed by exploiting the inability of bacterial sphingomyelinase to hydrolyze pools of sphingomyelin, which do not reside on the external leaflet of the plasma membrane. Treatment with bacterial sphingomyelinase hydrolyzed approximately 50% of total cellular sphingomyelin from p75 NTR -NIH-3T3 cells. This result is similar to those obtained from HL-60 leukemia cells (37). Thus, about 50% of the cellular sphingomyelin resides in a bacterial sphingomyelinase resistant pool. In HL-60 cells, the bacterial sphingomyelinase-resistant pool is the site of sphingomyelin hydrolysis induced by both tumor necrosis factor-␣ and vitamin D 3 (37). Similarly, NT-3 hydrolyzed a portion of the bacterial sphingomyelinase-resistant pool of sphingomyelin.
Based upon subcellular fractionation, the agonist-sensitive pool of sphingomyelin resides in the plasma membrane (37). However, since endosomes co-fractionate with plasma membrane, it was possible that agonist-sensitive sphingomyelin hydrolysis may occur in endosomal vesicles following receptor internalization. Although the internalization of p75 NTR is slow and unaffected by ligand binding (57), it is possible that receptor internalization via the endosomal pathway was a potential site for sphingomyelin degradation. Treatment of cells in hyperosmotic medium has been shown to effectively block the internalization of the transferrin, interleukin-1␤, and tumor necrosis factor-␣ receptors (46 -48). However, hyperosmolarity had no effect on NT-3-induced sphingomyelin hydrolysis, suggesting that receptor sequestration into endosomes is not a major site of p75 NTR -dependent sphingomyelin hydrolysis.
Neurotrophins did not specifically hydrolyze one molecular species of sphingomyelin. This result was independent of the diversity in the molecular species composition of sphingomyelin within a given cell. Diversity in the molecular species composition of sphingomyelin may have direct consequences for the bioactivity of the product of its degradation, ceramide. In this respect, the molecular species composition of ceramide has been shown to differentially stimulate protein phosphatase 2A activity (58,59).
NGF induced sphingomyelin hydrolysis in two cell lines which express solely p75 NTR , i.e. p75 NTR -NIH-3T3 cells and rat T9 glioma cells (33). However, in PC12 cells, which co-express Trk A and p75 NTR , NGF had little effect on sphingomyelin hydrolysis. On the other hand, NT-3 induced significant changes in sphingomyelin levels over a time course very similar to that seen in NT-3-treated p75 NTR -NIH-3T3 cells. Interestingly, although NT-3 induced sphingomyelin hydrolysis in 2 R. T. Dobrowsky, unpublished results.
FIG. 8. The EGFR and platelet-derived growth factor-receptor do not modulate NT-3-induced sphingomyelin hydrolysis. PC12 cells were labeled with [ 3 H]choline chloride for 3 days and placed in DMEM containing 3% serum for 4 h. The cells were transferred to a 37°C water bath and treated for 15 min with PBS, 10 ng/ml EGF, or 10 ng/ml platelet-derived growth factor in the presence or absence of 10 ng/ml NT-3. The lipids were extracted, and sphingomyelin was quantitated. Sphingomyelin levels were normalized to total phosphatidylcholine counts/min, and the results were expressed as the percent of time matched control values. Results shown are mean Ϯ S.E. from two experiments (n ϭ 6). PC12 cells, it has no effect on tyrosine phosphorylation and does not induce differentiation in these cells (49); suggesting that NT-3-induced sphingomyelin signaling is not sufficient for differentiation of PC12 cells. Moreover, the lack of sphingomyelin hydrolysis by NGF supports that sphingomyelin signaling is not involved in NGF-induced differentiation of PC12 cells.
The biologic role of NT-3-induced sphingomyelin metabolism in PC12 cells remains to be determined. It is possible that the product of sphingomyelin hydrolysis, ceramide, may initiate a pathway that inhibits NGF-induced neurite development in PC12 cells. In this respect, treatment of PC12 cells with either bacterial sphingomyelinase (60) or exogenous ceramide 2 inhibits NGF-induced neurite outgrowth, although we were unable to observe significant inhibition of NGF-induced neurite outgrowth in the presence of 10 ng/ml NT-3. However, since the presence of NGF inhibits NT-3-induced sphingomyelin hydrolysis in PC12 cells (Fig. 7), NT-3 may be unable to generate the bioactive mediator ceramide. Ceramide may attenuate NGF-induced neurite outgrowth through modulation of cellular protein phosphorylation/dephosphorylation cascades (58,59,61,62).
Recent studies have demonstrated that p75 NTR can modulate and/or potentiate the activity of Trk receptors. For example, overexpression of p75 NTR relative to Trk enhances tyrosine autophosphorylation of Trk and increases neuronal maturation in MAH cells (32). Furthermore, p75 NTR was found to potentiate masked autocrine loops in transfected fibrolasts (31). Our study suggests that the reciprocal situation also exists, i.e. that Trk tyrosine kinase can specifically modulate p75 NTR signaling. This regulation at least requires the tyrosine kinase activity of Trk, but it may also encompass other structural features of the Trk receptor.
At this point, we can not determine if Trk is directly or indirectly affecting p75 NTR -dependent sphingomyelin hydrolysis. p75 NTR is not significantly phosphorylated on tyrosine following NGF treatment of PC12 cells, suggesting that direct phosphorylation of p75 NTR by Trk is an unlikely point of regulation. It is possible that tyrosine phosphorylation/dephosphorylation of the signal-activated neutral sphingomyelinase inhibits its activity. However, other then cation requirements, little is known about factors regulating this enzyme (63). Alternatively, Trk may not inhibit NGF-induced sphingomyelin hydrolysis but stimulate rapid sphingomyelin resynthesis leading to no net change in overall sphingomyelin levels. However, that NT-3 induces sphingomyelin hydrolysis in PC12 cells suggests that Trk does not constitutively inhibit sphingomyelinase or stimulate sphingomyelin resynthesis. Finally, Trk may indirectly modulate p75 NTR -dependent signaling via coupling to immediate downstream effectors such as phosphatidylinositol 3-kinase, phospholipase C-␥, or proteins in the mitogen-activated or stress-activated protein kinase cascades (23). Regardless, our data provide biochemical evidence that neurotrophin receptors may undergo extensive cross-talk in coordinating signal transduction and support the notion that p75 NTR has a functional signaling capacity.