Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Ceramide and other sphingolipids have signaling roles in controlling proliferation, differentiation, and apoptosis in a wide variety of cell types, including keratinocytes. Specifically, in both primary human and mouse keratinocytes and a squamous carcinoma cell line, synthetic ceramides induce growth arrest and increase the markers of keratinocyte differentiation: transglutaminase activity and cornified envelope formation (Wakita et al., 1994; Pillai et al., 1996; Jung et al., 1998). However, studies using synthetic ceramides are limited due to the poor solubility of these compounds in aqueous solutions, their high degree of toxicity, and the differences in their acyl chain length versus that of natural ceramide. Natural ceramide may be generated intracellularly by specific agonists of sphingomyelinase; however, in keratinocytes few agonists of sphingomyelinase have been identified.

These limitations have led to alternative techniques to artificially increase or decrease ceramide and other sphingolipids within cells in culture. Of these techniques, modulation of sphingolipid biosynthesis with synthetic ceramide analogs is currently increasing in use. For example, D-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol (PDMP) inhibits glucosylceramide synthase activity (Fig. 1), resulting in the depletion of glycosphingolipids and, in some cells, the accumulation of ceramide (Inokuchi and Radin, 1987; Uemura et al., 1990; Shayman et al., 1991; Rani et al., 1995). These changes have been associated with decreased [³H]thymidine incorporation into DNA, decreased cell number, and decreased cell-substratum adherence (Inokuchi et al., 1989; Shayman et al., 1991; Barbour et al., 1992; Kan and Kolesnick, 1992; Rani et al., 1995). Additionally, PDMP is chemotherapeutic in a mouse model of cancer and is antimetastatic in an in vitro model of metastasis (Inokuchi et al., 1987, 1990).

View larger version (11K): [in this window] [in a new window]   Fig. 1.   The site of action of PDMP on sphingolipid biosynthesis.

A recent study examined the effects of PDMP on cultured human keratinocytes (Takami et al., 1998). In these cells, PDMP was shown to decrease glucosylceramide levels, cell number, and [³H]thymidine incorporation into DNA but not to alter ceramide levels. The goal of our current research using primary cultures of mouse epidermal keratinocytes was to examine the ability of the inhibitor of glucosylceramide synthase PDMP to increase ceramide levels and induce growth arrest. As anticipated, our studies revealed that PDMP increased ceramide levels and decreased the incorporation of [³H]thymidine into DNA. However, surprisingly, the decrease in [³H]thymidine incorporation was not due to growth arrest of the keratinocytes but was the result of an inhibition of nucleoside transport by PDMP. Thus, our observations suggest that data obtained using this compound must be interpreted with caution.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Calcium-free minimal essential medium and antibiotics were obtained from Biologos, Inc. (Maperville, IL). Bovine pituitary extract and epidermal growth factor were purchased from Life Technologies (Grand Island, NY). ITS+ (6.25 µg/ml insulin, 6.25 µg/ml transferrin, 6.25 ng/ml selenous acid, 5.35 µg/ml linoleic acid, and 0.125% BSA) was supplied by Collaborative Biomedical Products (Bedford, MA). PDMP was obtained from BIOMOL Research Laboratories (Plymouth Meeting, PA). Sphingomyelinases purified from Bacillus cereus and Streptomyces species were supplied by Sigma Chemical Co. (St. Louis, MO). [-³²P]ATP was purchased from NEN Life Science Products (Boston, MA). Silica gel 60 TLC plates with concentrating zone were obtained from EM Science (Gibbstown, NJ).

Cell Isolation and Culture. Epidermal keratinocytes harvested from neonatal ICR mice according to the method of Yuspa and Harris (1974) were used to initiate primary cultures. Cells were plated at a density of 25,000 cells/cm² and were allowed to attach overnight in the presence of keratinocyte plating media composed of calcium-free minimum essential medium supplemented with 2% dialyzed fetal bovine serum, 25 µM calcium chloride, 5 ng/ml epidermal growth factor, 2 mM glutamine, ITS+, and PSF (100 U/ml penicillin, 100 µg/ml streptomycin, and 0.25 µg/ml fungizone). After attachment, the cells were refed with keratinocyte growth media composed of the above medium in which serum was deleted and 90 µg/ml bovine pituitary extract was added.

Measurement of Cellular Ceramide Levels. Near-confluent cultures of primary mouse keratinocytes were refed with keratinocyte growth media supplemented with the indicated concentrations of PDMP or control vehicle. After the indicated time periods, the medium was aspirated and the cells were lysed with 0.2% SDS. Lipids were extracted from the SDS lysates according to the method of Bligh and Dyer (1959). Ceramide levels were assessed using the DAG kinase method according to Preiss et al. (1986) (reviewed by Bollag and Griner, 1998), and ³²P-labeled ceramide-1-phosphate spots were quantified using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Data were normalized to total phospholipid phosphate content for each sample. Total phospholipid phosphate was measured according to the method of Van Veldhoven and Mannaerts (1987).

Measurement of [³H]Thymidine Transport and Incorporation into DNA. Near-confluent cultures of primary keratinocytes were refed with keratinocyte growth media supplemented with the indicated concentrations of PDMP or control vehicle. The medium was supplemented with radiolabel during the final 1 h of exposure to either PDMP or vehicle, and at the indicated times, the cells were rinsed twice with ice-cold PBS followed by two rinses for 5 min each with ice-cold 5% trichloroacetic acid (TCA). Collection of the TCA wash and quantification by liquid scintillation spectrometry provided the amount of acid-soluble [³H]thymidine levels as shown in Fig. 5 without interference from the [³H]thymidine incorporated into DNA (acid-insoluble). To quantify the incorporation of [³H]thymidine into DNA as shown in Fig. 3, the cells were exposed to radiolabel, rinsed with PBS, and rinsed with TCA as described above. The cells were then rinsed with deionized water and solubilized in 0.3 M NaOH. Incorporation of [³H]thymidine into DNA was determined by counting an aliquot of the NaOH extract in a liquid scintillation spectrometer. To quantify the rate of nucleoside transport over a short time course as shown in Fig. 4, the medium was also supplemented with 1 µCi/ml [³H]thymidine, and at the indicated times, the cells were rinsed three times with ice-cold PBS and then solubilized in 0.3 M NaOH. Cellular transport of [³H]thymidine was determined by counting an aliquot of this NaOH extract in a liquid scintillation spectrometer.

Flow Cytometric Analysis. Near-confluent cultures of keratinocytes were refed with keratinocyte growth media supplemented with 40 µM PDMP, 5 ng/ml transforming growth factor-, or vehicle control. After 24 h, the media as well as one PBS wash (to obtain floating cells) were collected. The attached cells were then harvested with trypsin and pooled with the cells that had become detached. The combined cell pellet was washed once with ice-cold PBS, gently resuspended in ice-cold 70% ethanol, and stored at 4°C until processing. Nuclear DNA content was determined by staining with propidium iodide (50 µg/ml, 30 min) and analyzing the cells with a Becton Dickinson FACScalibur fluorescence-activated cell sorter. ModFit DNA and CellQuest analysis software were used to determine the proportion of nuclei in each phase of the cell cycle.

Statistics. The data are presented as means ± S.E. Each keratinocyte preparation represented a separate experiment (n = 1). Data were analyzed by ANOVA to determine whether individual time points or treatments differed significantly from control. The level of significance was a P value less than .05.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Initial studies sought to determine the ability of PDMP to increase intracellular ceramide levels. Therefore, cellular ceramide levels were measured after a 24-h exposure of primary mouse keratinocytes to the indicated concentrations of PDMP (Fig. 2A). PDMP caused a dose-dependent increase in ceramide levels (Fig. 2A), and as little as 20 µM PDMP was required to significantly elevate ceramide levels (136 ± 4% of control). To determine how rapidly ceramide levels increased in response to PDMP, keratinocytes were exposed to 40 µM PDMP for the indicated times (Fig. 2B). Significant increases in ceramide were observed in response to an exposure to 40 µM PDMP of only 15 min (190 ± 30% of control).

View larger version (12K): [in this window] [in a new window]   Fig. 2.   The concentration-dependent (A) and time-dependent (B) effects of PDMP on ceramide levels in primary mouse keratinocytes. Near-confluent cultures of primary mouse keratinocytes were treated with the indicated concentrations of PDMP or control vehicle for 24 h (A) or with 40 µM PDMP for the indicated times (B), and lipids were extracted and ceramide was measured as described under Experimental Procedures. Data are expressed as a percentage of control. Asterisks indicate that ceramide levels are significantly greater than control (means ± S.E., n = 6-11, *P < .05 versus control).

Because PDMP produced such rapid and significant increases in ceramide levels, we determined the ability of PDMP to induce growth arrest of primary cultures of mouse keratinocytes by measuring the incorporation of [³H]thymidine into the acid-insoluble fraction, indicative of DNA synthesis. PDMP caused a dose-dependent inhibition of [³H]thymidine incorporation into DNA after 24 h of exposure (Fig. 3A), and significant inhibition was achieved in the presence of only 10 µM PDMP (85.5 ± 3% of control). To determine how rapidly [³H]thymidine incorporation decreased in response to PDMP, keratinocytes were exposed to 40 µM PDMP for the indicated times (Fig. 3B). A significant decrease in [³H]thymidine incorporation was observed in response to 40 µM PDMP after only a 60-min exposure (40.1 ± 5.6% of control).

View larger version (11K): [in this window] [in a new window]   Fig. 3.   The concentration-dependent (A) and time-dependent (B) effects of PDMP on [3H]thymidine incorporation into DNA of primary mouse keratinocytes. Near-confluent cultures of primary mouse keratinocytes were treated with the indicated concentrations of PDMP or control vehicle for 24 h (A) or with 40 µM PDMP for the indicated times (B), and [3H]thymidine was present during the final hour. The incorporation of radiolabel into DNA was measured as described under Experimental Procedures, and data are expressed as a percentage of control. All treatments were significantly different from control (means ± S.E., n = 3-5, *P < .05 versus control).

The very rapid onset of inhibition of [³H]thymidine incorporation by PDMP is very atypical compared with inhibition of [³H]thymidine incorporation induced by keratinocyte growth inhibitors such as 1,25-dihydroxyvitamin D₃ (Griner et al., 1999). This strongly suggested that this effect was unrelated to growth arrest of the keratinocytes. In fact, initial experiments confirmed that acid-soluble [³H]thymidine levels were reduced in PDMP-treated keratinocytes (data not shown). Therefore, we investigated the ability of PDMP to inhibit the rate of transport of [³H]thymidine by primary mouse keratinocytes. [³H]Thymidine transport into primary mouse keratinocytes was measured over a range of time points from 30 s to 15 min after the addition of the radiolabel and 40 µM PDMP or vehicle control (Fig. 4). The rate of transport during this range of time points was found to be very linear with a mean correlation coefficient of 0.997. Therefore, a comparison of the slope of each curve of PDMP-treated cells versus control cells allows quantification of the inhibition of the rate of transport. Thus, the rate of transport of [³H]thymidine by 40 µM PDMP-treated cells was only 36.7 ± 3% of control cells. Pretreatment with PDMP for 1 h before the addition of radiolabel decreased the rate of [³H]thymidine transport to a statistically equivalent degree (33.1 ± 0.8% of control).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   The effects of PDMP on [3H]thymidine transport by primary mouse keratinocytes. Near-confluent cultures of primary mouse keratinocytes were treated with 40 µM PDMP or control vehicle and [3H]thymidine for the indicated times. The uptake of radiolabel was measured as described under Experimental Procedures, and data are expressed as disintegrations per minute. , control. , PDMP. All treatments are significantly less than control. Shown is one experiment representative of three.

To examine the mechanism by which PDMP inhibits [³H]thymidine transport, it was necessary to establish the routes through which [³H]thymidine gains entry into keratinocytes. Specific plasma membrane nucleoside transporters have been characterized well (Griffith and Jarvis, 1996), and nitrobenzylthioinosine (NBTI) is a specific inhibitor for at least one class of nucleoside transporter (Cass et al., 1974; Cass and Paterson, 1976). Therefore, the ability of NBTI to inhibit [³H]thymidine transport in primary mouse keratinocytes was measured over a range of time points from 30 s to 15 min after the addition of the radiolabel and 100 µM NBTI or vehicle control. The rate of transport during this range of time points was found to be very linear with a mean correlation coefficient of 0.993 (n = 4), and the rate of transport, as determined by the slope of the curve, was only 8.9 ± 2.7% of control cells (mean ± S.E., n = 4, P < .05). Furthermore, after a 1-h exposure to 100 µM NBTI, acid-soluble [³H]thymidine levels were similarly decreased to 8.1 ± 0.48% of control cells (mean ± S.E., n = 6, P < .05). Thus, at least 91% of [³H]thymidine uptake occurs through this specific transport mechanism in keratinocytes. Therefore, the effects of PDMP on [³H]thymidine transport are apparently due to a direct or indirect effect on the plasma membrane nucleoside transporter.

The rapid increase in ceramide levels in response to PDMP suggests that the effects of PDMP on nucleoside transport could be mediated by ceramide. However, compelling evidence against a role for ceramide in inhibition of [³H]thymidine transport was obtained by incubating keratinocytes with 0.1 U/ml bacterial sphingomyelinase for 1 h. Ceramide levels were increased greater than 10-fold in the presence of sphingomyelinase purified from B. cereus or from S. species. (Fig. 5). However, acid-soluble [³H]thymidine levels in keratinocytes were inhibited less than 20% by each enzyme. Thus, the dramatic increase in ceramide levels produced by the bacterial sphingomyelinases resulted in only a slight inhibition of [³H]thymidine uptake compared with PDMP.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   The effects of exogenous sphingomyelinase on ceramide levels and [3H]thymidine uptake. Near-confluent cultures of primary mouse keratinocytes were treated with 0.1 U/ml bacterial sphingomyelinase or control for 1 h. Ceramide levels (left y-axis and columns) and [3H]thymidine uptake (right y-axis and columns) were measured as described under Experimental Procedures, and data are expressed as a percentage of control [means ± S.E., n = 4 (ceramide), n = 7 ([3H]thymidine uptake), *P < .05 versus control]. , B. cereus; , S. species.

We have subsequently confirmed that PDMP is not a growth inhibitor for primary mouse keratinocytes through the use of flow cytometric analysis. Table 1 shows that a 24-h exposure to PDMP had no effect on the numbers of keratinocytes in each phase of the cell cycle. However, an established growth inhibitor of keratinocytes, transforming growth factor-, decreased the number of cells in S-phase and increased the number of cells in the resting phase (G₀/G₁), indicative of growth inhibition.

View this table: [in this window] [in a new window]   TABLE 1 Effects of PDMP and TGF- on the cell cycle profile of primary mouse keratinocytes Near-confluent cultures of primary mouse keratinocytes were treated with 40 µM PDMP, 5 ng/ml TGF-, or control vehicle for 24 h and then analyzed by flow cytometry as described under Experimental Procedures. Data are presented as the percentage of cells in each phase of the cell cycle. The small percentage of sub-G0/G1 cells is not included. Values are given as mean ± S.E. (n = 6).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Our findings reveal a rapid inhibition of both cellular uptake and DNA incorporation of [³H]thymidine by PDMP in primary cultures of mouse keratinocytes. Furthermore, experiments revealed a direct relationship between the intracellular levels of [³H]thymidine and its incorporation into DNA (data not shown). Thus, in keratinocytes treated with PDMP, the inhibition of uptake of [³H]thymidine (Fig. 4) is approximately sufficient to account for the inhibition of [³H]thymidine incorporation into DNA (Fig. 3). Although we have investigated this effect in depth in only primary mouse keratinocytes, we have conducted a limited number of experiments with several established cell lines. These results indicate that in HeLa and Madin-Darby canine kidney cells, PDMP decreases [³H]thymidine uptake in a manner similar to that observed in primary mouse keratinocytes (data not shown). Therefore, the ability of PDMP to induce growth arrest cannot accurately be assessed using the standard technique of [³H]thymidine incorporation.

Because PDMP and synthetic ceramide are each capable of inhibiting fluid-phase endocytosis (Chen et al., 1995), the mechanism of inhibition of [³H]thymidine uptake may be through inhibition of this vesicular transport pathway. However, the present results show that in keratinocytes, at least 91% of [³H]thymidine internalization is by a specific nucleoside transport protein. Therefore, the effects of PDMP on [³H]thymidine uptake must be due to inhibition of this specific transport protein and not due to inhibition of nonspecific uptake events.

In the present study, PDMP increased ceramide levels in primary mouse keratinocytes, and ceramide likely serves as a signaling molecule in keratinocyte growth arrest and differentiation (Wakita et al., 1994; Pillai et al., 1996; Jung et al., 1998). Therefore, we wished to determine whether the increase in ceramide levels was responsible for the inhibition of [³H]thymidine transport produced in response to PDMP. Using bacterial sphingomyelinases, we produced a 10-fold increase in ceramide levels, but this resulted in less than a 20% decrease in [³H]thymidine transport. By comparison, PDMP (40 µM) increased ceramide levels by only 90% while simultaneously decreasing [³H]thymidine transport by more than 63%. Thus, the levels of intracellular ceramide have no consistent correlation with the decrease in [³H]thymidine transport. Furthermore, it should be noted that the intracellular sites of production of ceramide by the bacterial sphingomyelinases and those by PDMP are likely different. Specifically, exogenously added sphingomyelinase hydrolyzes sphingomyelin in the plasma membrane to produce ceramide (Chatterjee, 1993), whereas PDMP increases ceramide levels by inhibiting glucosyltransferase, which is located within the Golgi (Coste et al., 1985; Futerman and Pagano, 1991; Trinchera et al., 1991; Paul et al., 1996; Madison and Howard, 1996; Allan and Obradors, 1999). Therefore, PDMP produces ceramide at a location distinct from the site of nucleoside transport: the plasma membrane. Although redistribution of ceramide to all cellular membranes may occur through normal vesicle trafficking, it is unlikely that ceramide is responsible for the inhibition of nucleoside transport.

The mechanism by which PDMP inhibits nucleoside transport remains unclear. The PDMP molecule could interact directly with the transport protein and/or the nucleoside phosphorylating enzymes. Furthermore, the inhibition of glucosylceramide synthase by PDMP could have numerous effects in addition to increasing ceramide. For example, PDMP treatment may result in depletion of glycolipids (Uemura et al., 1990), accumulation of both free sphingosine (Shayman et al., 1991) and methylated sphingosine (Felding-Habermann et al., 1990), and/or increased synthesis of sphingomyelin (Okada et al., 1988; Shayman et al., 1990). Changes in the levels or ratios of these membrane components could alter the physical characteristics of the plasma membrane. With the recent findings that membrane-bound proteins may be regulated based on the physical characteristics of their lipid microenvironment (for reviews, see Anderson, 1998, and Brown and London, 1998), it is not unreasonable to suggest that such changes may alter the activity of the nucleoside transporter. However, changes in these physical characteristics might be expected to occur over minutes to hours after the addition of PDMP, whereas a decrease in the rate of [³H]thymidine transport occurred as rapidly as 30 s after PDMP addition. This rapid onset of effect by PDMP also suggests that toxicity is not responsible for the decrease in [³H]thymidine transport, because the cells continued to live for at least 48 h after the addition of PDMP (data not shown). Thus, a direct effect of PDMP on the nucleoside transport apparatus seems to be the more likely scenario.

Additionally, there are increasing reports of other nonspecific actions of this inhibitor, such as the recent finding that PDMP alters calcium homeostasis and interferes with retrograde membrane transport from the Golgi to the endoplasmic reticulum independent of changes in glycosphingolipids (Kok et al., 1998). Moreover, recent studies have shown that PDMP inhibits not only glucosylceramide synthase but also lactosylceramide synthase and possibly other enzymes involved in ganglioside biosynthesis (Inokuchi et al., 1995; Chatterjee et al., 1996). Therefore, the nonspecificity of these compounds may limit their use as research tools to examine the role of sphingolipids in cellular processes. Recently, "improved inhibitors of glucosylceramide synthase" have been synthesized and partially characterized (Lee et al., 1999), but their effects on [³H]thymidine transport have not been determined. Thus, our results indicate that such inhibitors must be used with caution in any study to examine their effects on growth arrest.