5-fluorouracil modulation of dihydrofolate reductase RNA levels in methotrexate-resistant KB cells.

Cytotoxicity and growth inhibition by 5-fluorouracil in methotrexate-resistant dihydrofolate reductase gene-amplified KB cells in the presence of 30 microM thymidine correlates with incorporation of this fluorinated pyrimidine into RNA. Growth of these cells over several generations in the presence of inhibitory concentrations of 5-fluorouracil does not depress the steady state levels of either 18 or 28 S RNA but actually causes an increase. Similarly the rates of RNA and protein synthesis in 5-fluorouracil-treated cells are not decreased. The level of dihydrofolate reductase RNA from 5-fluorouracil-treated cells increases in a dose-dependent manner correlated with 5-fluorouracil incorporation into RNA. The qualitative size distribution of the dihydrofolate reductase RNA species is unaffected when examined by the Northern blotting technique indicating an RNA processing lesion is not induced by 5-fluorouracil incorporation into RNA. As the dose of dihydrofolate reductase RNA increases, there is no change in the level of dihydrofolate reductase specific activity, but the level of enzyme activity per cell increases. The relevance of these phenomena to the mechanism of 5-fluorouracil effect on RNA and relevance to combination chemotherapy with methotrexate are discussed.

5-Fluorouracil is a pyrimidine analog which is used alone or in combination for the treatment of several solid human tumors (1)(2)(3)(4). Many laboratories have studied FUra's' mechanisms of cytotoxicity (5). It is clear that in many systems, metabolism to FdUMP results in blockage of DNA synthesis through inhibition of thymidylate synthetase (5)(6)(7). Under certain conditions, FdUTP is formed and can be incorporated into DNA, an effect which is blocked by low concentrations of dThd (8).
In many in uitro and in uiuo systems, intracellular conversion of FUra to 5-fluorouracil ribonucleoside triphosphate results in incorporation of FUra residues into RNA. This incorporation has been correlated with cytotoxicity in tissue culture (9)(10)(11) and gastrointestinal mucosa (12). The effects of FUra incorporation into RNA on the different classes of RNA vary and their relationships to cytotoxicity are controversial. In the case of rRNA, inhibition of 45 S rRNA pro-' The abbreviations used are: FUra, 5-fluorouracil; FdUMP, 5-fluoro-2'-deoxyuridine 5'-monophosphate; FdUTP, 5-fluoro-2'-deoxyuridine 5'-triphosphate; H,PteGlu, dihydropteroylglutamic acid. cessing to mature 18 and 28 S RNA occurs with the administration of high concentrations of FUra (9,13,14) or by FUra in the presence of the metabolic modulator inosine (15). In the case of Escherichia coli tRNA, substitution of 90 to 95% of uridine and uridine-derived nucleosides with FUra leads to the substantial inhibition of aminoacylation of several tRNA species (16). In contrast, no effects were observed in tRNA where only 50% replacement had taken place (17). Incorporation of FUra residues into mRNA can cause miscoding during translation in E. coli (18) and homopolymers translated in uitro (19). The effect of FUra incorporation on mammalian mRNA is not well defined, in part due to the heterodisperse nature of poly(A) RNA. Translation of mammalian poly(A) RNA using an in uitro translation system demonstrated an enhanced degree of protein synthesis and the appearance of some high molecular weight protein species when poly(A) RNA isolated from the regenerating liver of FUra-treated rats was used as a messenger source (20), a result which can be interpreted as being due to FUra incorporation into RNA or the result of different mRNA synthesis. The interpretation of these results is important to the understanding of the RNAmediated effects of FUra and its metabolites. However, interpretation is limited by the lack of a single mRNA species to study as a model for the effect of FUra on this class of RNA.
This study was initiated to examine the effects of FUra on rRNA and a particular mRNA class under conditions of pharmacologic relevance. A clonally selected stable gene-amplified human cell line (KB7D) which overproduces dihydrofolate reductase and its mRNA has been employed to monitor FUra effects on a single class of mRNA (21)(22)(23). This report demonstrates a previously unobserved phenomenon, i.e. an increase in a particular mRNA species as a result of cell growth in the presence of FUra. In addition, this report demonstrates that cell kill by low levels of FUra during long term exposure correlates with its incorporation into RNA in the absence of a depression of the levels of 18 and 28 S rRNA and protein synthesis. The relevance of these observations to FUra's mechanism of action, clinical significance, and previous observations is discussed.

MATERIALS AND METHODS
Chemicak-DNase, RNase A, dThd, NADPH, methotrexate, and FUra were obtained from Sigma. FUra was obtained from Calbiochem-Behring; all other nucleosides and nucleotides were obtained from P-L Biochemicals. Poly(U) Sephadex, CsCl (optical grade), urea and XDNA were obtained from Bethesda Research Laboratories. Nitrocellulose (BA 85) was obtained from Schleicher & Schuell. H&e.Glu was synthesized and purified by the method of Blakely (24). Hind111 and DNA polymerase I were obtained from BioTec. [S-3H]FUra ( Cell Culture-KB and the methotrexate-resistant dihydrofolate reductase gene-amplified KB7D subline were maintained in RPMI 1640 medium supplemented with 5% dialyzed horse serum, 40 pg/ml of garamycin, and 30 p~ dThd at 37 "C in a 5% CO, atmosphere (22). KB7D cells were carried in the above medium with an additional 75 p~ methotrexate. Cells were plated in the absence of methotrexate for experiments unless otherwise indicated. KB and KB7D were passaged twice weekly by detachment with pancreatin and a subsequent 4-fold dilution in fresh medium. The doubling time for KB cells is 24 h and 30 h for KB7D.
For the estimation of growth inhibitory effects of drugs, cells were plated in duplicate vessels at 6.67 X lo" cells/cm' in complete medium (plus or minus dThd as indicated in the text) and allowed to attach at 37 "C overnight. After 24 h drug was added to the indicated concentration. Cells were allowed to grow for 6 days (an average of six cell doublings for KB cells and four to five for KB7D) a t which time they were harvested and counted to determine the extent of cell growth. For some experiments cytotoxicity by cloning efficiency was estimated by plating 100, 200, and 300 cells in 2 ml of fresh medium (minus drug) in duplicate 4.5-cm' wells. Cells were allowed to grow 14 to 21 days at which time colonies were stained with 2% (w/v) were deemed viable, cloning efficiency was 50 to 70% for KB and 20 methylene blue in methanol. Colonies containing a t least 50 cells to 30%: for KB7D cells. All tissue culture experiments were repeated at least twice with similar results.
Preparation and Gel Analysis of RNA-Total cellular RNA was prepared essentially by the urea extraction and CsCl discontinuous gradient centrifugation method of Ross (25) with minor modifications. RNAs were washed twice with 3.0 M sodium acetate, pH 4.5 (4 "C), twice with 70% ethanol, 100 mM NaCI, and finally once with 100% ethanol. The RNA was then dessicated, dissolved in sterile distilled water, and stored a t -70 "C. Yields were typically about 15 mg per 10' cells, using 50 pg of RNA per to determine amounts. RNA was analyzed by denaturation with deionized glyoxal and electrophoresis on 0.5-cm thick 1.4% agarose gels submerged in 10 mM sodium phosphate buffer, pH 7.0, with buffer recirculation as described (26). Typically 20 pg of total cellular RNA was used per lane. Gels were stained for 45 min with 10 pg/ml of acridine orange and destained on an enameled white pan for 2 to 3 days. The destained gels were photographed using a long wave UV transilluminator and Polaroid type 52 film. XDNA cut with Hind111 and labeled with [ a -,"P]dCTP using Klenow DNA polymerase was denatured by glyoxal treatment and included in all gels as a molecular weight standard and a marker for transfer efficiency during Northern blotting procedures (27, 28).
Species of RNA-containing regions of dihydrofolate reductase mRNA homology were identified by a modification of the Northern blotting procedure (28). Transfers were performed a t 4 "C for 72 h, which allows complete transfer. The filter was baked for 2 h a t 80 "C and prehybridized for 24 h as described by Wahl et al. (29) except the prehybridization buffer contained 0.6% (w/v) polyvinyl pyrrolidone.
The filter was then hybridized with lo6 cpm/ml of nick translated pDHFR26 (30) in the presence of 20 pg/ml of poly(rC) and hybridization buffer as described previously (27). pDHFR26 is a pBR322derived plasmid containing a 1.4-kilobase pair cDNA insert prepared from murine dihydrofolate reductase mRNA. It contains all the coding region and about 900 base pairs representing the 3"noncoding region of murine dihydrofolate reductase mRNA (30). Nick translation was employed to generate pDHFR26 labeled with [tv-:"P]dCTP to a radiospecificity of 1 to 2 X lo8 cpmlpg (31). RNA bands containing dihydrofolate reductase mRNA homology were visualized by exposing Kodak XRP-1 film to the washed filters a t -70 "C with Cronex Quanta I1 intensifying screens.
Quantitation of Dihydrofolate Reductase RNA-Dihydrofolate reductase RNA was quantitated by a modification of the dot blot method of Kafatos et al. (32). RNA was dissolved in 3 M NaC1, 0.15 M sodium citrate, pH 7.0. The RNA was spotted on a sheet of nitrocellulose presoaked with the same buffer using a Bethesda Research Laboratories Hybri-dot manifold to apply the RNA in 3-mm2 circles. In some experiments, 1 to 10 pg of tRNA were added to determine if the presence of a carrier was required to obtain linear results with the amount of RNA applied. Adding additional tRNA had no effect on the outcome ofexperiments. After applying the RNA baked for 2 h a t 70 "C, prehybridized and hybridized as described samples (amounts as indicated in the figures) the nitrocellulose was above. The resulting autoradiogram was scanned a t 600 nM using a Gilford 2600 microprocessor controlled spectrophotometer with a densitometer scanner-integrator attachment. The area under the absorbance curves was integrated and plotted as a function of micrograms of RNA applied. Results used for quantitation derived from areas of the curve which were linear with amounts of added RNA. The slope of these curves was used to estimate relative amounts of dihydrofolate reductase RNA.
Enzyme Assay and Preparation-Dibydrofolate reductase was assayed spectrophotometrically by the method of Osborne and Huennekens (33). Reaction mixtures contained in a final volume of 1 ml: 20 p M H,PteGlu, 100 p M NADPH, 150 mM KCI, 100 mM Tris-HCI, pH 7.5, and enzyme. Reactions were preincubated for 2 min a t 37 "C prior to initiation by the addition of H,PteGlu. All activities were linear with time and enzyme concentration. One unit of enzyme activity is defined as the amount of enzyme necessary to form 1 pmol of product per min. Cells were extracted for dihydrofolate reductase as described previously (34) except the cell sonicate was centrifuged at 110,000 X g a t 4 "C for 30 min to generate a cleared supernatant.

RESULTS
Our purpose was to utilize a cell line which overproduces a particular mRNA species to study the effects of FUra on rRNA and mRNA. To this end we determined some growth inhibition parameters of FUra under conditions where most of FUra's action could be attributable to incorporation into RNA. The IDso of FUra for KB and KB7D cells was determined after different times of continuous exposure. This method was chosen to determine a concentration range where FUra produces inhibition and an RNA complement which has a steady state percentage substitution of its pyrimidine components as FUra derivatives. Determinations of growth inhibitory effects were carried out in the presence of 30 PM dThd to circumvent possible inhibition of thymidylate synthetase by FdUMP biosynthesis as a mechanism of cell kill. Preliminary experiments indicated that 30 PM dThd promoted growth inhibition by FUra in both cell lines but did not inhibit growth itself (Fig. 1).
KB and KB7D cells were inoculated into duplicate wells of a Linbro dish (see under "Materials and Methods") and allowed to attach overnight. On the following day (day l), FUra was added to the concentrations indicated ( Fig. 2 ) . Dishes were harvested daily on days 3 to 6 (KB) or 4 to 7 (KB7D) and the cells were counted. The results presented in Fig. 2 demonstrate the similarity of the IDso of FUra in KB and KB7D cells. Cell viability, determined by the cloning efficiency of drug-treated cells yields similar results to the growth inhibition experiments (see Fig. 3). Fig. 2 also demonstrates that in KB and KB7D cells sensitivity to FUra remains unchanged during cell growth from mid-to late log (ie. 3.9 to 5.5 doublings for KB and 3.8 to 5.3 doublings for KB7D). To determine whether growth inhibition and cytotoxicity by FUra was associated with its incorporation into RNA, cells were grown in the presence of ["HIFUra. After 6 days of growth in medium containing various concentrations of ["HIFUra, cells were harvested and counted. A fraction of the cells was plated to determine viability and the remainder was extracted to determine the incorporation of ['HH]FUra into the acid-insoluble fraction. As Fig. 3 indicates, growth inhibition and cytotoxicity correlate well with each other and with the degree of FUra incorporation into the acid-insoluble fraction ( i e . RNA) (5). Net counts per min incorporated per cell were linear over the range studied (data not shown).
To determine whether there was any effect of FUra on the steady state level of dihydrofolate reductase RNA in cells grown in the presence of FUra, KB7D cells were seeded into flasks and incubated for 6 days in the presence of various concentrations of FUra. At this time the cells were harvested, counted, and RNA was prepared. The RNA was dot blotted on nitrocellulose and hybridized against '"P-labeled nick ing the area under a densitometric tracing of the autoradiogram and plotting the area as a function of micrograms of RNA. Slopes were taken from the linear portion of these curves, and the ratio of slopes from the drug versus control cells is taken as an estimate of the relative concentration of dihydrofolate reductase. As Fig. 5 indicates, the level of dihydrofolate reductase RNA appears to correlate inversely with inhibition of cell growth (up to 50% growth inhibition) reaching a peak change of approximately 3-fold.
It has been demonstrated by many researchers that FUra can inhibit maturation of 45 S precursor rRNA to the mature 18 and 28 S forms. It was anticipated that this phenomenon could be responsible for the observed increase in dihydrofolate reductase RNA. This might be expected as poly(A) RNA is present in low abundance compared to rRNA so a decrease in cellular rRNA would give an apparent increase in dihydrofolate reductase RNA (assuming that most of this RNA is polyadenylated) on a dot blot when total RNA is being used. We, therefore, grew cells in the presence of various concentrations of FUra (6 days growth in the presence of drug), prepared RNA, and analyzed the RNA by electrophoresis in agarose gels under denaturing conditions. A typical photograph of a UV-illuminated gel which has been stained with acridine orange is presented in Fig. 6. As Fig. 6  RNA extracted from cells grown for 6 days is about 3.6% for the control cells and about 19% for RNA from cells growth inhibited 50% (based upon 4.8 cell doublings for control and 2.4 for drug-treated cells). Therefore, the increase in dihydrofolate reduct.ase RNA is not due to a proportional decrease in cellular rRNA. This estimate is based on the assumption that RNA, respectively. Columns were loaded with 5 X lo5 dpm of ['HIRNA and 5.4 X 10' dpm of ["CIRNA. The number in parentheses represents the ratio of these two percentages. the amount of RNA per cell has not decreased significantly in drug-treated cells. In fact, in all experiments where the RNA recovery per cell has been estimated, the average recovery of RNA from control cells is found to be less than that of drug-treated cells (see Table I), indicating that the estimate of original RNA as a percentage of the RNA shown in Fig. 6 is an upper estimate.
A gel similar to that portrayed in Fig. 6 was blotted to nitrocellulose as described (see under "Materials and Meth-  Fig. 6. 10 pg of each RNA were denatured and electrophoresed as described (see under "Materials and Methods"). The gel was blotted to nitrocellulose and hybridized with the "P-labeled probe for dihydrofolate reductase RNA. The autoradiogram is labeled as described for Fig. 6 except the species of dihydrofolate reductase RNA (kilobases) are indicated in the left margin and the positions of the XDNA markers on the right (kilobases).
ods") and hybridized with ["P]pDHFR26. The autoradiogram of the hybridized filter is shown in Fig. 7. Fig. 7 shows that the RNA from control or FUra-treated KB7D cells has discrete RNA with homology to the murine dihydrofolate reductase cDNA probe. The sharp 3.5-and 1.0-kilobase bands observed in RNA from control cells is similar to that reported for HeLa cells (35). From this experiment it seems the pattern of discrete dihydrofolate reductase mRNAs (35) has not been qualitatively altered by treatment of cells with FUra. However, Fig. 7 shows that the RNA from cells grown in progressively higher concentrations of FUra displays a progressively higher 3.5-kilobase signal when hybridized with [,"PI pDHFR26. This change closely reflects the change seen by dot blot analysis as determined by densitometric tracing; while the 3.5-kilobase RNA changes substantially in amount, the 1.0-kilobase RNA does not change significantly until the 2.0 p M FUra concentration.
As we were using total cellular RNA (minus tRNA) to   Table I. Table I shows that the amount of RNA recovered from drug-treated cells is higher than controls. (Increases in RNA from drug-treated cells are found to be elevated routinely; increases range from 60 to several hundred per cent depending on the amount of drug used.) Poly(U) Sephadex chromatography was carried out by mixing [I4C]Urd-labeled RNA from control cells with either ['HIGuo-labeled RNA from control or FUra-treated cells to serve as an internal standard. As Table I indicates the per cent of RNA recovered as poly(A) RNA is not significantly different in control uersus FUra-treated cells.
Since previous reports have implicated processing of 45 S precursor rRNA as the lesion associated with RNA-mediated

TABLE I11
Effect of FUra on dihydrofolate reductase actioity KB7D cells grown 6 days in the presence of various concentrations of FUra were harvested and extracted as described (see under "Materials and Methods"). The enzyme activity and protein were determined (see under "Materials and Methods").

Enzyme activity FUra
Growth Units/gl (x 109 Units/&g Units/cell equiv- FUra toxicity, we decided to measure protein synthesis in drug-treated cells. This was particularly important since 1) we had not observed a change in the distribution of 18 and 28 S RNA from steady state drug-treated cells and 2) gel analysis of radiolabeled RNA from ["HIGuo-pulsed drug-treated and control cells (Table I) showed the loss of 18 S RNA and increase in 45 S RNA pattern as others had observed (data not shown, see Refs. 5, 9, 13, and 15). We reasoned that if protein synthesis was not significantly impaired in FUratreated cells, then the inhibition of 45 S RNA processing observed over the short term pulse of cells grown for several generations in FUra was probably not a mechanism of cell death. Table I1 shows an experiment which measures the amount of ["Hlleucine incorporated in control and drugtreated cells as a function of FUra concentration and time.
The amount of ["Hlleucine incorporated into protein on a per cell basis increases in a dose-dependent fashion over the concentration range of FUra shown. Thus, protein synthesis is not inhibited in FUra-treated cells, but is actually increased. The increase in dihydrofolate reductase RNA per p g of cellular RNA was an indication that the level of enzyme activity may be increased in FUra-treated cells. This was expected since protein synthesis in FUra-treated cells was not depressed. Cells grown for 6 days in the presence of various concentrations of FUra were extracted for dihydrofolate reductase enzyme activity determinations (see under "Materials and Methods"). A constant volume of extraction buffer to cell number was used so as to be able to relate activity to a cellular value. The results are presented in Table  111. As Table 111

DISCUSSION
FUra has been demonstrated over the years to be a drug with multiple mechanisms of action (5). In mammalian tissue culture systems where growth inhibition and/or cytotoxicity correlate with the incorporation of FUra into RNA, the primary mechanism of action has been ascribed to the inhibition of 45 S precursor rRNA maturation (5,9,13). Other effects on RNA have been reported, such as effects on in vitro translation rates of mRNA with FUra substitution as well as a perturbation in the size distribution of the products synthesized using in vitro translation syst,ems (20). However, no studies of FUra effect on a particular mRNA species related idrofolate Reductase R N A Levels to function or cellular levels have been reported for a mammalian cell line. Additionally, most studies on the effects of FUra on RNA metabolism and function have approached this problem from the standpoint of short term drug exposure (Le. less than one cell generation time). These studies generally require higher levels of drug than do experiments involving long term growth in the presence of drug to attain equivalent cell kill and growth inhibition. We have studied cells which have been treated with concentrations of FUra which inhibit cell growth over a long term period (i.e. greater than 2 generations) to achieve a steady state level of FUra substitution in RNA and to avoid possible epiphenomena which may be associated with high drug levels (e.g. possible inhibition of some RNA processing enzyme(s)). We have attempted to monitor a particular class of mRNA ( i e . dihydrofolate reductase RNA) which codes for the target enzyme of a drug (i.e. methotrexate) whose cytotoxic action is known to be affected by Furta. This study is the first to report an effect of FUra on the levels of a mRNA coding for a specific enzyme. Studies were carried out in the presence of 30 p~ dThd (see Fig. 1) in an effort to eliminate the inhibition of thymidylate synthetase by FdUMP as a mechanism of growth inhibition and thus channel growth inhibitory effects to the incorporation of FUra into RNA.
The results presented here show that in the gene-amplified methotrexate-resistant cell line KB7D, growth inhibition and cytotoxicity correlate with the incorporation of FUra into RNA. This RNA-directed effect was attained by the growth of cells in medium containing low concentrations of FUra (0 to 2 pM) and 30 p~ dThd to circumvent a block at thymidylate synthetase. The distribution and integrity of 18 and 28 S RNA from control and FUra-treated cells were found to be indistinguishable as determined by denaturing gel electrophoreesis. Since these RNA preparations were from cells which had grown at least several-fold in the presence of drug (see Fig. 6 and "Results"), the contribution of the RNA from the seeded cell population is calculated to be 19% at most in cells growth inhibited 50%. Experiments presented here (Table I) which indicate the FUra-treated cells contain more RNA per cell than controls suggest that the original cellular RNA probably contributes much less than 19% (even if RNA turnover is not taken into account) to the RNA being analyzed. This result contradicts the conclusion that the RNA-mediated cytotoxic effect of FUra is due to an inhibition of the processing of 45 S precursor rRNA to 18 and 28 S rRNA (9,13).
Other results from our laboratory confirm the observation that 45 S precursor rRNA maturation is affected but indicate the process is only delayed during continuous exposure to FUra. This conclusion is further substantiated by experiments which show that in cells grown in the presence of FUra, 45 S RNA processing was found to be inhibited during an hour pulse with ["HIGuo, but protein synthesis actually increased in FUra-treated cells (see under "Results" and Table 11). Additionally, there was no change in the per cent of cellular RNA as poly(A) RNA (Table I). Cellular RNA was, therefore, used for dot blots (Fig. 4) and Northern blotting (Fig. 7) as an estimation of the amount of dihydrofolate reductase RNA. This allowed the use of small amounts of cells for RNA analysis. These studies indicated that dihydrofolate reductase RNA increased per pg of RNA in FUra-treated cells (Figs. 4 and 5). Analysis of the dihydrofolate reductase RNA by denaturing gel electrophoresis, Northern blotting, and hybridization with a cloned ["PIcDNA probe indicates the major species of dihydrofolate reductase RNA appear unaltered in terms of molecular weight. Thus, FUra incorporation into RNA does not appear to have affected the processing of these dihydrofolate reductase RNAs which have been shown by others to possess mRNA activity (35). However, there is a progressive increase in the autoradiographic signal from the lanes containing RNA from the cells grown in the presence of increasing concentrations of FUra. This increase occurs mainly in the 3.5 kilobase species and is not observed for the 1.0 kilobase species until 2.0 p M FUra (Fig. 7). Since the percentage of cellular RNA as poly(A) RNA does not change in FUra-treated cells (Table I), this increase in dihydrofolate reductase RNA is not an artifact generated by a loss in rRNA content or an increase in the cellular level of poly(A) RNA. There are several explanations for this increase. It is possible that the increase in dihydrofolate reductase RNA reflects enhanced transcription of the dihydrofolate reductase gene or an enhanced distribution of cells in a phase of the cell cycle where dihydrofolate reductase RNA (and possibly other mRNAs) increases due to growth in the presence of FUra.
FUra has been reported to act at G, when its cytotoxicity is correlated with an RNA effect (36). Since in many cell systems dihydrofolate reductase is elevated in S phase, retardation of a cell in late G, may cause an increase in the mRNA for this enzyme. Increases in the mRNA for dihydrofolate reductase just prior to S phase has been demonstrated previously (37) although these increases may be the result of nuclear RNA turnover and not transcriptional rates (38). This generalized cell cycle phenomenon (i.e. an unequal distribution of FUratreated cells in a pre-S state) as an explanation for the observed increase in dihydrofolate reductase RNA per pg of RNA (Figs. 4 and 5) but without a substantial increase in the level of dihydrofolate reductase enzyme activity per pg of protein (Table 111) is an attractive hypothesis currently under investigation in our laboratory.
The mechanism of cell kill related to incorporation of FUra into RNA is unclear. Cell kill and growth inhibition correlate with incorporation of FUra into RNA in the presence of dThd. However, protein synthesis and steady state levels of rRNA are not decreased even though the short term maturation of 45 S RNA to 18 and 28 S RNA appears to be affected (see under "Results," Tables I and 11, and Fig. 6). We conclude from these observations that the inhibition of 45 S RNA processing is probably an epiphenomenon of the effect of FUra incorporation into RNA not responsible for cytotoxicity in this system (although this may be a contributing factor with high FUra doses). Our interpretation is that the incorporation of FUra into RNA which correlates with growth inhibition of these cells may be due to impairment of mRNA function or aberrant translation of mRNA in general to form miscoded translation products such that the cell must commit excessive resources to synthesize greater than normal amounts of RNA to achieve normal levels of enzyme-mediated cell functions. This is suggested by recent experiments in our laboratory which show an enhanced ability of total cellular RNA from FUra-treated cells to stimulate protein synthesis in an in vitro protein synthesis system.2 Alternative explanations such as an effect on some critical nuclear RNA species are not excluded. This hypothesis (Le. impaired RNA function due to FUra incorporation) is also consistent with the disproportionate increase in dihydrofolate reductase RNA compared to the lack of change in dihydrofolate reductase enzyme activity and cannot be distinguished from a late G1/S block by FUra as an explanation by these studies. The elevated levels of dihydrofolate reductase RNA and enzyme per cell may be a partial explanation for the antagonism observed between FUra and methotrexate during their combined use in certain schedules of tumor treatment. It has ' B. ,J. Dolnick and J . d. Pink, unpublished results. ~~ been demonstrated that in regimens where methotrexate precedes FUra, enhanced tumor kill is observed in vivo (39,40). However, in vitro antagonism is observed when the two drugs are coadministered or when FUra precedes methotrexate (41-43). Some explanations revolve about the thymidylate synthetase site of FUra action (42) and do not account for RNArelated effects. In this report we present evidence that FUra can cause increases in the intracellular levels of dihydrofolate reductase RNA and enzyme activity. These increases may result in the enhanced resistance of FUra-treated cells to methotrexate. Although a maximal elevation of dihydrofolate reductase activity of 36% was observed, removal of FUra may result in an even greater increase if the elevated dihydrofolate reductase RNA contributes to enzyme levels due to a released block in cell cycle progression which allows translation of the accumulated message.
Further studies are necessary to evaluate the change in dihydrofolate reductase RNA as a possible mechanism of FUra antagonism to methotrexate. The functionality of the elevated dihydrofolate reductase RNA is also a question which needs to be answered. These areas are currently under investigation and it is hoped they will provide more insight as to the possible role of FUra incorporation into RNA as a mechanism of this drug's action as well as basic information about RNA function.
Chi Cheng for his gift of KB and KB7D cells.