Effect of Serum on Phenylalanine Hydroxylase Levels in Cultured Hepatoma Cells*

high levels of phenylalanine hydroxylase in hepatoma with a half-life of 22 hours for hours, a constant enzyme content persists for at 96 hours. This decline of phenylalanine hydroxylase is reversible; levels are restored in a time-and dosage-dependent fashion upon of serum to basal cultures. The serum factor nondialyzable moderately heat-stable. The by serum of the phenylalanine hydroxylase content of basal cultures is blocked by and requires ongoing cellular protein synthesis. When added to the enzyme-assay mixture irz vitro, serum does not alter the phenylalanine hydroxylase activity of extracts from basal cultures. Three lines of evidence suggest that serum contains a nonsteroidal phenylalanine hydroxylase

Continued high levels of phenylalanine hydroxylase in cultured H4-II-E-C3 rat hepatoma cells require either serum or glucocorticoids in the culture medium. Upon withdrawal of serum, cellular phenylalanine hydroxylase levels decay exponentially with a half-life of 22 hours for about 60 hours, after which time a low, constant enzyme content persists for at least 96 hours. This decline of phenylalanine hydroxylase is fully reversible; normal enzyme levels are restored in a time-and dosage-dependent fashion upon addition of serum to basal cultures. The serum factor is nondialyzable and moderately heat-stable. The stimulation by serum of the phenylalanine hydroxylase content of basal cultures is blocked by 3-[2-(3,,5-dimethyl-2-oxocyclohexyl)-2-h~~drox~~ethyl]glutarimide and requires ongoing cellular protein synthesis. When added to the enzyme-assay mixture irz vitro, serum does not alter the phenylalanine hydroxylase activity of extracts from basal cultures. Three lines of evidence suggest that serum contains a nonsteroidal phenylalanine hydroxylase stimulatory component(s): (a) glucocorticoid antagonists inhibit less than one-half of the biological activity of serum; (b) exhaustive extraction of endogenous serum glucocorticoids with charcoal reduces the activity of serum to about one-half of control values; and (c) the stimulatory effects of charcoal-extracted serum and hydrocortisone are additive. The phenylalanine hydroxylase stimulatory activities of the charcoal-extracted sera from four mammalian species and from three stages in development in one mammalian species are comparable. A survey of partially purified preparations of a number of known hormones failed to reveal any one capable of elevating the phenylalanine hydroxylase levels of basal cultures in a manner comparable to that of charcoal-extracted serum.
Phenylalanine hydroxylase, found in only a few mammalian tissues (liver (l-4), kidney (l-4), and pancreas (l)), is thought to be a differentiated function characteristic of those tissues. With the aim of finding a model system for the study of the mechanisms regulating the expression of the phenylalanine hydroxylase gene in differentiated tissues, a number of cell culture lines of hepatic origin were surveyed for phenylalanine hydroxylase.
tion by hydrocortisone depended on protein synthesis, as the action of hydrocortisone was inhibited by cycloheximide (7). The responses of phenylalanine hydroxylase in the H4 and MH,C, cells to serum' and glucocorticoids (7) are similar, but the H4 cells are more suitable for experimentation than the MH,C, cells because they have a shorter doubling time (about 24 hours) and are more readily detached from the substratum than the MH,C, cells. For this reason, the H4 cells were used for all further experiments.
We report now on the effects of serum and of several hormones on the phenylalanine hydroxylase content of H4 cells. It will be shown that, apart from glucocorticoids and insulin, the serum contains an as yet unidentified component that raises the phenylalanine hydroxylase content of the H4 cells and that this effect, like that of glucocorticoids, requires active protein synthesis.

EXPERIMENTAL PROCEDURE
Cell Culture-The growth of stock cultures of the H3 cells and their subcultivation were carried out as described previously (7) except that the standard growth medium consisted of' a modified Swim's Medium S-77 supplemented with 5% (v/v) f'etal bovine serum plus 1O"c (V/VI NaOH solution bv incubation first overnight at 37" and then for 2 hours at 60". Radioactivity in the various fractions was measured by liquid scintillation spectrometry in 10 ml of Aquasol (New England Nuclear) as described (7). Quenching was estimated by the channelsratio method.
Cultures killed with trichloroacetic  acid solution  before  the addition  of radioactive  substrates  served as controls,  and values for  the incorporation  of label into the trichloroacetic  acid-soluble  and  -insoluble  fractions  of these cultures  were subtracted  from (7). This decline in cellular enzyme content in serum-free medium continues at an exponential rate for about 60 hours (Fig. 11 (7), and extracts of cells from 2 to 6 culture flasks were prepared by the disruption of 175 to 250 mg of packed cells/ml of buffer containing Triton X-100 detergent as described (8). The particle-free extracts were assayed for phenylalanine hydroxylase by the method of Ayling et al. (9,10); 1 enzyme unit is defined as the phenylalanine-dependent oxidation of 1 nmol of 6,7-dimethyltetrahydropteridine to the dihydro form, or the formation of 1 nmol of tyrosine, per min at 37" (7). The reproducibility of the enzyme assay was monitored on the day of each determination by the use of thawed samples of a rat liver extract of known phenylalanine. hydroxylase activity that had been stored frozen in liquid nitrogen (8 interaction between hydrocortisone and charcoal-extracted serum was examined. In Experiment 1 of Table I, maximally  stimulatory concentrations of hydrocortisone and charcoalextracted serum in combination elevated the phenylalanine hydroxylase levels of basal cultures by an amount about equal to that calculated on the basis of numerical additivity, whereas the stimulation evoked by hydrocortisone plus native serum was significantly greater. However, in Experiment 2 of Table I, although the effect of hydrocortisone alone was not as pronounced as in Experiment 1, the effect of either native or charcoal-extracted serum in combination with hydrocortisone was equivalent in magnitude to the stimulation obtained with native serum plus hydrocortisone in Experiment 1; and this same degree of stimulation was also observed under each of these conditions in Experiment 3. It would seem, therefore, that the extraction of serum with charcoal under these conditions impairs little, if any, of the ability of serum to interact with exogenously added hydrocortisone to produce a maximum elevation of cellular phenylalanine hydroxylase content. Although from these experiments the combined effect of maximal concentrations of hydrocortisone and charcoal-extracted serum would appear to be supra-additive, the results from a single additional experiment gave no indication of a similar synergistic interaction between low concentrations of the hormone (4.0 Charcoal-The data from the experiments of Fig. 3 had indicated that native horse serum stimulated phenylalanine hydroxylase in the cells more than did native fetal bovine serum. Because this difference could have resulted from differing concentrations of endogenous glucocorticoids in the two sera, and not from a difference in the contents of their nonsteroidal phenylalanine hydroxylase stimulatory components, the biological activities of several mammalian sera were surveyed after extraction with charcoal. Table II summarizes the results obtained. We conclude from these data that the nonsteroidal phenylalanine hydroxylase stimulatory activities of porcine, bovine, calf, fetal bovine, and human sera are comparable to that of Serum-In order to examine the possibility that the nonsteroidal component(s) of serum acting on cellular phenylalanine hydroxylase might be any one of a number of already known hormones, various hormones were tested for their ability to stimulate the phenylalanine hydroxylase levels of basal cultures. The following peptide hormones, each tested at concentrations of 0.1, 1.0, and 10 pg/ml, had no effect on the level of phenylalanine hydroxylase in basal cultures (postconfluent cultures preincubated in serum-free medium for 4 days): bovine and rat somatotropin and follicle-stimulating hormone and bovine luteinizing hormone, thyrotropin, and prolactin. The level of phenylalanine hydroxylase in all the experimental cultures remained at the control basal level of 16 * 5 units/g after 40 hours of exposure to the hormones. In the same experiment the charcoal-extracted horse serum (20%) caused a rise to 103 units/g. Epinephrine at a concentration of 1Om6 M was toxic and at 10m7 and 10e8 M had no effect; triiodothyronine was also without effect in the range of 10e9 to 10m5 M. Some stimulatory effects were noted with glucagon and with insulin; the results of these experiments are summarized in Table III. The weak stimulatory effect of glucagon, seen only at the highest concentration tested, was thought to be of questionable physiological significance and was not further studied.
On the other hand, since insulin exhibited a modest phenyalanine hydroxylase stimulatory activity that was equivalent over the range of concentrations examined, the possibility was explored that the presence of endogenous insulin might be responsible for at least part of the effect of charcoal-extracted serum. First, the stimulatory activities of insulin and charcoalextracted serum, both alone and in combination, were determined to ascertain whether or not additive effects could be observed. The data from Experiment 1 of Table IV  dosage range had not been tested in the earlier experiments, the effect of insulin on the phenylalanine hydroxylase content of basal cultures was examined at a concentration about 5-fold greater than this value (14.6 microunits/ml); and in the same experiment the ability of a preparation of guinea pig anti-insulin serum of known immune titer to antagonize both the effect of this level of insulin and that of 20% charcoal-extracted serum was also determined.
The data from Experiment 2 of Table IV argue strongly against the possibility that the presence of endogenous insulin contributes significantly to the phenylalanine hydroxylase stimulatory activity of charcoal-extracted serum. First, when insulin was added to the medium at a concentration of 14.6 microunits/ml, only a marginal elevation of cellular enzyme content was seen. Second, whereas this small stimulation was completely blocked by the anti-insulin serum, this antiserum, when present at a concentration of 50 (anti) microunits/ml, had no antagonistic effect on the phenylalanine hydroxylase stimulatory activity of charcoal-extracted serum. From all these experiments it was concluded that the presence of physiological levels of either somatotropin, thyrotropin, follicle-stimulating hormone, luteinizing hormone, prolactin, epinephrine, triiodothyronine, glucagon, insulin, or testosterone alone' could not be responsible for the phenylalanine hydroxylase stimulatory activity of charcoal-extracted serum.
Dependence on de NOLJO Protein Synthesis of Stimulation of 'While the formal possibility still cannot he excluded that a subtle combination of two or more of these hormones might he acting synergistically at low levels to produce the observed effects of serum, to approach this question through an experimental protocol such as the one used above would be prohibitively complex without any prior information as to the endogenous concentrations of these hormones in serum by means of appropriate radioimmunoassays.  Table  V), and also,  Table V, the final specific activity of the labeled precursor will be higher in the serum-containing medium than in the serum-free medium, thus producing a greater apparent rate of protein synthesis in the cultures previously exposed to serum.
From the experiments of Fig. 5 and Table V, it is concluded (a) that the protein synthetic rates of postconfluent basal cultures incubated with and without serum are equivalent, (b) that the effect of serum on cellular phenylalanine hydroxylase content cannot be attributed to a nonspecific increase in cellular protein anabolism evoked by serum under these conditions, and (c) that some component(s) of serum evokes a relative stimulation of phenylalanine hydroxylase levels in the cells by a mechanism that depends on ongoing protein synthesis. Consistent with this view is the observation that horse serum or cycloheximide, when added in. vitro to the enzymeassay mixture, does not affect the phenylalanine hydroxylase activity of soluble extracts from cells exposed to serum-free or serum-containing medium, respectively. * The experiment of Table V also provides additional verification of the physiological integrity of H4 cells cultured for up to 5 days in serum-free medium since the protein synthetic activity of these cultures is comparable to that of sister cultures after re-exposure to serum for the final 23 to 27 hours. DISCUSSION The effects of serum on cultured animal cells are manifold and complex; indeed, only a few established cell lines are capable of replication, or even survival, in culture without serum or some macromolecular component derived from serum (19,(22)(23)(24)(25)(26)(27)(28)(29)(30)(31). Thus, within the context of the diverse biological activities of serum, a major concern in attempting to evaluate the significance of an effect of serum on any given cellular function is the extent to which that effect is specific for the function in question. The experimental evidence presented here is consistent with the following conclusions.
Cultivation of H4-II-E-C3 cells in serum-free medium results in an exponential decay of their phenylalanine hydroxylase content to values that probably reflect the constitutive level of expression of the phenylalanine hydroxylase gene in these cells. Re-exposure of the resulting basal cultures to serum evokes a stimulation of their enzyme content that is both time-and dosage-dependent and requires protein synthesis. The stimulatory factor in serum is not dialyzable and is stable at 65" for 15 min. Evidence that serum contains a phenylalanine hydroxylase stimulatory component(s) in addition to the endogenous glucocorticoids was garnered from three observations: (a) glucocorticoid antagonists ("anti-inducer" steroids) inhibited less than one-half of the biological activity of serum in the cell culture system; (b) serum that had been exhaustively extracted with charcoal still retained about one-half of the phenylalanine hydroxylase stimulatory activity of native serum and this activity was not antagonized by the antiinducer fluoxymesterone; and (c) the effects of maximally stimulatory concentrations of charcoal-extracted serum and hydrocortisone were additive. From the last of these findings, it follows that both glucocorticoids and the active component(s) in serum are required for the maximum expression of cellular phenylalanine hydroxylase levels. A survey of partially purified preparations of a number of known hormones failed to reveal any one capable of elevating the phenylalanine hydroxylase levels of basal cultures in a manner comparable to that of charcoal-extracted serum. Finally, whereas the phenylalanine hydroxylase stimulatory activities of various mammalian sera in the native state were variable, the activities of the same sera after charcoal extraction were indistinguishable from one another.
The data presented in Table V and Fig. 5 suggest that the effect of serum on cellular phenylalanine hydroxylase content, while requiring concomitant protein synthesis, cannot be attributed to a nonspecific increase in protein anabolism. In this connection, it may be noted that, in marked contrast to the pronounced pleiotypic effects' of serum on 3T3 fibroblasts (32), a serum "step-down" from 10 to 0.5% for up to 72 hours caused little decrease, if any, in the rate of cellular protein synthesis in BHK 21/13 fibroblasts, as measured both by the incorporation of labeled valine into cultures and by the protein synthetic activity of cell extracts in a cell-free system in. vitro (33). On the other hand, in hepatoma tissue culture cells preincubated in serum-free medium plus dexamethasone for 16 hours, after exposure to 5% bovine serum for 1 hour, the rate of incorporation of labeled amino acid into total cellular protein was increased by 43% (34) and the rate of total protein degradation decreased by 18% (35).
A comparison of the effects of serum on tyrosine aminotransferase and phenylalanine hydroxylase in cultured rat hepatoma cells warrants some discussion. First, whereas in hepatoma tissue culture (34) cells the stimulatory effects of serum and dexamethasone on tyrosine aminotransferase are additive, the stimulation by serum requires the continuous presence of dexamethasone.
In contrast, an increase in the phenylalanine hydroxylase levels of H4 cells can be effected by serum in the absence of glucocorticoides (Fig. 4). Conversely, through the use of a strain of RLC cells (RLC-GAI) adapted to growth in serum-free medium (19), Gerschenson et al. have shown that the paradoxical elevation by actinomycin D of tyrosine aminotransferase activity in dexamethasone-preinduced cells (i.e. "superinduction") requires serum (36). It would be of interest to ascertain whether or not superinduction by actinomycin D, if also seen with the phenylalanine hydroxylase system of H4 cells, would likewise be serum-dependent.
Finally, the tyrosine aminotransferase stimulatory activity of bovine serum, like the phenylalanine hydroxylase stimulatory activity of horse serum, was moderately heat-stable and nondialyzable (34). The following lines of evidence would argue that serum and hydrocortisone elevate the phenylalanine hydroxylase levels of H4 cells by increasing the amount of phenylalanine hydroxylase protein and not by activating the enzyme, either directly or indirectly: (a) the effects of these agents are not detectable before 8 to 10 hours of exposure; (b) both effects require continuous cellular protein synthesis; (c) neither agent has any effect on the phenylalanine hydroxylase activity of cell extracts when added to the enzyme-assay mixture in uitro; (d) a mixture of extracts from basal and from serum plus hydrocortisone-stimulated cells contained the phenylalanine hydroxylase levels expected on the basis of numerical additivity'; and (e) through the use of sheep anti-phenylalanine hydroxylase serum, generously provided by Dr. Seymour Kaufman (37), extracts from cells exposed either to charcoal-extracted horse serum alone or to hydrocortisone alone were found to contain greater amounts of immunoprecipitable material than did 'I.e. coordinate increases in the rates of uptake of macromolecular precursors and of macromolecular synthesis and decrease in the rate of macromolecular degradation, as defined by Tomkins and co-workers (32). extracts from basal cells, as visualized by Ouchterlony double immunodiffusion.'-' It remains to be determined whether each of these agents elevates the amount of phenylalanine hydroxylase in the cells by increasing the rate of synthesis of the enzyme, decreasing the rate of its degradation, or both. In this connection, if the rates of formation and decay of enzyme are assumed to be zero and first order, respectively, in accordance with the model first proposed by Schimke (40), the fact that the half-times of stimulation and decline of cellular phenylalanine hydroxylase content upon exposure to charcoal-extracted serum (Fig. 2) and upon withdrawal of serum (Fig. l), respectively, are comparable (about 22 hours) would support the first of these alternatives with respect to serum.