Effect of plasma concentrations of uridine on pyrimidine biosynthesis in cultured L1210 cells.

The concentration of uridine in the media of cultured L1210 cells was maintained within the concentration range found in plasma (1 to 10 microM) to determine if such concentrations are sufficient to satisfy the pyrimidine requirements of a population of dividing cells and to determine if cells utilize de novo and/or salvage pathways when exposed to plasma concentrations of uridine. When cells were incubated in the presence of N-(phosphonacetyl)-L-aspartate to block de novo biosynthesis, plasma concentrations of uridine maintained normal cell growth. De novo pyrimidine biosynthesis, as determined by [14C]sodium bicarbonate incorporation into uracil nucleotides, was affected by the low concentrations of uridine found in the plasma. Below 1 microM uridine, de novo biosynthesis was not affected; between 3 and 5 microM uridine, de novo biosynthesis was inhibited by approximately 50%; and above 12 microM uridine, de novo biosynthesis was inhibited by greater than 95%. Inhibition of de novo biosynthesis correlated with an increase in the uracil nucleotide pool. The de novo pathway was much more sensitive to the uracil nucleotide pool size than was the salvage pathway, such that when de novo biosynthesis was inhibited by greater than 95% the uracil nucleotide pool continued to expand and the cells continued to take up [14C]uridine. Thus, the pyrimidine requirements of cultured L1210 cells can be met by concentrations of uridine found in the plasma and, when exposed to such physiologic concentrations, L1210 cells decrease their dependency on de novo biosynthesis and utilize their salvage pathway. Circulating uridine, therefore, may be of physiologic importance and could be an important determinant in anti-pyrimidine chemotherapy.

The concentration of uridine in the media of cultured L1210 cells was maintained within the concentration range found in plasma (1 to 10 PM) to determine if such concentrations are sufficient to satisfy the pyrimidine requirements of a population of dividing cells and to determine if cells utilize de novo and/or salvage pathways when exposed to plasma concentrations of uridine. When cells were incubated in the presence of N-(phosphonacety1)-L-aspartate to block de novo biosynthesis, plasma concentrations of uridine maintained normal cell growth. De novo pyrimidine biosynthesis, as determined by [14CJsodium bicarbonate incorporation into uracil nucleotides, was affected by the low concentrations of uridine found in the plasma. Below 1 PM uridine, de nouo biosynthesis was not affected; between 3 and 5 NM uridine, de novo biosynthesis was inhibited by approximately 50%; and above 12 uridine, de novo biosynthesis was inhibited by >95%.
Inhibition of de novo biosynthesis correlated with an increase in the uracil nucleotide pool. The de novo pathway was much more sensitive to the uracil nucleotide pool size than was the salvage pathway, such that when de novo biosynthesis was inhibited by ~9 5 % the uracil nucleotide pool continued to expand and the cells continued to take up [I4C]uridine. Thus, the pyrimidine requirements of cultured L1210 cells can be met by concentrations of uridine found in the plasma and, when exposed to such physiologic concentrations, L1210 cells decrease their dependency on de novo biosynthesis and utilize their salvage pathway. Circulating uridine, therefore, may be of physiologic importance and could be an important determinant in antipyrimidine chemotherapy.
The importance of the plasma as a source of salvageable pyrimidines is not firmly established. The in vivo toxicity and anti-tumor effectiveness of inhibitors of de novo pyrimidine biosynthesis such as PALA' and pyrazofurin (1,2) can be reversed by uridine administration, indicating that uridine derived from plasma can satisfy the pyrimidine requirements of cells with an intact salvage pathway. Likewise, uridine has been used to manage patients with hereditary orotic aciduria, a disease in which the enzymes orotate phosphoribosyltransferase and orotidylate decarboxylase of the de novo biosynthetic pathway are deficient (3). However, in each of these * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
tate; PBS, a 0.09% NaCl solution containing 0.075% K,HPO, and examples, plasma uridine concentrations would be expected to be elevated above the normal state because of the exogenous uridine administered; thus, the capacity of physiologic concentrations of plasma pyrimidines to satisfy cellular pyrimidine requirements under normal, unperturbed conditions remains to be established. Indeed, determining the relative dependences of normal and tumorous tissues on de novo versus salvage pathways i n vivo could greatly influence the future of antipyrimidine cancer chemotherapy.
Recent advances in analytical methodology has allowed quantitation of the low concentrations of pyrimidines in plasma and serum. The circulating plasma concentrations of uridine in humans, mice, and rats are between 1 and 10 FM and are relatively constant for each species (4). Furthermore, the liver has been identified as an important source and modulator of circulating uridine (5,6). Studies to date, however, suggest that circulating concentrations of uridine are too low to be physiologically significant. For example, the concentration of uridine required to reverse the growth-inhibitory effects of PALA in cultured SV40-transformed kidney and human colonic HT-29 cells is 100 ptM ( 7 , 8 ) , a concentration at least 10-fold higher than that found in plasma. Likewise, Hoogenraad and Lee (9) reported a 70% decrease in de novo pyrimidine biosynthesis when cultured rat hepatoma cells were incubated with 500 p~ uridine; a uridine concentration that is 50to 500-fold higher than plasma concentrations. It is not clear from these studies whether the high concentration of uridine was required to offset a high rate of uridine utilization and/or catabolism or to an inability of lower concentrations to satisfy cellular pyrimidine requirements.
In the present study, cultured L1210 cells were exposed to fixed concentrations of uridine by either adjusting the cell number for minimal uridine removal or by infusion into the culture at a rate to offset uridine removal. The data demonstrate that plasma concentrations of uridine can satisfy the cellular requirements for pyrimidines in cells where de novo biosynthesis is blocked; that cells exposed to plasma concentrations of uridine utilize their salvage pathways and turn off de novo biosynthesis; and that the de novo pyrimidine biosynthetic pathway is more sensitive to cellular uracil nucleotide concentrations than is the salvage pathway. These data indicate that circulating uridine is a factor in antipyrimidine cancer chemotherapy and possibly a determinant of the marginal clinical effectiveness so far achieved with inhibitors of de novo pyrimidine biosynthesis. 1630 medium supplemented with penicillin, streptomycin, and 10% fetal calf serum. The stocks were diluted and given fresh media a t least 3 times a week in order to maintain cells in logarithmical growth. All experiments were performed using 20% fetal calf serum unless otherwise stated.
Quantitation of Media Uridine-To 0.5 to 2.0 ml of media, 10 nmol of the internal standard 5-methylcytidine were added and the sample volume was brought up to 2 to 3 ml with water. The samples were then centrifuged through Amicon Centriflo CF25 membrane cones (Amicon Corp., Cambridge, MA) and the filtrate containing the nucleosides was lyophilized. Radioactive media were treated with barium hydroxide prior to centrifugation to precipitate the excess ["C]bicarbonate as barium carbonate.
The lyophilized samples were redissolved in 200 pl of water and 5 p1 of xanthine oxidase (Grade 111) were added to oxidize xanthine and hypoxanthine. A 100-p1 aliquot was analyzed on an Altex Model 312 high pressure liquid chromatograph equipped with a Whatman Partisil PXS 5/25 ODs-3 column. The samples were eluted with an acetate buffer (0.01 M sodium acetate plus 0.01 M acetic acid, p H 4.5) at 1.7 ml/min. Between each run, the column was washed with 95% methanol and 5% buffer. Uridine was detected a t both 254 and 280 nm and peak heights were used for uridine quantitation (4).
Reversal of PALA Growth Inhibition by Uridine"L1210 cells in conditioned media were distributed in 25-cm2 culture flasks (1300 cells/ml, 5 ml/flask). The conditioned media were prepared by incubating cells (>1 X lo5 cells/ml) overnight, spinning down the cells, and collecting the media. The flasks, in triplicate, were made 2 mM in PALA and 10, 5, 1, 0.5, 0.25, or 0 pM in uridine. At 48 h, the cells were counted and the media uridine concentration was measured.
Uridine Infusion Studies"L1210 cells were placed into 25-cm2 culture flasks (0.5 X lo5 cells/ml, 4 ml/flask) and were incubated overnight in order to deplete the uridine present in the serum. The following morning, solutions of uridine and PALA were sonicated for 5 min in warm water to prevent air bubbles from forming in the infusion tubing. All remaining procedures were performed in a 37 "C warm room. Twenty-four disposable 1-ml syringes were filled through a 23 gauge, 1-inch needle. With a 20-gauge needle, a hole was made in 24 tissue culture flask caps of the soft plastic variety. Using tweezers, a 30-cm piece of PE50 tubing was threaded through the cap hole leaving 10 cm on the back side of the cap. The tubing was then attached to the syringe needle and the syringes were set into a modified infusion pump rack capable of holding multiple syringes. The infusion pump plate was pushed up behind the syringes filling the PE50 tubing with uridine solution, and the pump motor was turned on. PALA (80 p1 of a 5 mM solution, in H20 neutralized with NaOH) and uridine (to give an initial concentration of 2.5 to 10 pM) were added to 12 flasks and uridine alone was added to another 12 flasks. These 24 flasks were attached to the infusion pump ensuring that the end of the tubing was resting in the media. A set of noninfused, PALA-treated (100 p~) flasks was also prepared. All uridine solutions were in H 2 0 and the infusion pump (Harvard Apparatus Compact Infusion Pump Model 975, Harvard Apparatus Company, Inc., South Natick, MA) was run on the slowest speed throughout the entire 72-h period delivering 4.8 pl/h. The cell number/flask and the media uridine concentration were determined a t 0, 4, 24, 48, and 72 h, in triplicate.
Measurement of de Novo Pyrimidine Biosynthesis"L1210 cells which had been grown for a t least 24 h in RPMI 1630 media with 10% fetal calf serum were distributed into fifteen 15-ml culture tubes (5 ml/tube, 7 X lo5 to 1 x IO6 cells/ml). The tubes were immersed horizontally in a 37 "C shaking water bath. After 2 h of incubation, all tubes were made 2 mM in L-glutamine and 100 pl of a 0.5 mM uridine solution were added to all but the control tubes. The culture tubes were connected to the infusion pump as described in the previous section with the control tubes being infused with only PBS and the remaining tubes being infused with uridine in the range of 1 to 6 mM in PBS. The pump was run at a speed which delivered 7.2 or 10.8 pl/h. After 1 h of continued incubation, 3 tubes each of the different infused uridine concentrations were disconnected from the pump, the cells were pelleted at 200 X g for 10 min, and 2 ml of the tration. At the same time, 25 pCi (50 p1 in HZO) of NaH14C03 were media were removed for measurement of the media uridine concenadded to the remaining tubes which were then reattached to the pump. After incubating with NaH14C03 for 1 h in the shaking water bath, the tubes were centrifuged for 10 min at 200 X g and 2 ml of the media were saved for analysis. The remainder of the media was poured into a waste bottle containing barium hydroxide. The cells were washed with 5 ml of PBS and recentrifuged. The cells were lysed in 1 ml of H20 and precipitated with 0.8 ml of a cold 10% trichloroacetic acid solution, and 10 nmol of 5-methylcytidine were added. The trichloroacetic acid precipitate was centrifuged a t 800 X g for 15 min and the aqueous layer was collected. Excess trichloroacetic acid was removed by vortexing the aqueous layer with 2 ml of a trich1orotrifluoroethane:tri-n-octylamine solution (3:l). The aqueous layer was transferred to a vial and incubated overnight a t 37 "C with phosphodiesterase I (Type VI) and alkaline phosphatase (Type I) to convert all uracil nucleotides and uridine diphosphate esters to uridine. The lyophilized samples were analyzed by high pressure liquid chromatography in the same manner as for the measurement of media uridine concentrations. Each run was collected in fractions and counted.
Measurement of Uridine Saluage"L1210 cells a t 6 X lo5 cells/ml were distributed into 15-ml culture tubes (5 ml/tube) and incubated in a shaking 37 "C water bath with 400, 100, 20, and 0 p~ uridine. After 2 h, the cells were washed twice with media and incubated in   5 p~, 0.25 pCi). The cells were then centrifuged, the pellet was resuspended in 2 ml of H20, the sample was filtered through DE81 filter disks, and the disks were counted. An aliquot of cells taken just prior to the final incubation was analyzed to determine the size of the uracil nucleotide pool.

RESULTS
Effect of Uridine on the Inhibition by PALA of L1210 Cell Growth-The data in Table I show that plasma concentrations of uridine (1 to 10 p~) are ineffective in rescuing cultured L1210 cells from growth inhibition by PALA. When  (Table I). These data are similar to the observations of Swyryd et al. (7) and Tsuboi et al. (8) who found that 100 p~ uridine was required to reverse the growthinhibitory effects of PALA in cultured SV40-transformed baby hamster kidney cells and in cultured HT-29 human colonic epithelial cancer cells, respectively. In each of the above studies, the uridine concentration required to maintain normal growth is at least lO-fold higher than plasma concentrations. However, depletion of uridine from the culture media may explain the high concentration of uridine required in the experiment detailed in Table I.
When L1210 cell cultures at 1 to 3 x lo5 cells/ml were made 100 p~ in uridine, uridine disappeared from the media a t a rate of 0.3 to 1.0 nmol/h/l@ cells during the first 24 h of growth. By 48 h, media uridine concentrations fell below 0.2 FM. Addition of 100 p~ PALA did not significantly affect the rate of depletion of media uridine during the first 24 h when the initial media uridine concentration was 100 p~. Uridine was not detectable in the media (<0.2 p~) after 24 h of incubation when the initial concentration of media uridine was 10 pM or less. Less than a 5% loss of uridine was detected when 20 pM uridine was incubated at 37 "C in RPMI 1630 media. Fetal calf serum was found to be approximately 1 p~ in uridine and the presence of 20% fetal calf serum did not alter the loss of media uridine.
The addition of tracer amounts of [I4C]uridine to the media of L1210 cells demonstrated 67 to 85% of the radiolabel was retained in the media after 24 h. By collecting and counting the uridine peak from the high performance liquid chromatography analyses of the media, it was determined that the specific activity of the media uridine remained unchanged after 24 h of incubation. Since the specific activity of uridine remains unchanged and the amount of media uridine at 24 h is approximately 25% of its original level, the majority of the radiolabel retained in the media after 24 h is no longer in the form of [14C]uridine. No significant amounts of radiolabeled uracil or dihydrouraeil were found in the media.
Since L1210 cells at high density (1 X lo5 cells/ml) rapidly deplete uridine from the media, the uridine rescue of PALAtreated L1210 cells was examined at low cell density to minimize the loss of uridine from the media. L1210 cells at 1300 cells/ml were incubated with 2 mM PALA and 0.25 to 10 p M uridine for 48 h (Fig. 1). Media uridine concentration measurements at 48 h revealed that cultures initial11 10 pM in uridine were 6.2 p~ k 0.6 (S.D.) and those initially 5 pM were 1.7 p~ f 0.2 (S.D.) in the absence of PALA and were 5.7 p~ k 0.4 (S.D.) and 1.5 p~ 0.2 (S.D.), respectively, in the presence of PALA. No uridine (<0.2 p~) was detected at 48 h in the cultures in which the media uridine was originally 1 plw or less. The results in Fig. 1 demonstrate that even at the lowest concentrations of uridine, the 48-h growth of PALAtreated cells was almost completely rescued at 48 h. Thus, at low cell density, plasma concentrations of uridine will maintain normal growth in cultures when de novo synthesis is blocked.
In order to examine the effect of constant levels of uridine on the growth inhibition caused by PALA of L1210 cells at high cell density, it was necessary to infuse uridine into the cultures to offset the removal of media uridine. An infusion system was developed to supply uridine to the cultured cells at a rate that approximated the rate of uridine depletion by the cells. Although it was not possible to maintain constant uridine concentrations, the infusion system maintained uri-dine concentrations close to those found in plasma (Fig. 2). The rate of consumption of infused uridine generally correlated with the rate of cell division between each time point. In the uridine infusions experiments, it was found through experimental trial to be necessary to infuse more uridine into the PALA-treated cultures than into the untreated cultures in order to maintain the media uridine concentration in the PALA-treated cultures close to the level of the untreated cultures. The results in Fig. 2 show that infusion of L1210 cultures with plasma concentrations of uridine (1 to 10 p~) reverse the growth-inhibitory effects of PALA.
Cadman and Benz (11) determined the minimum cellular requirement of exogenous uridine necessary to maintain normal cell growth in cells where de nouo pyrimidine biosynthesis had been inhibited by pyrazofurin and reported that the amount of uridine required for cell division for L1210 cells under these conditions is 53.3 fmol/cell. The 4 experiments shown in Fig. 2  Effect of Exogenous Uridine on de Novo Pyrimidine Biosynthesis-The effect of exogenous uridine on de nouo pyrimidine biosynthesis in L E 1 0 cells was examined by measuring the incorporation of ['4C]bicarbonate into uracil nucleotides at different media uridine concentrations. These experiments were carried out using 7 X lo5 to 1 X lo6 cells/ml. At 1 X lo6 cells/ml, uridine depletion from the media was found to be 12 to 17.5 nmol/ml/h at an initial uridine concentration of 20 to 50 p~. In order to maintain relatively constant media uridine levels, cell cultures were infused with uridine as described under "Experimental Procedures." Effect of Uridine on PJ have no effect on de novo pyrimidine biosynthesis in L1210 cells, while media uridine concentrations in the range of 3 to 5 FM generally produce approximately a 50% inhibition of de novo pyrimidine biosynthesis and media uridine concentrations above 12 PM cause a >95% inhibition of de novo pyrimidine biosynthesis. Thus, in the range of plasma uridine concentrations (2 to 12 NM), the activity of the de nouo pathway in cultured L1210 cells is decreased by 15% to more than 95%. The inhibition of de novo pyrimidine biosynthesis by uridine is related to the expansion of the uracil nucleotide pool. The data in Fig. 4 show that the de nouo pathway is inhibited by 95% when the uracil nucleotide pool is doubled.  Fig. 3 were measured as described under "Experimental Procedures." The size of the intracellular uracil nucleotide pool of cells exposed to uridine relative to the uracil nucleotide pool size of cells not exposed to exogenous uridine is plotted on the abscissa such that an expansion of 1.5 represents a 50% increase in the uracil nucleotide pool size. In 13 control samples, the uracil nucleotide pool level was 2.0 nmol/106 cells f 0.2 (S.E.).

TABLE I1
Effect of expansion of uracil nucleotide pool on the incorporation of p'C]uridine L1210 cells at 6 X lo5 cells/ml were preincubated with uridine to expand the uracil nucleotide pool. An expansion of 1.5 represents a 50% increase in the uracil nucleotide pool size. The cells were then washed, incubated with [U-"Cluridine, and the amount of phosphorylated uridine was measured as described under "Experimental Procedures." The control level of uracil nucleotides was 1.12 f 0.16 nmol/ lo6 cells (mean + S.E. of 6 samdes). Expansion of the intracellular uracil nucleotide pool was associated with a decrease in the phosphorylation of ["C] uridine by L1210 cells. When the uracil nucleotide pool was expanded 2-fold or greater, the phosphorylation of [I4C]uridine was reduced by 41 to 48% (Table 11). The de novo pathway is much more sensitive to the intracellular uracil nucleotide pool size than is the salvage pathway since a doubling of the pool was associated with inhibition of de novo biosynthesis by >95% (Fig. 4) but [14C]uridine phosphorylation was inhibited by only 40 to 50%. In addition, the uracil nucleotide pool continued to increase beyond the size necessary to inhibit de nouo pyrimidine biosynthesis by >95%.

DISCUSSION
Mammalian cells have two separate pathways to maintain cellular pools of pyrimidine nucleotides: a de nouo pathway and a salvage pathway by which they can utilize preformed nucleosides or bases (17). The relative dependencies of tissues on de novo versus salvage pathways for generating pyrimidine nucleotides in vivo is unknown. The cat brain has shown a requirement for preformed pyrimidines (18) and mature erythrocytes lack the first four enzymes of the de nouo pyrimidine pathway (19), indicating that these tissues must rely on a salvage mechanism for maintaining intracellular uracil nucleotide pools. However, for other mammalian tissues that have both de novo and salvage pathways, the balance between the two pathways could be determined by the availability of salvageable nucleosides. Weber and co-workers have measured an increase in the specific activities of the enzymes of both de nouo and salvage biosynthesis isolated from human colon carcinoma uersus normal mucosa (20) and from rat sarcoma as compared to skeletal muscle (21). In order to study the salvage of plasma uridine in the healthy rat, Moyer et al.
(22) infused [5-3H]uridine into the rat intravenously to provide a steady state concentration of circulating labeled uridine and found that >70% of the circulating uridine is catabolized rather than salvaged. Direct determination of the simultaneous flux through the de nouo and salvage pathways has not been done in uiuo because of the lack of analytical technology. A pertinent question that can be answered with existing methodologies is whether or not circulating concentrations of preformed pyrimidines are adequate to satisfy cellular pyrimidine requirements. In the present study, we duplicated, i n uitro, concentrations of uridine found in plasma to determine if these concentrations are adequate to maintain cellular growth.
Previous studies in our laboratory have shown that plasma uridine concentrations in man, mice, and rats are in the range of 1 to 10 PM (4). Uridine concentrations remain within this range throughout the day and are unaffected by a 24-h fast. Previous studies have demonstrated that >lo0 FM uridine is necessary to rescue cultured cells from growth inhibition by PALA (7,8). Accordingly, Table I (Figs. 1 and 2). Thus, the concentration of uridine in plasma is sufficient to maintain cellular uracil nucleotide pools when cells are forced to use their salvage pathway, as in the presence of PALA.
The data in Fig. 3 demonstrate that L1210 cells incubated in the presence of physiologic concentrations of uridine, in the absence of PALA, utilize their salvage pathway in preference to their de novo pathway. The decrease in de novo synthesis in the presence of plasma concentrations of uridine is directly related to the size of the uracil nucleotide pool (Fig.  4). UTP is known to be an allosteric inhibitor of both carbamyl phosphate synthetase I1 (the first enzyme of the de novo pathway) and uridine/cytidine kinase (12-16). UTP inhibition of carbamyl phosphate synthetase I1 and uridine/cytidine kinase has never been compared in the same system, so the relative affinity of UTP for the 2 pathways cannot be assessed from values reported in the literature. Anderson and Brockman (16) reported 84% inhibition of uridine/cytidine kinase from P815 mouse tumor cells when both uridine and UTP were a t a 10 mM concentration. Carbamyl phosphate synthetase 11, isolated from rat liver (15), rat hepatomas (12, 15), and mouse spleen (13,14) is inhibited by 70 to 90% when the ATP to UTP ratio (each at millimolar concentrations) ranged from 4:l to 1:l. LevinP tal. (13) reported a K, of 0.11 mM for UTP for mouse spleen carbamyl phosphate synthetase I1 when ATP varied between 1 and 25 mM and UTP ranged from 0 to 4 mM. Our data demonstrate that in intact L1210 cells the de novo pathway is considerably more sensitive to intracellular uracil nucleotides than is the salvage pathway. When the uracil nucleotide pool is doubled (Fig. 4) and de novo synthesis is nearly completely inhibited (Fig. 3), the uracil nucleotide pool continues to expand in the presence of exogenous uridine and the salvage of [14C]uridine is inhibited by less than 50% (Table 11).
Since the concentration of uridine in the plasma is adequate to satisfy the requirements of L1210 cells for uracil nucleotides and since L1210 cells preferentially utilize their salvage pathway over their de novo pathway when exposed to plasma concentrations of uridine, plasma uridine may be an important factor in antipyrimidine chemotherapy. It would appear that the highly vascularized areas of a tumor are exposed to a concentration of uridine sufficient to circumvent growth inhibition by inhibitors of de novo biosynthesis. These data support the development of inhibitors of uridine/cytidine kinase as agents for use in combination with inhibitors of de novo pyrimidine biosynthesis in the treatment of cancer.