The polyamines spermidine and spermine, and their diamine precursor, putrescine, are essential constituents of all eukaryotic cells(1, 2, 3) . These polycationic compounds are required for growth-related processes, such as the transcription of specific genes (4) , the unique hypusine modification of the eukaryotic translation initiation factor eIF-5A(5) , and other aspects of macromolecular biosynthesis(6) . Although the enzymatic machinery required for polyamine biosynthesis is expressed in the vast majority of nucleated cells(1, 2, 3) , a high affinity polyamine transport system has also been detected in most, if not all cell types examined(7) .

There is substantial evidence that uptake from extracellular sources plays a quantitatively important role in the homeostasis of polyamine pools in vivo. For instance, tumor-forming ability of L1210 mouse leukemia mutant cells defective in polyamine transport is strongly impaired in hosts treated with -difluoromethylornithine (DFMO), ()a suicide inhibitor of ornithine decarboxylase, unlike parental cells(8, 9) . The main source of circulating polyamines is thought to be the gastrointestinal tract, where the microbial flora generates millimolar concentrations of putrescine and cadaverine which are made available to the plasma compartment through enterohepatic circulation(10, 11, 12) . Moreover, DFMO does not inhibit bacterial pathways of putrescine and cadaverine biosynthesis(13) . The potential therapeutic implications of plasma-borne polyamines are made clear by the demonstration that several invasive tumor types grow very poorly in DFMO-treated rodents fed with antibiotics and a polyamine-deficient diet, whereas treatment with DFMO alone has little, if any, antitumor activity in such models (9, 14) . Cell transformation is in fact characterized by a strikingly enhanced capacity for diamine and polyamine uptake(7) . This is reflected in vivo by a strongly preferential accumulation of radiolabeled polyamines in tumors relative to healthy tissues(10, 15, 16) . Thus, the clinical usefulness of polyamine antimetabolites like DFMO may ultimately be limited by the capacity of tumors to salvage extracellular polyamines.

However, the kinetic and physiological properties of the mammalian polyamine uptake system are poorly understood, and its structure is as yet completely unknown. There is no consensus on the number of different carrier species involved in the specific uptake of polyamines, although at least one high affinity, saturable transport system appears to be shared by putrescine, spermidine, and spermine in several mammalian cell types(7) . One consistently observed property of polyamine uptake is its negative regulation by intracellular polyamine levels(7) . This is first reflected by the up-regulation of polyamine uptake activity which follows the depletion of intracellular polyamine pools by enzyme inhibitors such as DFMO(2, 7, 17) , and which likely contributes to the resistance of many tumors to growth inhibition by these agents in vivo(8, 9, 14) . Conversely, the rate of polyamine uptake quickly decays following preincubation with exogenous polyamines(17, 18, 19, 20) . The mechanism responsible for the feedback inhibition of polyamine transport is incompletely understood. An apparent relief of basal and/or spermidine-induced inhibition of polyamine uptake by protein synthesis inhibitors has been noted earlier in mammalian cells (21, 22) and in the fungus Neurospora crassa(23) , suggesting a requirement for the synthesis of a short-lived polyamine transport repressor (PTR) protein(22, 23) . Recent evidence strongly suggests that the latter repressor might be the ornithine decarboxylase antizyme or a closely related protein(24, 25) .

Polyamines play major physiological functions in the mouse mammary gland (21, 26) as well as in human breast tumors(27, 28, 29, 30, 31) . Breast tumors have elevated polyamine levels as compared with healthy tissues (32, 33) and are exquisitely sensitive to growth inhibition by DFMO in vitro(27, 28, 29, 30, 31) . ()Estradiol [E] has been reported to stimulate ornithine decarboxylase expression in estrogen-responsive human breast cancer cell lines(27, 29, 34) , but more recent evidence suggests that estrogens act mostly to cooperatively enhance the induction of ornithine decarboxylase by polypeptide growth factors such as insulin and insulin-like growth factor-1. ()Earlier studies have demonstrated that polyamine uptake is developmentally regulated and under control by insulin and prolactin in the normal mouse mammary gland(21) . Although the hormonal regulation of polyamine biosynthesis in human breast cancer cells has been the focus of several studies(27, 28, 29, 30, 34) , the polyamine transport system has not been previously characterized in these tumors. The fact that ornithine decarboxylase expression is characteristically low in many breast epithelial carcinoma cells(27) suggests that these tumors might actively rely on uptake of exogenous polyamines, or on an efficient system for salvaging polyamines spontaneously released (35, 36) in the tumor microenvironment.

We are now reporting the kinetic characteristics, substrate specificity, as well as the hormonal and feedback regulation of the polyamine transport system in estrogen-responsive human breast cancer cells. These cells are shown to possess a very active and specific polyamine uptake system with quite narrow structural constraints for substrate recognition. Moreover, estrogens and insulin are not only potent, cooperative inducers of polyamine transport, but are also required for its up-regulation by polyamine depletion in the absence of cell growth. Finally, evidence is presented that net polyamine accumulation is restricted not only by an unstable PTR, but also by other feedback inhibition mechanisms acting on carrier activity as well as at a post-transporter level. The physicochemical requirements of this transport system are described in the companion article(37) .

EXPERIMENTAL PROCEDURES

Materials

N-(3-Aminopropyl)hexahydropyrimidine (APHHP; cf.Table 1) was synthesized in 70% yield by treating sym-norspermidine (N,N`-{bis(aminopropyl)}-1,3-diaminopropane) with a 37% (w/w) aqueous solution of formaldehyde in tetrahydrofuran at room temperature(40) ; H NMR (in CDCl): 1.33 (bs, 2 H, NH), 1.53-1.63 (m, 4 H, CH), 2.26 (t, 2 H, CHN, J = 7.5 Hz), 2.53 (bs, 1 H, NH), 2.70 (t, 4H, CHN, J = 6.7 Hz), 2.78 (t, 2 H, CHN, J = 5.7 Hz), 3.34 (bs, 2H, NCHN); C NMR (CDCl): 69.71, 52.90, 52.82, 44.86, 40.42, 30.27, 26.80. 3-Cyclohexylpropylamine (cf. Table 1) was synthesized as follows. Briefly, a mixture of 3-cyclohexylpropyl chloride (16 g, 100 mmol) and potassium phthalimide (18.5 g, 100 mmol) in dimethylformamide (200 ml) was refluxed for 3 h, cooled to room temperature to give N-3-cyclohexylpropyl phthalimide (25.3 g, 93%) H NMR (CDCl): 0.83-0.86 (m, 2 H), 108-1.34 (m, 6 H), 1.45-1.8 (m, 7 H), 3.65 (t, 2H, J = 7.3 Hz), 7.68-7.77 (m, 2H), 7.77-7.87 (M, 2H); C NMR (CDCl): 168.42, 133.79, 132.16, 123.10, 38.33, 37.30, 34.41, 33.21, 26.59, 26.30, 25.97. The crude phthalimide was treated with 5 ml of hydrazine monohydrate in 100 ml of ethanol and refluxed for 40 min. The mixture was cooled to room temperature, filtered, washed with CHCl to give the crude product which, following in vacuo distillation, yielded 1 g of the pure 3-cyclohexylpropylamine (6.4% yield); H NMR (CDCl): 0.84-0.96 (m, 2 H), 105-1.27 (m, 6 H), 1.39-1.75 (m, 9 H), 2.65 (t, 2H, J = 6.9 Hz); C NMR (CDCl): 42.36, 37.33, 34.44, 33.18, 30.87, 26.46, 26.15.

Cell Culture

Radiometric Determination of Putrescine and Spermidine Uptake

For the determination of kinetic parameters of transport, the substrate concentration was varied by adding increasing concentrations of nonradioactive substrate to a fixed amount of [H]putrescine and [H]spermidine, and the K and V values were determined by Lineweaver-Burk analysis. Competition for uptake by various compounds was measured by adding increasing concentrations of the test substance in the presence of a fixed amount of radioactive substrate (3 µM [H]putrescine or 1 µM [H]spermidine, respectively). The effective concentration of competitor needed for half-maximal uptake inhibition (IC) was obtained by iterative curve fitting of a sigmoidal equation using the SigmaPlot software program (Jandel Scientific, Corte Madera, CA), and K values were calculated using Dixon plots(43) .

Effect of Hormones on Cell Proliferation and Polyamine Transport Activity

Determination of Intracellular Polyamine Contents

RESULTS

Kinetic Parameters of Putrescine and Spermidine Uptake in ZR-75-1 Cells

Figure 1: Kinetic analysis of putrescine and spermidine uptake in ZR-75-1 human breast cancer cells. [H]Putrescine () and [H]spermidine uptake () was determined as described under ``Experimental Procedures.'' Each point is the mean ± S.D. of determinations from triplicate cultures. Insets, Lineweaver-Burk plots of the results.

Figure 2: Inhibition of putrescine and spermidine uptake by common substrates of the transport system in ZR-75-1 cells. [H]Putrescine (A) and [H]spermidine uptake (B) were determined using 10 and 1 µM of substrate, respectively, in the presence of increasing concentrations of spermidine (A, ), spermine (), or putrescine (B, ). Each point is the mean ± S.D. of determinations from triplicate cultures. K values were determined using Dixon plots(41) .

Up-regulation of Polyamine Transport Activity by Estrogens, Insulin, and DFMO

Figure 3: Time course of the effect of E and insulin on cell proliferation and spermidine uptake in ZR-75-1 cells. ZR-75-1 cells were grown for the indicated period in RPMI 1640 medium supplemented with 5% (v/v) dextran-coated charcoal-treated fetal bovine serum containing either no hormone addition (), 10 µg/ml of insulin (), 1 nM E (), or the combination thereof (), and analyzed for DNA content (A) and spermidine uptake activity (B). Each point is the mean ± S.D. of determinations from triplicate cultures. When no bar is shown, the experimental deviation was smaller than the symbol used.

Figure 4: Effect of DFMO, insulin and E on growth and putrescine uptake in ZR-75-1 cells. A and B, ZR-75-1 cells were grown for the indicated time in RPMI 1640 medium supplemented with 5% (v/v) dextran-coated charcoal-treated fetal bovine serum containing either no hormone addition (control, ), 500 µM DFMO (), 10 µg/ml insulin and 1 nM E (), or both DFMO and the hormone combination (), and assayed for DNA content (A) and putrescine uptake (B). C, Lineweaver-Burk analysis of putrescine uptake characteristics in ZR-75-1 cells treated for 4 days as in B.

Mitogenic stimulation has been temporally associated with increased polyamine transport in other cell types(19, 46, 47) . On the other hand, depletion of intracellular putrescine and spermidine by DFMO also increases the rate of polyamine uptake, presumably by relieving endogenous feedback inhibition of transport(2, 7, 17) . Since DFMO potently inhibits breast cancer cell proliferation(27, 28, 29, 34) , the relationship between the enhancement of polyamine transport by E and insulin and their mitogenic effect was further examined by comparing the effect of DFMO on putrescine uptake activity in the presence or absence of these hormones. Preliminary experiments had shown that the hormone ablation-induced decay in putrescine uptake activity could be accelerated by increasing the plating density (data not shown). Thus, this stepdown strategy was used here to better evaluate the up-regulatory effect of hormone treatment on putrescine uptake activity. DFMO had a limited inhibitory action on the already low growth rate observed in the absence of mitogens (Fig. 4A), and stably increased the rate of putrescine transport by about 33% as early as 24 h after its addition (Fig. 4B). On the other hand, DFMO almost completely blocked the mitogenic effect of insulin and E, while potentiating the associated increase in putrescine uptake activity. As shown in Fig. 4C, elevations in putrescine transport activity induced by insulin and E as well as by DFMO after a 4-day treatment solely resulted from an increase in the V of the uptake process. Likewise, the hormonal and DFMO-induced stimulation of spermidine uptake did not modify substrate affinity (data not shown).

Down-regulation of Polyamine Transport Activity by Retinoids

Figure 5: Effect of retinoic acid on cell growth and spermidine transport in ZR-75-1 cells. Three days after seeding in RPMI 1640 medium supplemented with 5% (v/v) dextran-coated charcoal-treated fetal bovine serum, cells were preincubated for 2 days in the presence (, ) or absence (, ) of insulin (10 µg/ml) and E (1 nM). Retinoic acid (RA, 1 µM) (solidsymbols) or the ethanol vehicle (0.1%, v/v) (plainsymbols) was then further added at time 0 of the time course shown, and cells were analyzed for DNA content (A) and spermidine uptake activity (B). Each point is the mean ± S.D. of determinations from triplicate cultures.

Dependence of Feedback Inhibition of Polyamine Transport on Protein Synthesis and Structural Requirements For Its Induction

In order to first address the requirement for protein synthesis for feedback transport inhibition in ZR-75-1 cells, the acute effect of spermidine or putrescine on polyamine uptake activity was determined in the presence or absence of cycloheximide (CHX) (Fig. 6A). Exogenous spermidine (20 µM) caused a nearly maximal down-regulation (90% inhibition) of its own uptake rate within 2 h, with an approximately 50% reduction already observed after 1 h, thus indicating the presence of a rapid and efficient mechanism of feedback transport repression in these cells. An equimolar concentration of putrescine was much less potent in that respect, slowly decreasing the velocity of spermidine uptake by about 60% after 6 h. While CHX alone slightly inhibited spermidine uptake activity, it delayed the onset of feedback transport inhibition so that a comparable degree of polyamine-induced repression required about 6 h in its presence. CHX blocked the down-regulation of polyamine transport induced by putrescine much more potently than that caused by spermidine (Fig. 6A).

Figure 6: Dependence on protein synthesis and structural specificity of feedback inhibition of spermidine transport in ZR-75-1 cells. A, at time 0, 20 µM spermidine (, ), 20 µM putrescine (, ), or no amine supplement (, ) was added to ZR-75-1 cell cultures in standard growth medium containing 1 mM aminoguanidine, in the presence (solidsymbols) or absence (plainsymbols) of 200 µM CHX. [H]Spermidine uptake activity was measured at the indicated times after rinsing cell monolayers twice with serum- and amine-free RPMI 1640 medium. B, prior to the experiment, cells were grown for 6 days in standard growth medium containing 1 mM aminoguanidine, with 10 µM AbeAdo being added to one group of cells for the last 24 h. Medium was then supplemented with the indicated concentrations of spermine (), spermidine (), putrescine (), or putrescine + 10 µM AbeAdo (). After a 3-h incubation period, cell monolayers were rinsed twice with serum- and amine-free RPMI 1640 medium, and [H]spermidine uptake activity then determined. Each point is the mean ± S.D. of determinations from triplicate cultures.

The ability of various substrates to down-regulate polyamine transport was compared by preincubating ZR-75-1 cells for 3 h with putrescine (0.3-30 µM), spermidine (0.03-10 µM), or spermine (0.01-3 µM) prior to the determination of uptake activity (Fig. 6B). Spermidine and spermine were both very potent feedback repressors of polyamine transport, with IC values of 0.09 ± 0.03 and 0.21 ± 0.07 µM, respectively, while putrescine was 25-60-fold less potent (IC = 5.4 ± 1.2 µM). Furthermore, pretreating the cells with 10 µM AbeAdo, an irreversible inhibitor of S-adenosylmethionine decarboxylase (39) which depletes spermidine and spermine while causing substantial putrescine accumulation(52, 53) , did not notably decrease the effective concentration of exogenous putrescine needed to repress spermidine uptake (IC 3.6 ± 0.7 µM; Fig. 6B). The latter observation, and the fact that AbeAdo treatment alone increased the basal rate of spermidine uptake by 40-50%, strongly suggest that spermidine and spermine are the major effectors for the feedback inhibition of polyamine transport.

Influence of the Feedback Repression of Polyamine Transport on Intracellular Accumulation of Exogenous Polyamines

()

The effect of CHX on the time course of internalization of H-labeled spermidine (5 µM) in ZR-75-1 cells is shown in Fig. 7. [H]Spermidine accumulation reached a maximal, steady-state level within 2 h, i.e. the time required for a maximal, protein synthesis-dependent decrease in the uptake rate (cf.Fig. 6A). CHX prevented this stabilization and led to a linear substrate accumulation (up to 4-fold), which abruptly stopped between 4 and 6 h, in keeping with the time course observed for the protein synthesis-independent component of feedback transport inhibition. In order to assess whether maintenance of the feedback suppression of polyamine uptake required ongoing protein synthesis, the effect of CHX on [H]spermidine accumulation was also monitored in parallel cell cultures preincubated with 10 µM 1-methylspermidine (MeSpd) for the last 3 h before addition of the labeled precursor. While MeSpd can functionally replace spermidine in mammalian cells, it is a very poor substrate for spermine synthase and is completely resistant to acetylation by spermidine/spermine N-acetyltransferase(38, 57) , an enzyme involved in polyamine catabolism and excretion(80) . This spermidine homolog was thus expected to exert its intracellular effects longer than the natural polyamines. As shown in Fig. 7, the rate of [H]spermidine accumulation was decreased by about 70% in cells preincubated with MeSpd as a result of feedback transport inhibition. However, [H]spermidine accumulation in MeSpd-treated cells was increased 3-5-fold by CHX, a significant stimulation being noted as early as after 30 min, and stopped between 4 and 6 h when intracellular [H]spermidine levels had reached approximately 120 nmol/mg DNA. When taking into account the initial MeSpd pool already internalized during the preincubation period (cf.Fig. 8A), the total polyamine (MeSpd + spermidine) accumulation at which net [H]spermidine uptake ceased in these cells was comparable to that measured in CHX-treated cells incubated with spermidine only (i.e. about 200 nmol/mg DNA).

Figure 7: Effect of CHX on the time course of [H]spermidine accumulation in ZR-75-1 cells. ZR-75-1 cells were preincubated for 3 h in the absence (, ) or presence (, ) of 10 µM MeSpd in standard growth medium containing 1 mM aminoguanidine. At time 0, medium was substituted with serum-free RPMI 1640 medium containing 5 µM [H]spermidine plus (solidsymbols) or minus (plainsymbols) 200 µM CHX, and intracellular radioactive content was determined at the intervals shown. C, control cells. Each point is the mean ± S.D. of determinations from triplicate cultures.

Figure 8: Effect of CHX on MeSpd accumulation in ZR-75-1 cells. Cells were incubated in standard growth medium containing 10 µM MeSpd in the presence () or absence () of 200 µM CHX and harvested at the times shown for determination of MeSpd content by high performance liquid chromatography (A), and total trichloroacetic acid-insoluble protein content (B). Each point is the mean ± S.D. of determinations from triplicate cultures.

As for [H]spermidine, MeSpd accumulation rapidly reached a maximum within 2 h, thus indicating that MeSpd compares to spermidine as a feedback transport inhibitor (Fig. 8A). On the other hand, CHX caused an almost linear rate of MeSpd internalization (1 nmol/h/culture) for up to 8 h, resulting in a 4-fold higher accumulation of the analog. The maximal, net MeSpd accumulation thus reached (equivalent to 180 nmol/mg DNA) was similar to that observed with [H]spermidine after only 4 h (cf.Fig. 7). This buildup was followed by an abrupt decrease in total MeSpd content per culture (Fig. 8A) coinciding with an acceleration of net protein loss detected in CHX-treated cells (Fig. 8B). No significant effect on total cellular protein content was observed in cells incubated for up to 12 h with CHX in the absence of exogenous polyamines (data not shown), strongly suggesting that the observed toxicity was related to the accumulation of the polyamine rather than to the effect of CHX per se.

Effect of DFMO-induced Polyamine Depletion on the Feedback Inhibition of Polyamine Transport

Figure 9: Effect of DFMO on feedback inhibition of spermidine uptake activity. Cells were grown in standard growth medium for 5 days, with 1 mM DFMO being added to half of the cell cultures during the last 3 days. Medium was then replaced with serum-free RPMI 1640 medium plus (B) or minus (A) 1 mM DFMO, and containing either 20 µM spermidine (, ) or no amine supplement (, ), in the absence (plainsymbols) or presence (solidsymbols) of 200 µM CHX, and [H]spermidine uptake activity was assayed at the indicated times. Each point is the mean ± S.D. of determinations from triplicate cultures.

In order to compare the effect of DFMO and CHX on intracellular polyamine accumulation, ZR-75-1 cells were preincubated for 4 days in the presence or absence of 1 mM DFMO, and the differential effect of CHX on putrescine and spermidine internalization then determined (Table 3). The main change in polyamine content induced by DFMO in ZR-75-1 cells was a 20% decrease in the level of spermine, the major polyamine species in this cell line, while the already low putrescine and spermidine contents fell below detection limits. Despite this apparently modest decrease in total polyamine content, DFMO pretreatment alone increased the net accumulation of exogenous putrescine and spermidine by 77 and 94%, respectively. While CHX increased net putrescine and spermidine internalization 2.0- and 6.8-fold, respectively, its differential effect on putrescine accumulation was much weaker (+30%, p < 0.05) in the presence of DFMO. On the other hand, spermidine accumulation was still markedly sensitive to CHX (3.2-fold increase) in DFMO-treated cells. Interestingly, putrescine accumulation led to an increase in spermidine content in both control and DFMO-treated cells, even in the presence of CHX. This observation is consistent with the low ornithine decarboxylase expression found in ZR-75-1 cells being a major limiting factor for polyamine biosynthesis in this cell line. Low amounts of N-acetylspermidine (1.0 ± 0.1 nmol/mg protein) were also formed upon spermidine addition in DFMO-treated, but not in control or CHX-treated cells (data not shown). As shown in Fig. 10, although DFMO markedly increased the initial velocity of [H]spermidine internalization, feedback inhibition of polyamine accumulation followed a similar time course in control and DFMO-treated cells, with a virtual cessation in net internalization observed after about 2 h. Likewise, DFMO did not qualitatively alter the pattern of [H]spermidine internalization observed upon derepression of transport by CHX, except that the rate of accumulation was increased 1.5-2-fold.

Figure 10: Effect of DFMO and CHX on [H]spermidine accumulation in ZR-75-1 cells. Prior to the experiment, cells were grown in the presence or absence of 1 mM DFMO as described in Fig. 9. At time 0, serum-free medium containing 5 µM [H]spermidine was added in the absence (plainsymbols) or presence (solidsymbols) of 200 µM CHX to control (, ) or DFMO-treated cells (, ). Intracellular radioactive content was determined at the indicated times. The ornithine decarboxylase inhibitor was present throughout the experimental period. Each point is the mean ± S.D. of determinations from triplicate cultures.

DISCUSSION

A differential dependence on Na for putrescine and polyamine transport has led to postulate the existence of multiple classes of polyamine carriers in some models (for review, see Refs. 7, 58). As documented in the companion paper(37) , the physicochemical requirements for putrescine and spermidine uptake are qualitatively very similar in ZR-75-1 cells, and both processes are Na-independent. On the other hand, evidence for multiple putrescine and polyamine uptake systems has also been based on the pattern of uptake inhibition by mutual competitors(58, 59, 60) , whereas a single class of carrier could account for di- and polyamine transport in other models(12, 19, 61) . As shown by genetic complementation(62) , although at least two different loci are involved in putrescine and polyamine transport-deficient phenotypes, it is not clear whether more than one structural transport gene is involved.

On the basis of the kinetic parameters herein presented for breast cancer cells, putrescine and spermidine separately behaved as substrates for single, saturable transporters with a respective affinity (putrescine < spermidine spermine) typical of most mammalian cells(7) . The findings that (i) spermidine inhibited putrescine uptake with a K very close to its K as a substrate, and that [ii] spermine and APHHP, respectively, blocked putrescine and spermidine uptake with closely similar K values for both substrates, are consistent with competition for a common carrier. On the other hand, the K of putrescine toward spermidine uptake was much lower than expected from its own affinity as a substrate. Such a behavior could reflect the existence of distinct high affinity carriers, one being mainly responsible for polyamine transport, and the second being shared by putrescine as well as polyamines, as proposed for human endothelial cells(58) . However, this would predict a biphasic inhibition of spermidine uptake by increasing concentrations of putrescine, unlike the simple sigmoidal relationship observed (Fig. 2B). Moreover, Hill plots of spermidine uptake up to 100 µM reproducibly yielded coefficients of unity (data not shown), thus arguing against the existence of multiple classes of spermidine transporters in ZR-75-1 cells. Therefore, if distinct carriers exist for putrescine and polyamine transport, non-substrates should inhibit the activity of a given transporter type through a purely antagonistic interaction, e.g. polyamines should inhibit putrescine binding to its specific carrier while not being themselves substrates.

Previous studies have shown that aliphatic compounds with two unhindered protonated amine centers separated by a distance roughly corresponding to that found between the terminal amino groups of spermidine is optimal for an efficient interaction with both diamine and polyamine uptake systems in mammalian cells(63, 64, 65) . This was confirmed confirmed here by the relative ability of 1,6-diaminohexane, 1,7-diami-noheptane, and spermidine to inhibit putrescine uptake (Table 1). Moreover, the central imino group and one of the terminal amino groups of spermidine and its homologs can be alkylated without major loss of affinity for the transport system, as found with APHHP (Table 1; (40) ). The inability of 3-cyclohexylpropylamine or 4-(3-aminopropyl)morpholine to compete for spermidine uptake in ZR-75-1 cells further indicates that the secondary amino group of APHHP is essential to preserve a high affinity for the polyamine uptake system, while the sterically hindered tertiary amino nitrogen is most likely dispensable in that respect. Indeed, substantial structural modifications of the methylene backbone between the two terminal amino groups of polyamines can be tolerated, as with 1,4-bis(3-aminopropyl)piperazine, but at the expense of substrate affinity, as noted earlier for N-substituted spermidine derivatives(66) . In addition to the key role of properly distant and unhindered amine centers, specific hydrophobic interactions must also largely contribute to the higher affinity of spermidine for the polyamine transporter as compared to putrescine. This is strongly suggested by the fact that N-alkylated 1,3-diaminopropane derivatives are much better competitors than 1,3-diaminopropane against putrescine uptake in ZR-75-1 cells(31) . Similar hydrophobic interactions likely explain the increased ability of aliphatic amines to inhibit spermidine uptake with a lengthening of the methylene carbon chain (Table 1; Refs. 40, 63).

A single putrescine/spermidine carrier model might reconcile these observations if one postulates that spermidine, which differs from putrescine by the addition of an aminopropyl group, binds to the same recognition site as putrescine through hydrophobic as well as electrostatic interactions unavailable for the diamine as a substrate. Thus, a single putrescine/polyamine carrier could exhibit qualitative differences in substrate binding characteristics toward the two classes of substrates, which could account for difference between the K and K of putrescine as a substrate and a spermidine competitor, respectively. Interestingly, cross-inhibition observed between putrescine and the herbicide paraquat for their uptake in rat alveolar type II cells (67, 68) has been described as partial competitive inhibition (43) resulting from a decrease in substrate affinity caused by prior binding of the competitor to a distinct site on the same carrier(69) .

Estrogens exert major mitogenic effects in human breast cancer cells expressing functional cognate receptors(70) . Several targets such as the progesterone receptor, pS2, and procathepsin D genes have initially been identified as being under specific estrogenic control(70) . More recent evidence, however, indicates that polypeptide growth factors can induce these same genes by activating common transcriptional elements involved in the transduction pathway of estrogens(71, 72) . The present data provide the first evidence that a membrane transport system can also be under dual, synergistic regulation by estrogens and a polypeptide growth factor. Moreover, the E- and insulin-induced increase in polyamine uptake activity is an early event in the cascade of events leading to exponential cell growth, which is cooperatively stimulated by both hormones in ZR-75-1 as well as other breast cancer cell lines(71) . Although putrescine and spermidine depletion blocks the mitogenic action of insulin and E, it does not interfere with the hormonal stimulation of polyamine transport activity. The fact that the hormone-dependent induction of the polyamine transport system was independent of the cell growth status shows that it does not merely result from the elevated ion fluxes accompanying mitogenic activation, but may rather represent a novel, genuine target of estrogen and growth factor action in mammary tumor cells. Whether the hormonal induction of polyamine transport results from the increased activity of a constant number of transporters or in the net addition of new carrier molecules to the membrane is as yet undetermined, but insulin is known to control the expression of the mammalian transporters GLUT-1 and GLUT-4 through both mechanisms(73) .

Polyamine uptake in other tissue types can be stimulated by peptide hormones, serum, and other mitogenic stimuli(7) . However, with few exceptions(74) , most of these effects were not clearly dissociated from the general trophic effect of these agents. It is noteworthy that insulin and prolactin induce polyamine uptake activity in the lactating mouse mammary gland(21) , suggesting that the hormonal regulation of polyamine transport may be a more general feature of mammary epithelial cells. The dual and cooperative regulation of ornithine decarboxylase expression and polyamine transport by estrogens and insulin in human breast cancer cells may constitute a coordinate response to increase the polyamine pool at early steps in the hormonal activation of macromolecular synthesis. It should be noted that the rate of polyamine biosynthesis is low, even under optimal growth conditions, in most human breast cancer cell lines examined(27, 31) . Polyamine transport in breast tumor cells may thus play a key role in polyamine homeostasis in vivo where amino acid precursors for endogenous polyamine synthesis are likely more limiting than in the in vitro context. This idea is supported by the recent finding that arginine limitation up-regulates polyamine transport in human endothelial cells, an effect which is neither suppressed by ornithine or putrescine(58) .

Interestingly, polyamine transport was found to be under acute negative regulation by retinoic acid in ZR-75-1 cells. Very few early biological responses to retinoids have yet been reported in human breast cancer cells(50, 51) . The rapid decay of spermidine uptake activity induced by retinoic acid, which preceded any detectable effect on ZR-75-1 cell growth, suggests that this membrane transport system might be a proximal target of retinoid action in human breast cancer cells. Since retinoic acid antagonizes growth factor action and inhibits estrogen-dependent gene expression in human breast carcinoma cells(50, 51) , down-regulation of polyamine uptake might result from a similar effect on the expression of the E- and insulin-regulated carrier. However, prior hormonal induction of spermidine uptake activity markedly postponed any detectable effect of retinoic acid on the latter parameter. Thus, retinoic acid does not appear to interfere efficiently with the transduction pathway synergistically activated by estrogens and insulin, but might rather increase the turnover of the polyamine carrier, which has a relatively long half-life(20) . Retinoic acid has been reported to prevent the induction of putrescine transport by blocking the G/G transition in initially quiescent mouse hepatocytes(47) . It remains to be determined whether retinoids may modulate polyamine transport by affecting cell cycle distribution in breast tumor cells.

The present data clearly indicate that polyamine transport activity in human breast cancer cells is under stringent and rapid feedback inhibition by internalized polyamines, as in other mammalian cell types (17, 18, 19, 20, 21, 22) . The early feedback repression of polyamine transport in ZR-75-1 cells requires the de novo synthesis of a short-lived PTR protein, which represses carrier activity independently of its intrinsic transport capacity, which is metabolically stable (Fig. 7). Similar characteristics have been reported for the feedback inhibition of polyamine uptake in rat hepatoma cells and Chinese hamster ovary cells(22) . This PTR might be identical to the ornithine decarboxylase antizyme(75, 76) , as elegantly suggested by transfection experiments with expression vectors encoding a functional antizyme cDNA(24, 25) . Accordingly, antizyme, in addition to its role in directing the rate of ornithine decarboxylase degradation(76) , may function in the reversible down-regulation of polyamine carrier activity. In virtue of its short half-life and its rapid inducibility by polyamines, presumably at the translational level through a unique mode of ribosomal frameshifting(77) , antizyme would be especially well suited for the acute repression of polyamine transport.

Nevertheless, the feedback system regulating polyamine carrier activity in human breast cancer cells shows several novel regulatory features. First, polyamine carrier activity was not only repressed by the rapid induction of an unstable, antizyme-like PTR, but also by an as yet undescribed, more latent mechanism which does not require protein synthesis. Since preventing PTR synthesis deregulates polyamine accumulation, the down-regulation of polyamine carrier activity observed in CHX-treated cells might have resulted from a nonspecific, perhaps toxic effect of excessive polyamine levels. However, polyamine carrier activity was under feedback inhibition almost independently from de novo protein synthesis in DFMO-treated ZR-75-1 cells (Fig. 9) under conditions of controlled spermidine accumulation ( Table 3and Fig. 10). Thus, internalized polyamines can repress carrier activity through an additional mechanism which does not require de novo protein synthesis nor result nonspecifically from their potential cytotoxicity. Such a mechanism could be analogous to the trans-inhibition effect of internalized substrates on system A amino acid carrier activity, which is thought to result from the ``locking'' of the substrate-bound carrier complex in the cytoplasmic orientation(78) . Alternatively, internalized polyamines might activate a relatively stable modifier molecule such as casein kinase II, which is known to be strongly stimulated by polyamines(79) , and thus inhibit carrier activity through post-translational modifications.

Second, no evidence for an unstable PTR activity could be detected in ZR-75-1 human breast cancer cells without ongoing polyamine accumulation, unlike findings reported in rat and hamster cell lines (22) . Up-regulation of putrescine and spermidine uptake activity by CHX without prior incubation with exogenous substrate has also been reported in the fungus N. crassa(23) and in the lactating mouse mammary gland(21) , respectively. If intracellular polyamine levels can regulate PTR synthesis, the endogenous or free polyamine pool prevailing in ZR-75-1 cells would thus appear to be too low to sustain a basal repression of polyamine uptake activity. Therefore, the up-regulation of polyamine transport activity caused by DFMO in ZR-75-1 cells does not likely result from the suppression of a basal rate of PTR synthesis by polyamine depletion, but rather from an increase in the total polyamine transport capacity.

As a third, novel feature of the present model, the inducibility of the putative PTR by internalized polyamines was found to be markedly impaired by DFMO. A significant delay in PTR induction due to the time needed to restore a normal, subthreshold spermidine content in DFMO-treated cells is unlikely, since endogenous spermidine level in control cells (1 nmol/mg DNA) represents only a minor fraction of the amount accumulated within the first hour ( Fig. 8and Fig. 10). Since polyamines rapidly favor antizyme translation(77) , sustained polyamine starvation could decrease steady-state antizyme mRNA levels in ZR-75-1 cells, assuming the identity of the PTR with antizyme. It should be noted that antizyme mRNA expression is apparently constitutive in most vertebrate tissues and does not respond to putrescine addition(75) , although its possible regulation by endogenous polyamines has not been reported to our knowledge. Alternatively, the onset of PTR induction could be delayed in DFMO-treated ZR-75-1 cells due to a nonspecific effect of polyamine depletion on the overall rate of protein synthesis (6) .

In addition to PTR induction and trans-inhibition of carrier activity by internalized polyamines, polyamine accumulation may also be regulated by other homeostatic mechanisms beyond membrane carrier activity. This is suggested by the fact that, although CHX did not markedly alter the kinetics of down-regulation of spermidine carrier activity in DFMO-treated ZR-75-1 cells, protein synthesis inhibition dramatically increased spermidine, but not putrescine accumulation. Thus, although DFMO impairs PTR induction and up-regulates polyamine transport activity, spermidine overaccumulation in DFMO-treated cells is still largely prevented by at least one additional mechanism requiring protein synthesis. For instance, the synthesis of a protein regulating polyamine efflux might be induced when intracellular polyamine levels exceed a given threshold, which could account for a decrease in net substrate accumulation despite ongoing uptake. However, our preliminary results strongly indicate that the rate of polyamine efflux is not affected by the extent of substrate accumulation (data not shown). An alternative candidate might be spermidine/spermine N-acetyltransferase, which is rapidly induced by polyamines and analogs, but much less so by putrescine, and is likely involved in the homeostasis of polyamine pools(80) , as suggested by the low amounts of N-acetylspermidine formed in DFMO-treated cells incubated with spermidine. Whether spermidine/spermine N-acetyltransferase induction can solely account for the post-carrier regulation of polyamine accumulation remains to be determined by comparison with acetylation-resistant polyamine analogs.

Upon inhibition of de novo protein synthesis, ZR-75-1 cells accumulate 3-6-fold higher levels of various substrates of the polyamine transport system, as in other eukaryotic cells(22) . There is now substantial evidence that excessive intracellular polyamine levels are highly toxic to eukaryotic cells(54, 55, 56, 57) . In fact, polyamine overaccumulation leads to rapid cell death by apoptosis in ornithine decarboxylase-overexpressing L1210 cells, without prior oxidation of the accumulated substrate(57) . A similar defect in the feedback regulation of spermidine uptake leading to lethal polyamine accumulation has been reported in ornithine decarboxylase-overproducing HTC cell variants(55) . If antizyme is indeed the polyamine-induced PTR, the intriguing connection between high ornithine decarboxylase expression and an altered feedback regulation of polyamine transport might find an explanation in an abnormally high mobilization of antizyme molecules by the large pool of ornithine decarboxylase, thus decreasing their availability for feedback transport inhibition. Thus, the present data support the notion that multiple mechanisms of feedback inhibition might be needed to prevent apoptosis or other forms of cytotoxicity (57) caused by polyamine overaccumulation. Quite notably, putrescine induced PTR protein activity only after a long lag period and caused virtually no down-regulation of spermidine uptake in the presence of CHX. Thus, the diamine is much less active than spermidine or spermine both as an inducer of PTR and as a trans-inhibitor of polyamine uptake, and the present data suggest that its limited repressing effect might in fact be due to its conversion to spermidine. It is noteworthy that putrescine overaccumulation caused by inhibition of S-adenosylmethionine decarboxylase(53) , ornithine decarboxylase overproduction(52) , or the relief of feedback inhibition of putrescine uptake (55) is not overtly toxic, perhaps with some exceptions (56) or may even be beneficial in certain physiological situations(52) . The comparatively low cytotoxicity of putrescine could therefore account for its much weaker activity in feedback transport inhibition.

The present finding that DFMO-induced polyamine depletion strongly interferes with the induction of an unstable PTR by internalized polyamines, but not with the overall regulation of net polyamine accumulation, raises new questions as to the exact role of this putative protein in the regulation of polyamine accumulation. PTR induction and, at later stages, trans-inhibition by internalized substrates might mostly govern the initial velocity of polyamine transport, with post-carrier mechanisms dependent on de novo protein synthesis such as spermidine/spermine N-acetyltransferase induction for the final adjustment of the size of the polyamine pool.

FOOTNOTES

*

This research was supported by grants from the Cancer Research Society Inc. and by the Fonds de la Recherche en Santé du Québec. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked ``advertisement'' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§

Supported by Endorecherche Inc.

¶

To whom correspondence should be addressed.

()

The abbreviations used are: DFMO, DL--difluoromethylornithine; AbeAdo, 5`-{[(Z)-4-amino-2-butenyl]methylamino}-5`-deoxyadenosine; APHHP, N-(3-aminopropyl)hexahydropyrimidine; CHX, cycloheximide; E, 17-estradiol; MeSpd, 1-methylspermidine; PTR, polyamine transport repressor.

()

R. Poulin and M. Huber, manuscript in preparation.

()

M. Huber and R. Poulin, manuscript submitted for publication.

()

R. Poulin, A. Paterakis, G. Pelletier, and A. E. Pegg, manuscript submitted for publication.

ACKNOWLEDGEMENTS

REFERENCES

1. Tabor, C. W., and Tabor, H. (1984) Annu. Rev. Biochem. 53, 749-790 [CrossRef][Medline] [Order article via Infotrieve]
2. Pegg, A. E. (1988) Cancer Res. 48, 759-774 [Abstract/Free Full Text]
3. Heby, O., and Persson, L. (1990) Trends Biochem. Sci. 15, 153-158 [CrossRef][Medline] [Order article via Infotrieve]
4. Celano, P., Baylin, S. B., and Casero, R. A., Jr. (1989) J. Biol. Chem. 264, 8922-8927 [Abstract/Free Full Text]
5. Park, M. H., Wolff, E. C., and Folk, J. E. (1993) Trends Biochem. Sci. 18, 475-479 [CrossRef][Medline] [Order article via Infotrieve]
6. Marton, L. J., and D. R. Morris. (1987) in Inhibition of Polyamine Metabolism : Biological Significance and Basis for New Therapies (McCann, P. P., Pegg, A. E., and Sjoerdsma. A., eds) pp. 79-105, Academic Press, New York
7. Seiler, N., and Dezeure, F. (1990) Int. J. Biochem. 22, 211-218 [CrossRef][Medline] [Order article via Infotrieve]
8. Persson, L., Holm, I., Ask, A., and Heby, O. (1988) Cancer Res. 48, 4807-4811 [Abstract/Free Full Text]
9. Ask, A. Persson, L., and Heby, O. (1992) Cancer Lett. 66, 29-34 [CrossRef][Medline] [Order article via Infotrieve]
10. Sarhan, S., Knödgen, B., and Seiler, N. (1989) Anticancer Res. 9, 215-224 [Medline] [Order article via Infotrieve]
11. Hessels, J., Kingma, A. W., Ferwerda, H., Keij, J., van der Berg, G. A., and Muskiet, F. A. J. (1989) Int. J. Cancer 43, 1155-1166 [Medline] [Order article via Infotrieve]
12. Osborne, D. L., and Seidel, E. R. (1990) Am. J. Physiol. 258, G576-G584
13. Bey, P., Danzin, C., and Jung, M. (1987) in Inhibition of Polyamine Metabolism : Biological Significance and Basis for New Therapies (McCann, P. P., Pegg, A. E., and Sjoerdsma, A., eds) pp. 1-32, Academic Press, Orlando, FL
14. Seiler, N., Sarhan, S., Grauffel, C., Jones, R., Knödgen, B., and Moulinoux, J.-P. (1990) Cancer Res. 50, 5077-5083 [Abstract/Free Full Text]
15. Chaney, J. E., Kobayashi, K., Goto, R., and Digenis, G. A. (1983) Life Sci. 32, 1237-1241 [CrossRef][Medline] [Order article via Infotrieve]
16. Volkow, N., Goldman, S. S., Flamm, E. S., Cravioto, H., Wolf, A. P., and Brodie, J. D. (1983) Science 221, 673-675 [Abstract/Free Full Text]
17. Alhonen-Hongisto, L. Seppänen, P., and Jänne, J. (1980) Biochem. J. 192, 941-945 [Medline] [Order article via Infotrieve]
18. Rinehart, C. A., and Chen, K. Y. (1984) J. Biol. Chem. 259, 4750-4756 [Abstract/Free Full Text]
19. Kakinuma, Y., Hoshino, K., and Igarashi, K. (1988) Eur. J. Biochem. 176, 409-414 [Medline] [Order article via Infotrieve]
20. Byers, T. L., and Pegg, A. E. (1990) J. Cell. Physiol. 143, 460-467 [CrossRef][Medline] [Order article via Infotrieve]
21. Kano, K. and Oka, T. (1976) J. Biol. Chem. 251, 2795-2800 [Abstract/Free Full Text]
22. Mitchell, J. L. A., Diveley, R. R., Jr., and Bareyal-Leyser, A. (1992) Biochem. Biophys. Res. Commun. 186, 81-88 [CrossRef][Medline] [Order article via Infotrieve]
23. Davis, R. L., Ristow, J. L., Howard, A. D., and Barnett, G. R. (1991) Arch. Biochem. Biophys. 285, 297-305 [CrossRef]