A Novel Class of Genetic Variants of the L1210 Cell Up-regulated for Folate Analogue Transport Inward ISOLATION, CHARACTERIZATION, AND DEGREE OF METABOLIC INSTABILITY OF THE SYSTEM*

We have isolated stable variants of the L1210 cell exhibiting increased transport inward of the folate analog, methotrexate. These variants show 3- to 14- fold increases in [‘Hlmethotrexate influx compared to parental cells but are unaltered for [‘Hlmethotrexate efflux. This increased influx in each variant is quan-titatively reflected in corresponding elevations in in- tracellular exchangeable levels of drug at steady state, but there is no alteration in membrane potential. The increases in influx are associated with increased values for influx V,, for a system normally transporting reduced folates and the same increase in the amount of a specific binding component at the cell surface. Oth-erwise, values for influx K, and specificity for various folate structures are unchanged. This alteration in [‘HI methotrexate influx is biochemically and genetically stable, since it is expressed in isolated plasma membrane vesicles and is retained during growth in non- selective medium. Following addition of cycloheximide, the same rate of decay of this transport activity (tllz = 126 f 24 to 137 f 26 min) was shown for parental and variant cells. From these results we con-clude that turnover of this transport property occurs in these cells which is genetically regulated. Also, the elevated transport activity inward for this folate ana- log in these variant cells is probably the result of a genetic alteration up-regulating the rate of synthesis of the “putative“ carrier protein itself. The absence of any effect on efflux of [’Hlmethotrexate methotrexate. subopti-mum procedures adequate for the derivation of the desired variants. agreement was derived from measure- ments of cell chloride which varied only from 2.69 22 to 2.98 51 dry the transport experiments were intracellular exchangeable is unbound dihydrofolate reductase. cellular levels of this

We have isolated stable variants of the L1210 cell exhibiting increased transport inward of the folate analog, methotrexate. These variants show 3-to 14fold increases in ['Hlmethotrexate influx compared to parental cells but are unaltered for ['Hlmethotrexate efflux. This increased influx in each variant is quantitatively reflected in corresponding elevations in intracellular exchangeable levels of drug at steady state, but there is no alteration in membrane potential. The increases in influx are associated with increased values for influx V , , for a system normally transporting reduced folates and the same increase in the amount of a specific binding component at the cell surface. Otherwise, values for influx K, and specificity for various folate structures are unchanged. This alteration in ['HI methotrexate influx is biochemically and genetically stable, since it is expressed in isolated plasma membrane vesicles and is retained during growth in nonselective medium. Following addition of cycloheximide, the same rate of decay of this transport activity (tllz = 126 f 24 to 137 f 26 min) was shown for parental and variant cells. From these results we conclude that turnover of this transport property occurs in these cells which is genetically regulated. Also, the elevated transport activity inward for this folate analog in these variant cells is probably the result of a genetic alteration up-regulating the rate of synthesis of the "putative" carrier protein itself. The absence of any effect on efflux of ['Hlmethotrexate in these variants in the face of evidence for increased synthesis of the carrier protein for the system mediating influx of this folate analog is construed as further evidence for the nonidentity of systems mediating each flux that we proposed on the basis of earlier kinetic studies.
Methotrexate and other folate analogs are accumulated in tumor cells by a transport system (reviewed in Refs. 1 and 2) normally mediating accumulation from plasma of the coenzyme, 5-methyltetrahydrofolate. Since transport of folate analogs is a determining factor in their chemotherapeutic effectiveness (reviewed in Ref. 21, knowledge as to the properties of this system is not only of biochemical interest but has pharmacologic significance as well. To date, biochemical stud-ies of this specific membrane property, particularly those which focus on the putative carrier component, have been hampered by the fact that it appears to exist in the plasma membrane at very low levels (3) when compared to systems mediating transport of many other nutrients (4, 5). For this reason we have sought methods for the isolation of tumor cell variant6 which are "up-regulated" for this specific membrane property and, hopefully, have elevated levels of the putative carrier in the plasma membrane. The availability of this type of mutant phenotype would also be of value along with phenotypes exhibiting down-regulation of transport in delineating the mechanism of reduced folate transport in these tumor cells. Similar genetic approaches have been successfully applied (6-11) to the study of mechanisms transporting amino acids and cations in mammalian cells.
We now report on the isolation and properties of three variants of the L1210 leukemia cell which show modest to substantial elevation in folate analog transport inward. These increases are associated with increased values for influx V-, but unchanged values for influx K,,, or specificity for various folate compounds, and the same increase in the amount of a specific binding component at the cell surface. The same kinetic alteration is fully expressed in isolated plasma membrane vesicles and appears to be genetically stable. Data was also obtained showing that this property was metabolically unstable with the same rate of decay of this transport actvity in parental and variant cells treated with cycloheximide. These results suggest that the elevated transport activity inward for this folate analog in these variant cells is a result of an increased rate of synthesis of the "putative" carrier itself.

EXPERIMENTAL PROCEDURES
Rationale for Variant Cell Selectwn-This was based upon our earlier (12, 13) studies of methotrexate-resistant L1210 cells which were reduced in folate compound transport inward and collaterally sensitive to the lipophilic nonclassical antifolate, metoprine. These variants exhibited a 5-fold reduction in influx V, , for methotrexate and reduced folate compounds which partially accounted for their resistance. Although this relatively finite decrease in influx V, . for reduced folate compounds was not growth limiting, when such a compound was used as the sole "folate" source, it was growth limiting in the presence of metoprine. This agent, like methotrexate (reviewed in Ref. 141, blocks regeneration of Hzfolate' from H,folate and, thus, in its presence must be replenished exogenously. Also, since metoprine is highly lipophilic, its entry into tumor cells, in contrast to folate analogs, requires only simple diffusion (15). This deficiency in reduced folate transport served to explain (13) the increased sensitivobserved (13) when only folic acid was in the growth medium (this ity to metoprine, since increased sensitivity of these variants was not folate has (16, 17) a more preferred route of entry into L1210 cells).
We reasoned, therefore, that it might be possible to obtain variants with increased capacity for transport of reduced folate compounds by selecting for metoprine resistance with 5-formyl-H4folate as the sole folate source. This phenotype could be readily identified by examining for increased sensitivity to methotrexate. We also reasoned that the same variant phenotype might be selected during growth in suboptimum concentrations of 5-formyl-H4folate alone. Both procedures proved adequate for the derivation of the desired variants.
Isolation Procedures-(a) By methods described previously (18), L1210 cells were grown in RPMI medium supplemented with dialyzed calf serum and the minimum concentration of dl-5-formyl-H4folate required for maximum growth. Cell suspensions were transferred to the same medium containing an ICm concentration of metoprine. After cell density approached stationary phase of growth, cells were transferred again in this medium. This was repeated until the doubling time of the culture was similar to controls. Cells from drug medium were then cloned three times by plating in drug medium containing 0.3% agar. After 7-10 days of incubation at 37 "C colonies were picked and screened for increased sensitivity to methotrexate using 2.2 p~ folic acid as the sole folate source. Subcultures showing increased sensitivity were examined (2) for methotrexate transport. (b) Cell suspensions grown in the same medium as above were transferred to medium containing dialyzed calf serum and dl-5formyltetrahydrofolate at the EC. concentration. After the cell number approached stationary phase, additional transfers were made until the doubling time was similar to controls. Cloning and screening of subcultures was carried out as described above.
Transport Measurements with Intact Cells-A rapid sampling procedure modified (19) from Plagemann and co-workers (20) was employed during these studies for processing of samples by radioactive scintillation counting. This procedure employed a correction for extracellular radioactivity (17) which yielded the intracellular radioactivity. Values for intracellular water were derived (17) for all of the cell types used in this study and varied from 2.9 k 0.3 to 3.5 3~ 0.4 ml/g dry wt in agreement with values also determined from wet and dry weight determinations (2). The derivation of values for membrane potential has been described (2,2.1). This was derived from measurements of cell chloride which varied only from 2.69 * 22 to 2.98 f. 51 pmol/g dry weight, among the cell types studied. Data derived during transport experiments were expressed as nanomoles/g dry weight, in accordance with conventions established in our earlier studies (2). Data for intracellular drug concentration are expressed as values for exchangeable drug, that is unbound to dihydrofolate reductase. Intracellular levels of this enzyme were determined as described earlier (2).
Transport experiments were carried out at 37 "C with cell suspensions (2-3 X lo7 cells/ml) prepared in buffer-salts solution containing 107 nM NaC1, 20 mM Tris-HC1, 26.2 mM NaHC03, 5.3 mM KCl, 1.9 mM CaC12, 1 mM M&12 with 7 mM D-glucose at pH 7.4. In some experiments, we used a buffer containing 150 mM Hepes (Sigma) and 2 mM MgC12. The pH was adjusted to 7.4 with 66.2 mM KOH. Conditions employed (2) for these determinations ensured measurements of unidirectional influx and efflux of [3H]methotrexate. Experimental procedures used to derive values for influx V-, influx K,, and efflux rate constant have been described in detail (2).
Binding Assay-In a modification of that described by Henderson et al. (31, the assay mixture used contained 3-5 X lo7 L1210 cells and [3H]aminopterin in 150 mM Hepes buffer plus 2 mM M&~z adjusted to pH 7.4 with 66.2 mM KOH. Following incubation for 5 min at 0-4 "C samples were centrifuged in an Eppendorf microcentrifuge at 12,000 X g for 1 min (19). Supernatant was removed by suction and residual fluid removed with cotton swabs. Pellets were then resuspended in water and prepared for counting of radioactivity. Results were corrected for a control in which 500 p~ aminopterin was added prior to [3H]aminopterin to eliminate specific binding and expressed as picomoles for 10' cells. Transport Measurements with Membrane Vesicles-The method of Hochstadt et al. (22) was employed with some modification (17,23,24) for the isolation at 0 "C of L1210 cell plasma membranes following mechanical disruption of washed cells. A determination of purity of membrane preparations by marker enzyme analysis and electron microscopy and the transport procedures have been described (17,23,24).

RESULTS
Growth Inhibition of Parental and Variant L1210 Cells by Folate Antagonists-Three variant cell types were isolated by the selective procedure described above. The sensitivity of these variants compared to parental L1210 cells to growth inhibition by metoprine and methotrexate are shown in Table  I. The variants (R26, R69, and Rs2) show 3-to 26-fold increases in resistance to metoprine compared to parental L1210 cells. Conversely, their sensitivity to methotrexate was increased by 3-to 7-fold. It should be noted that both R69 and R82 cells exhibit 2-fold elevation in the intracellular target, dihydrofolate reductase, which would account for some of the increased resistance to metoprine and tend to offset the increased sensitivity to methotrexate due to transport alterations. For comparison, we also show data for a cell line (R1) shown in our earlier (12) studies to have a reduced V, , for methotrexate influx. This cell line shows increased sensitivity to metoprine but decreased sensitivity to methotrexate.
Characteristics of Methotrexate Transport in Parental and Variant L1210 Cells-Time courses for [3H]methotrexate accumulation at 37 "C in variant L1210 cells exhibited characteristics similar to those already reported (2, 12, 13) from our laboratory for parental L1210 cells. Representative data for one variant (R69) and the parental cell type following exposure to 1 p~ [3H]methotrexate are shown in Fig. lA. Following an initial linear phase of uptake, intracellular accumulation was characterized by an exponential approach to steady state (2-50 min). Measured values of tllz were in the range of 3-4 min, and accumulation is described by a function of the form cr = C,(Iexp(-IGr(tto))) (1) where C , is the steady state, Kef is the rate constant for efflux of drug, and to is the onset of free drug accumulation. In this variant (R69) and the other variants (data not shown), initial uptake and levels of accumulation of exchangeable drug (not bound to dihydrofolate reductase) at steady state were greater than in the parental cell type (Fig. lA). In contrast to the  methotrexate efflux at 37 "C from variant and parental L1210 cells (Fig. 1B) loaded with this compound were indistinguishable. Like accumulation, efflux was characterized by the same exponential, and measured values of tl/P varied between 3-4 min. Based on calculations of membrane potential for these cells from the Nernst equation (2), steady state levels of accumulation for exchangeable drug represent electrochemical gradients of approximately 2-and M-fold, respectively, for parental and R69 cells.  Fig. 1A) were within 10% of those calculated using the following empirical equation, and the values derived for each constant shown in Table 11.

Specific Surface Membrane Binding of [3H]Aminopterin by Parental and Variant L1210 Cells-The various cell types
were examined for their ability to specifically bind [3H]aminopterin at the cell surface. Data derived from this binding assay appears (3, 26) to be a measure of the putative carrier component of the reduced folate transport system, since this binding exhibits (3) structural specificity similar to influx and is decreased (26) in transport-deficient cell lines. Data is shown in Fig. 2    rocal plot (binding/concentration) of the specifically bound fraction we derived data on the association constant (KO) and number of binding sites in each cell type. These are summarized in Table IV. Only small differences in binding affinity (KO) were observed among these cell types. This group of values compared closely to the values for influx Ki (Table IV) derived for each cell type in this same buffer. However, compared to parental L1210 cells, the value derived for total binding sites in variant cells was appreciably higher (3-to 15fold compared to parental cells). This increase in the number of binding sites was proportional to the increase in influx Vgax shown (Table 11) for these variants.   might possibly be a result of perturbations involving cytosolmediated metabolism affecting the rate of carrier translocation or increasing the asymmetry (21, 27) in distribution of carrier at each membrane surface. It was of interest, therefore,

Methotrexate Transport in L1210 Cell Variants
to determine to what extent the altered properties in these variants were expressed in isolated plasma membranes. To do this, we employed methodology for preparing plasma membrane vesicles from the various cell types under study. In these membrane preparations, and unlike intact L1210 cells Metabolic Turnouer of [ 3H]Methutrexate Influx Capacity in Cells Treated with Cycloheximide-We also sought information on the basis for the apparent increase in plasma membrane content of the reduced folate transport system in these variant cell types suggested by the results described above. If this system was metabolically unstable, then depending upon its rate of degradation in parental and variant cell types, the increase in membrane content of this system could be attributed to either increased synthesis or decreased degradation.
To determine this, we measured initial influx of [3H]methotrexate in parental L1210 cells during growth in cell culture with 5 mg/ml of cycloheximide. At this concentration of cycloheximide, protein synthesis, as measured by [3H]leucine incorporation into acid-insoluble cellular material, was inhibited nearly 100% (data not shown). Initial influx was reduced within 30 min after cycloheximide treatment, and further reduction occurred during the 5-h period in which measurements were made. An analysis of this data is shown in Fig. 4, where it can be seen that influx exhibited an exponential decline with a decay time (tlI2) of slightly above 2 h. During this period only minimal effects were seen on cell viability as determined by nigrosin exclusion or clonogenicity assays. Cell viability decreased from 92-95% to 90-92%. The same experiment was also carried out with cultures of each of the variant cell types. In contrast to that seen with intact LE10 cells, initial efflux of [3H]methotrexate from Rg2-derived vesicles was also 10-to 12-fold greater than for parentally derived vesicles. This is consistent with a simple carrier model for facilitated diffusion and bidirectional flux proposed for these vesicle preparations on the basis of our earlier studies (21). DISCUSSION The method of selection employed during these studies appear to provide a means for the isolation of the desired genetic variants of the L1210 cell with elevated transport of reduced folates and folate analogs. This elevated membrane activity could be attributed to a transport system which was a counterpart of the same system in parental cells, since both systems exhibited the same specificity for various folate compounds. Although genetic analysis has not been carried out with these variants, from the manner of their isolation and the stability of the variant property in the absence of selective pressure, they are assumed to have originated by alteration of some genomic element in parental L1210 cells. The phenotype characterized in these studies could have originated in one of a number of ways. The explanation, most likely in view of the data, is that they arose form an alteration of some regulatory component which constitutively controls the rate of synthesis of the putative transport protein. We found that the elevation in influx Vmax characteristic of these variants is associated with a proportional increase in surface membrane content of a specific binding component. Also, both the influx K , and the binding KO were essentially unaltered in these variants, and the system elevated in variant cells exhibited the same specificity for various folate compounds as does the system in parental cells. The full extent of expression of this elevated transport property in plasma membrane vesicles tends to eliminate the involvement of altered cytosol-associated energy metabolism and shows that this increased transport capacity is fixed in the isolated membrane itself. Since directional asymmetry for transport by intact L1210 cells is not maintained (17,23,24,30) in membrane vesicles, it is unlikely that greater asymmetry in intact variant cells is the reason for this increased transport activity. Finally, since we have shown that the rate at which influx of [3H]methotrexate decays is the same in both variant and parental cells, the increased membrane content of this system can probably be explained by increased rate of synthesis of some protein component.
Metabolic turnover of this transport system in 11210 cells was implied from the results of our earlier (18) studies in cell culture. In these studies we showed that during progression in growth of L1210 cells from early logarithmic phase to stationary phase, the Vmax for [3H]methotrexate influx showed a 3-fold fluctuation with a maximum during the midlogarithmic phase of growth. The decrease in [3H]methotrexate influx V,, in L1210 cells observed in the present study following inhibition of protein synthesis would appear to suggest that degradation and turnover of some protein component (putative carrier?) of this transport system normally occurs. Thus, in the face of this continuing degradation, levels of influx might conceivably be modulated by variable rates of synthesis.
The degree to which the increased synthesis of this transport component might be due to an alteration at the transcriptional, translational, or post-translational level of gene expression and membrane protein synthesis must await further study. However, this appears to be the first example in mammalian cells of a genetic variant with stable up-regulation of a membrane transport property. Although it is not known, as yet, how this type of variant is related to others showing possible down-regulation (12, 26) of synthesis of a putative carrier for this transport system, the availability of genetic variants with alterations in the direction of increased and decreased synthesis may offer a unique opportunity to explore aspects of the regulation of synthesis of transport-associated membrane proteins. Moreover, the availability of cell lines with elevated levels of transport protein will undoubtedly facilitate the isolation of this component. This could then lead the way to the development of cloning procedures necessary for probing the molecular genetics of this system. Finally, we believe the biochemical data derived in these studies is further evidence for the notion of separate systems mediating influx and efflux of these analogs in L1210 cells which was proposed on the basis of our earlier kinetic studies (21,27). Evidence derived for an increased rate of synthesis of a specific binding protein in these transport-elevated variants would appear to make it less likely that the absence of an effect on efflux was a kinetic anomaly.