Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The novel antimetabolite gemcitabine is a deoxycytidine analog (2',2'-difluoro-2'-deoxycytidine) that behaves similarly to cytosine arabinoside (ara-C) at the biochemical level but has the advantages of greater potency and slower clearance from tumor cells and from the body (Heinemann et al., 1988; Hui and Reitz, 1997). Compared with other antimetabolites, which are most active against leukemias, gemcitabine is unusual in that it displays cytotoxic activity against solid tumors. Specifically, gemcitabine has shown promise against head and neck, ovarian, pancreatic, non-small cell lung, and breast carcinomas that were unresponsive to other chemotherapeutic agents (Lund et al., 1993; Guchelaar et al., 1996; Hui and Reitz, 1997; Noble and Goa, 1997).

The ultimate cytotoxic action of gemcitabine involves inhibition of DNA synthesis by incorporation of 2',2'-difluoro-2'-deoxycytidine triphosphate (dFdCTP) (Ross and Cuddy, 1994; Plunkett et al., 1995). Factors that can affect the level of DNA incorporation, and hence cytotoxic efficacy, include the rate of intracellular accumulation, phosphorylation, and deamination of gemcitabine and its nucleotide derivatives, as well as competition from endogenous substrates and changes in polymerase/exonuclease activities. A change in any of these determinants could confer tumor cell resistance or enhanced drug sensitivity depending on the direction of change. The present study focused solely on the first step in the metabolism of gemcitabine, namely the mechanisms by which cells transport the drug across the plasma membrane. Changes in membrane transport processes have been shown to contribute to the cellular resistance to nucleoside antimetabolites such as ara-C (White et al., 1987; Cass, 1995b), and we hypothesized that these same transporters are involved in the cellular uptake of gemcitabine.

Mammalian cells express two subtypes of equilibrative nucleoside transporters that can be delineated by their sensitivity to inhibition by nitrobenzylthioinosine (NBMPR) and at least three distinct Na⁺-dependent concentrative transporters with differing substrate specificities (Belt et al., 1993; Cass, 1995a; Griffith and Jarvis, 1996). There is evidence that gemcitabine can interact with the substrate binding site of at least a subset of these transporters. High concentrations of gemcitabine (>1 mM) have been reported to inhibit [³H]uridine uptake by recombinant human es (equilibrative, inhibitor-sensitive; hENT1) transporters expressed in Xenopus oocytes (Griffiths et al., 1997) and a recombinant pyrimidine-selective concentrative transporter (rCNT1) expressed in COS-1 cells (Fang et al., 1996). We have also shown recently that gemcitabine can inhibit the uptake of [³H]formycin B by the both the es and ei (equilibrative, inhibitor-insensitive) transporters of mouse Ehrlich ascites tumor cells and appears to have a relatively higher affinity for the es transporter subtype (Burke et al., 1998). Although the aforementioned results indicate that gemcitabine can act as an inhibitor of nucleoside transporters, a recent preliminary meeting report (Mackey et al., 1998) suggests that gemcitabine may also serve as a substrate for these systems. This is supported by the finding that the nucleoside transport inhibitor dipyridamole and its congener BIBW22BS can inhibit the antiproliferative activity of gemcitabine in various cancer cell lines (Jansen et al., 1995).

The model system used for the present study, a head and neck squamous carcinoma cell line (designated HN-5a), is representative of a class of tumors shown to be responsive to gemcitabine therapy in preclinical trials (Braakhuis et al., 1991). Because the nucleoside transporter phenotype of this cell line had not been established previously, initial studies involved an assessment of the transporter subtypes mediating the uptake of the well defined permeants [³H]uridine and [³H]formycin B (Plagemann and Woffendin, 1989; Cass, 1995a). Parallel studies were conducted using a gemcitabine-resistant variant of this cell line (GEM-8e) to determine whether changes in transporter activities contributed to the observed resistance. Once the transporter phenotypes were defined, the capacity of these systems to mediate the cellular uptake of [³H]gemcitabine was assessed.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. [G-³H]Formycin B (14 Ci/mmol) and [5,6-³H]uridine (35-50 Ci/mmol) were purchased from Moravek Biochemicals (Brea, CA) and ICN Biomedicals, Inc. (Montreal, Quebec, Canada), respectively. [³H]Water (1 mCi/g) and [carboxyl-¹⁴C]dextran-carboxyl (0.58 mCi/g) were purchased from DuPont Canada Inc. (Markham, Ontario, Canada). Gemcitabine and [³H]gemcitabine (22 Ci/mmol) were generously provided by Eli Lilly Inc. (Scarborough, Ontario, Canada). Nonradiolabeled formycin B, uridine, NBMPR, and dipyridamole [2,6-bis(diethanolamino)-4,8-dipiperidinopyrimido-[5,4-day]pyrimidine] were supplied by Sigma Chemical Co. (St. Louis, MO). Fetal bovine serum (FBS) and Dulbecco's modified Eagle's medium (DMEM) was from GIBCO BRL (Burlington, Ontario, Canada). Penicillin G and streptomycin sulfate were purchased from ICN (Montreal, Quebec, Canada). All other compounds were of reagent grade.

Cell Culture. The head and neck squamous carcinoma cell line used in this study (HN-5a) is one of several clonal cell lines established at the St. Joseph's Health Center (London, Canada) (Lapointe et al., 1992) from patients who had not received any prior treatment. Cells were cultured in DMEM plus 10% FBS supplemented with penicillin G (100 units/ml) and streptomycin sulfate (100 µg/ml) and were maintained in a humidified atmosphere of 5% CO₂ at 37°C. Gemcitabine-resistant cell lines were selected by culturing HN-5a cells in the presence of gemcitabine (5-10 nM) for 3 weeks with the drug-containing medium changed weekly. For cytotoxicity assays, rapidly proliferating cells (10⁵ cells/25-cm² flask) were exposed to varying concentrations of drug for 4 days. Cell numbers were determined using an electronic particle counter (Coulter Electronics, Hialeah, FL) at the beginning and end of the drug-exposure period.

[³H]Nucleoside Uptake Assays. Preliminary studies showed that the activity of the nucleoside transporters in these cells declined significantly on reaching confluence (data not shown). Therefore, all transport studies were conducted using cells grown to 75% of confluence in 175-cm² flasks. Cells were removed from the flasks by trypsinization (0.5% for 15 min at 37°C) and then diluted 6-fold with DMEM plus 10% FBS and pelleted by centrifugation (0.3 ml packed cell volume). The cell pellets were washed once by resuspension/centrifugation in either normal Dulbecco's PBS or a modified Na⁺-free PBS (iso-osmotic replacement by Li⁺) and then resuspended in 13 ml of the same buffer for use in the uptake assays. It has been shown in a variety of systems that Li⁺ is unable to substitute for Na⁺ at Na⁺-dependent nucleoside transporters, and hence differences in [³H]nucleoside uptake observed in Na⁺ versus Li⁺ medium have generally been taken to represent the operation of Na⁺-dependent nucleoside transporters (Spector and Huntoon, 1984; Jarvis, 1989; Plagemann and Aran, 1990; Dagnino et al., 1991; Williams and Jarvis, 1991; Baer et al., 1992; Doherty and Jarvis, 1993). In some cases (as indicated in Results), cells were depleted of ATP by sequential incubation with rotenone (20 ng/ml for 15 min at 37°C) and 2-deoxyglucose (2 mM for 15 min at 37°C). This procedure has been shown to reduce cellular ATP content by 95%, which is sufficient to prevent [³H]uridine metabolism over the time course of these studies (Hammond and Johnstone, 1989; Hammond, 1991). All uptake assays were conducted at room temperature (22°C). Uptake was initiated by the addition of cell suspension (1 × 10⁶ cells) to [³H]substrate layered over a 200-µl cushion of silicone oil/mineral oil (21:4 v/v) in 1.5-ml microcentrifuge tubes. Assays were terminated after a defined incubation time (minimum of 3 s) through centrifugation of cells through the oil for 10 s at 12,000g. The supernatant and oil were removed, and the cell pellets were digested with 1 M sodium hydroxide for 16 h at room temperature. The digest was analyzed for ³H content by standard liquid scintillation counting techniques. The estimated time required to pellet the cells through the oil layer (2 s) is included in all reported incubation times.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Establishment of Gemcitabine-Resistant Cell Lines. HN-5a cells cultured in 5 or 6 nM gemcitabine did not form colonies but were overgrown. No colonies grew in cultures exposed to 10 nM gemcitabine. However, exposure of cultures to 7, 8, and 9 nM gemcitabine resulted in the isolation of several stable gemcitabine-resistant cell lines. One of these cell lines, designated GEM-8e (selected using 8 nM gemcitabine), was further characterized and found to have an IC₅₀ value for gemcitabine (15.9 ± 1.6 nM) that was 6.03 ± 0.88 times higher (n = 3) than that determined for the parent cell line (2.66 ± 0.21 nM).

[³H]Uridine Uptake. Initial studies were conducted to establish the types of nucleoside transporters expressed by the HN-5a cells. These experiments used the well established permeant [³H]uridine, a substrate for all of the known classes of nucleoside transporter, and were conducted using ATP-depleted cells to prevent the intracellular trapping of [³H]uridine as its phosphorylated derivatives. The time course of the uptake of 10 µM [³H]uridine by HN-5a cells was best described by an hyperbolic relationship and was mediated by a process that was sensitive to inhibition by dipyridamole and NBMPR (Fig. 1A). The initial rate of transporter-mediated influx was 0.50 ± 0.05 pmol·µl¹·s¹, and steady state was achieved at an intracellular concentration of 12.3 ± 1.0 µM. Preincubation of the cells for 30 min with a range of concentrations of NBMPR before assessment of the uptake of 10 µM [³H]uridine (20 s) resulted in the inhibition profile shown in Fig. 1B. Most of the transporter-mediated uptake was sensitive to inhibition by NBMPR with an IC₅₀ value of 0.41 nM. However, a small component (5.2%) of the dipyridamole-sensitive uptake was resistant to inhibition by NBMPR even at concentrations in excess of 1 µM (Fig. 1B). Gemcitabine inhibited the uptake of [³H]uridine by both HN-5a and GEM-8e cells (ATP-depleted, normal Na⁺) in a concentration-dependent manner with IC₅₀ values of 231 and 330 µM, respectively (Fig. 2).

View larger version (23K): [in this window] [in a new window]   Fig. 1.   [3H]Uridine accumulation by ATP-depleted HN-5a cells. A, time course of uptake. Cells were incubated with 10 µM [3H]uridine in the absence (; total uptake) and presence (; nonmediated uptake) of 10 µM dipyridamole/NBMPR for times indicated, and assays were terminated as described in text. Transporter-mediated uptake () was calculated in each experiment as difference between total and nonmediated uptake curves. Each point represents mean ± S.E. from four experiments. B, inhibition of [3H]uridine uptake by NBMPR. ATP-depleted HN-5a cells were incubated with indicated concentrations of NBMPR for 15 min and then exposed to [3H]uridine (10 µM) for 20 s. Results are presented as a percentage of transporter-mediated accumulation observed in the absence of NBMPR (control). Dotted line shows relative amount of NBMPR-resistant uptake of [3H]uridine. IC50 value for NBMPR inhibition of uptake, with its 95% confidence interval, was calculated from the sigmoid curve, as shown, fitted to average data from three independent experiments.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Inhibition of [3H]uridine uptake by gemcitabine. ATP-depleted HN-5a () and GEM-8e () cells were exposed concurrently to 10 µM [3H]uridine and indicated concentrations of gemcitabine for 20 s and then terminated by oil-stop method described in text. Results are presented as a percentage of transporter-mediated accumulation observed in the absence of gemcitabine (control). Each point represents the mean ± S.E. from four experiments. IC50 values for gemcitabine inhibition of [3H]uridine uptake, along with their 95% confidence intervals, were calculated from the sigmoid curves shown. *Significant difference between HN-5a and GEM-8e uptake data (Student's t test, p < .05).

[³H]Formycin B Uptake. To assess the potential contribution of Na⁺-dependent transport processes to the uptake of nucleosides by HN-5a and GEM-8e cells, the uptake of the poorly metabolized (Plagemann and Woffendin, 1989) inosine analog [³H]formycin B (10 µM) was studied in both the presence and absence of Na⁺ in ATP-replete (normal) cells (Table 1, Fig. 3). [³H]Formycin B has been established as a permeant for both forms of equilibrative nucleoside transporter (es, ei) as well as the cif (purine-selective) and cib (nonselective) subtypes of concentrative transporter (Cass, 1995a; Griffith and Jarvis, 1996). The time course of transporter-mediated [³H]formycin B uptake by HN-5a cells in the absence of Na⁺ was saturable (steady state achieved at 14.3 µM intracellular concentration) with an initial rate of influx of 0.53 pmol·µl¹·s¹. These results are similar to those obtained using [³H]uridine as the substrate (see above). When cells were preincubated with 100 nM NBMPR to block uptake by the es transporter, the initial rate of [³H]formycin B influx was reduced by 90% to 0.05 pmol·µl¹·s¹. In both cell lines, the accumulation of [³H]formycin B in the presence of Na⁺ (±100 nM NBMPR) was similar to that seen in its absence. The only effect of Na⁺ was a slight increase in the initial rate of influx of [³H]formycin B by the HN-5a cells (Table 1). This may indicate the presence of a small component of Na⁺-dependent nucleoside transporter-mediated activity. However, Na⁺ had no effect when similar assays were conducted in the presence of 100 nM NBMPR, a condition that should enhance the detection of concentrative transporter activity by blocking [³H]formycin B efflux via the major equilibrative transporter (es). Therefore, this minor effect of Na⁺ on the initial rate of influx was probably due to factors, as yet undefined, unrelated to the activity of concentrative nucleoside transporters.

View this table: [in this window] [in a new window]   TABLE 1 Transporter-mediated uptake of [3H]formycin B by HN-5a and GEM-8e cells Cells were incubated with 10 µM [3H]formycin B in presence and absence of 100 nM NBMPR or 10 µM NBMPR/dipyridamole for various times as shown in Fig. 3. Cells were equilibrated in either normal Dulbecco's PBS (+Na) or in an Na+-free buffer (Na; iso-osmotic replacement with Li+). Total transporter-mediated uptake (Total) was defined as that sensitive to inhibition by 10 µM NBMPR/dipyridamole. NBMPR-sensitive uptake (es) was calculated as difference between uptake in presence and absence of 100 nM NBMPR, whereas NBMPR-resistant uptake (ei) was defined as difference between uptake in presence of 100 nM NBMPR and that observed in the presence of 10 µM NBMPR/dipyridamole. Hyperbolic curves were fitted to resulting time course data and extrapolated to obtain estimates of initial rate of influx (Vi) and, for total and es-mediated components, steady-state intracellular concentrations of [3H]formycin B (Max); for ei-mediated uptake, which did not approach steady state over time course of these assays, intracellular concentration after 5-min incubation is shown. Each value represents mean ± S.E. of four independent experiments.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Time courses of [3H]formycin B accumulation by human head and neck squamous carcinoma cells. ATP-replete HN-5a ( and ) and GEM-8e ( and ) cells were incubated with 10 µM [3H]formycin B in the absence (total uptake) and presence of 100 nM NBMPR (NBMPR-resistant uptake) or 10 µM dipyridamole/NBMPR (to inhibit all equilibrative transporter activity; nonmediated), for times indicated. All uptake assays were conducted in Dulbecco's PBS buffer (normal Na+) and terminated by oil-stop method described in text. Each point represents mean ± S.E. from four experiments.

[³H]Gemcitabine Uptake. Concentrations of gemcitabine of <3 µM had no significant effect on the uptake of [³H]uridine in either cell line (Fig. 2). This indicated that if gemcitabine was a substrate for nucleoside transporters in these cells, it likely had a K_m value of >3 µM. Therefore, the cellular uptake of 10 µM [³H]gemcitabine was measured using both ATP-replete and ATP-depleted cells in the presence and absence of Na⁺. In ATP-depleted cells in the absence of Na⁺, the total uptake time course fit best to an hyperbolic relationship with initial rates of 0.48 ± 0.03 and 0.57 ± 0.06 pmol/µl/s in the HN-5a and GEM-8e cells, respectively (Fig. 4, Table 2). These rates are not significantly different from each other and are similar to the initial rates obtained for both [³H]uridine and [³H]formycin B uptake by these cells (see above). In addition, as expected under these ATP/Na⁺-depleted conditions, the intracellular concentration of [³H]gemcitabine never exceeded the extracellular medium concentration of 10 µM (Table 2). Preincubation of ATP/Na⁺-depleted cells with 100 nM NBMPR, to selectively block uptake via the es transporters, resulted in a near-complete inhibition of transporter-mediated uptake (Fig. 4), suggesting that the ei transporter plays only a minor role in the cellular uptake of [³H]gemcitabine by these cells (compare the "ei + es-mediated uptake" and "es-mediated uptake" columns in Table 2). When all equilibrative nucleoside transporters were blocked with a combination of NBMPR and dipyridamole, a significant amount of cell-associated [³H]gemcitabine was still observed even at 5-s incubation (1.5 pmol/µl intracellular water), the shortest time point measured (Fig. 4). This contrasts with the results obtained using [³H]formycin B where there was essentially no cell associated radiolabel at 5 s in the presence of NBMPR and dipyridamole (Fig. 3). Nucleoside transporter-mediated uptake of [³H]gemcitabine (es + ei mediated), calculated as the total accumulation minus that observed in the presence 10 µM NBMPR and dipyridamole ("nonmediated"), was not significantly different between the two cell lines in terms of either the steady-state accumulation or the initial rate of influx (Fig. 5, Table 2).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Time courses of [3H]gemcitabine accumulation by ATP-depleted HN-5a (, , and ) and GEM-8e cells (, , and ). All assays were conducted in Na+-free PBS (iso-osmotic replacement with Li+). Cells were incubated with 10 µM [3H]gemcitabine in the absence (total uptake) and presence of 100 nM NBMPR (NBMPR-resistant uptake) or 10 µM dipyridamole/NBMPR (NBMPR/DY resistant; non-transporter-mediated influx) for times indicated. Uptake assays were terminated by centrifugation of cells through an oil layer as described in text. Dotted line represents nonmediated uptake of 10 µM formycin B by HN-5a cells (from Fig. 3), shown for comparison. Each point represents mean ± S.E. from four experiments.

View this table: [in this window] [in a new window]   TABLE 2 Uptake of [3H]gemcitabine by HN-5a and GEM-8e cells ATP-depleted (ATP) and ATP-replete (+ATP) cells were equilibrated in either normal Dulbecco's PBS (+Na) or in an Na+-free PBS buffer (Na; iso-osmotic replacement with Li+) and then incubated with 10 µM [3H]gemcitabine in absence (total cell associated) and presence of 100 nM NBMPR or 10 µM NBMPR/dipyridamole for various times as shown in Fig. 4. Equilibrative transporter (ei + es)-mediated uptake was calculated as total cell associated [3H]gemcitabine minus that observed in presence of 10 µM NBMPR/dipyridamole. NBMPR-sensitive (es) transporter-mediated uptake was calculated as difference in cell accumulation of [3H]gemcitabine in presence and absence of 100 nM NBMPR. Hyperbolic curves were fitted to resulting time course data and extrapolated to obtain estimates of steady-state intracellular concentrations of [3H]gemcitabine (Max) and initial rate of influx (Vi). Each value represents mean ± S.E. of four independent experiments.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Effect of cellular ATP depletion on time courses of transporter-mediated uptake of 10 µM [3H]gemcitabine by HN-5a ( and ) and GEM-8e ( and ) cells. Equilibrative transporter-mediated uptake was calculated as difference between total uptake and that observed in the presence of 10 µM NBMPR/dipyridamole, as shown in Fig. 4, for incubation times ranging from 5 s to 5 min. Cells were depleted of ATP as described in text, and all assays were conducted using nominally Na+-free media. Values for steady-state accumulation and initial rates of influx of [3H]gemcitabine, derived by extrapolation from these hyperbolic relationships, are summarized in Table 2. Each point represents mean ± S.E. from four experiments.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

A significant problem in the clinical management of cancers is the development of cellular resistance to the cytotoxic effects of anticancer drugs. Most of these drugs act via intracellular targets, and hence drug resistance may arise through changes in membrane transport processes resulting in a reduced cellular accumulation of drug (Cass, 1995b). This is particularly true for nucleoside analogs that, on being phosphorylated intracellularly, interfere with DNA synthesis. These agents are generally hydrophilic in nature and require a mediated transport system for efficient transfer across the cell membrane. Consequently, the effectiveness of nucleoside analogs in a given clinical situation will depend, to some extent, on the nucleoside transport characteristics of the tumor cells. It has been shown that cultured cells with reduced nucleoside transporter activity, due to either genetic mutations or pharmacological intervention, are resistant to the cytotoxic effects of a variety of nucleoside antimetabolites (White et al., 1987; Belt et al., 1993; Cass, 1995b).

Because gemcitabine is a cytidine analog, it is not unreasonable to expect it to gain entry into the tumor cells via one or more of the plasma membrane-located nucleoside transporters. Therefore, a reduction in nucleoside transporter activity could potentially contribute to the resistance to the cytotoxic effects of gemcitabine observed in the GEM-8e cell line developed in the present study. It would appear from the results obtained, however, that resistance in this line is not due to altered nucleoside transport activity. The GEM-8e cells actually accumulated nucleosides at a slightly greater rate than did the HN-5a parent cell line (Table 1). More than 90% of the uptake in both cell lines was mediated by the es transporter subtype (NBMPR IC₅₀ = 0.41 nM, see Fig. 1B). The remaining influx, in both the parent HN-5a and gemcitabine-resistant GEM-8e cell lines, was mediated by the ei transporter. In general, cells accumulated [³H]uridine and [³H]formycin B to steady-state levels that were 30% higher than the initial medium concentration in both the presence and absence of Na⁺. This non-Na⁺-dependent and non-ATP-dependent concentrative effect has been observed in other cell lines and has been postulated to be due to substrate binding to intracellular components (Plagemann and Woffendin, 1989; Hammond, 1991).

Previous studies have shown that high concentrations of gemcitabine (>1 mM) inhibit nucleoside influx via both es transporters (Griffiths et al., 1997) and Na⁺-dependent concentrative nucleoside transporters (Fang et al., 1996). In addition, we have shown that gemcitabine can inhibit [³H]formycin B uptake by both the es and ei transporters of mouse Ehrlich ascites tumor cells (Burke et al., 1998). The present study showed that gemcitabine can also inhibit [³H]uridine uptake via es transporters of a human cell line with a potency comparable to that observed in mouse Ehrlich cells(Burke et al., 1998). The IC₅₀ value for gemcitabine inhibition of uptake by GEM-8e cells (330 µM) was higher than the IC₅₀ value observed in the parent HN-5a cells, suggesting that it had a lower affinity for the GEM-8e transporters. However, this difference was not sufficient to account for the 6-fold decrease in sensitivity of the GEM-8e cells to the cytotoxic effects of gemcitabine.

Inhibition of nucleoside influx by a compound provides only an indication of the affinity of that compound for the permeant site of the transporter and, as such, does not distinguish between a transport inhibitor and a transporter substrate. Therefore, the capacity of gemcitabine to serve as a substrate for the nucleoside transporters of head and neck squamous carcinoma cells (HN-5A and GEM-8e) was assessed directly through measurement of the influx of [³H]gemcitabine (10 µM) in the presence and absence of Na⁺, NBMPR, and ATP. [³H]Gemcitabine was accumulated by ATP-depleted HN-5A and GEM-8e cells at rates similar to those obtained using [³H]uridine or [³H]formycin B as substrates. However, unlike the latter substrates, a large proportion of the [³H]gemcitabine uptake was resistant to both NBMPR and dipyridamole (30%). This "nonequilibrative-transporter-mediated" uptake was observed in ATP-depleted cells in the absence of Na⁺ and was likely due to passive diffusion of gemcitabine across the cell membrane or nonspecific association with components of the cell membrane. This conclusion is supported by the results of Heinemann et al. (1988), who showed that the octanol/water partition coefficient for gemcitabine was 5-fold higher than that for ara-C. These investigators also showed that Chinese hamster ovary cells accumulated dFdCTP at a significantly greater rate than ara-CTP when cells were incubated for 4 h with the respective nucleosides and that unlike ara-CTP, dFdCTP accumulation was proportional to the gemcitabine concentration over a 100-fold range. Gemcitabine does not appear to be as dependent on nucleoside transport processes for entry into cells as are other nucleoside analogs, such as ara-C.

Cellular uptake studies using metabolizable nucleosides like uridine require that cells be depleted of ATP to prevent the intracellular trapping of the nucleoside as its phosphorylated derivatives (Cass, 1995a; Griffith and Jarvis, 1996). When uptake of 10 µM uridine was measured in HN-5a cells without first depleting the cells of ATP, the apparent steady-state [³H]uridine concentration was in excess of 30 µM, reflecting the activity of uridine kinases (data not shown). The same experiments performed using ATP-depleted cells resulted in an intracellular concentration of 12 µM [³H]uridine. To determine whether intracellular trapping of [³H]gemcitabine metabolites was likely to be a problem over the time course of these experiments, the uptake of [³H]gemcitabine was compared in normal ATP-replete and ATP-depleted cells. Surprisingly, ATP had exactly the opposite effect on [³H]gemcitabine uptake of that expected based on the metabolic rationale described above. ATP-replete cells accumulated [³H]gemcitabine to a significantly lesser extent than did ATP-depleted cells. This phenomenon was evident in both the HN-5a and GEM-8e cell lines and may be indicative of the operation of an ATP-dependent efflux mechanism for [³H]gemcitabine. It should be noted, however, that the parent HN-5a cells had relatively more of this putative efflux activity than did the gemcitabine-resistant GEM-8e cells (Fig. 5), suggesting that this mechanism is not responsible for resistance to gemcitabine in these cells. The best characterized efflux pump for cytotoxic drugs is the multidrug resistance (MDR) transporter P-glycoprotein (Stein, 1997). Preliminary studies from our laboratories indicate that HN-5a cells do express a basel level of MDR activity because the MDR modifier verapamil increased the uptake of [³H]vincristine in HN-5a cells by 60% (P. Ferguson unpublished data). Similar studies with verapamil could not be conducted in combination with [³H]gemcitabine because verapamil has been shown to inhibit nucleoside transporters (Hammond et al., 1985), which would complicate the interpretation of such studies. To our knowledge, there is no direct evidence in the literature to suggest that gemcitabine is a substrate for the MDR system. Indeed, MDR P388 leukemia cells have been shown to exhibit no cross-resistance to gemcitabine (Waud et al., 1996). However, it has been noted that induction of gemcitabine resistance in the ovarian cancer cell line A2780 results in a cross-resistance to the MDR drugs doxorubicin and vincristine, although this was not associated with induction of P-glycoprotein (Ruiz van Haperen et al., 1995).

In summary, [³H]gemcitabine has been established as a substrate for both es and ei nucleoside transporters in the HN-5a cell line and its gemcitabine-resistant variant, GEM-8e. There were no significant differences between these cell lines in terms of their nucleoside transport processes, suggesting that the resistance of the GEM-8e cells to gemcitabine does not involve changes in nucleoside transporter activity. Both the HN-5a and GEM-8e cells express primarily the es subtype of equilibrative nucleoside transporter, with <10% of the total [³H]nucleoside influx mediated by NBMPR-resistant mechanisms. Preliminary data were obtained to suggest that the HN-5a and GEM-8e cells possess an ATP-dependent mechanism for removing gemcitabine from the cells. Further studies are required to define the characteristics of this efflux mechanism, but the presence of such a system would clearly have an affect on the clinical efficacy of gemcitabine.