Effect of cobalt on synthesis of heme and cytochrome P-450 in the liver. Studies of adult rat hepatocytes in primary monolayer culture and in vivo.

Synthesis of heme and cytochrome P-450 have been studied in adult rat hepatocytes in primary monolayer culture. Incubation of cells with delta-aminolevulinic acid increases both cellular heme and cytochrome P-450 relative to that found in cells incubated under control conditions. Formation of heme is proportional to the concentration of delta-aminolevulinic acid in the culture medium and is not saturable. By contrast, formation of cytochrome P-450 is saturable; excess intracellular heme appears as a new absorption band at 420 nm in the carbon monoxide-reduced difference spectrum. We have studied the effect of cobalt on heme and hemoprotein formation in this cell system. The metal blocks formation of cytochrome P-450 but fails to affect heme synthesis. In contrast to previous findings with isolated mitochondria, no cobalt protoporphyrin formed in hepatocytes cultured in the presence of the metal. In studies of rats in vivo, it was confirmed that cobalt acts to reduce the amount of [14C]heme in the liver after administration of delta-amino[14C]levulinic acid. However, the present findings suggest that this effect of cobalt represents accelerated breakdown of newly labeled hepatic heme rather than inhibition of synthesis. We conclude that cobalt interferes with formation of cytochrome P-450 not by direct inhibition of heme synthesis but most likely by blocking the association of heme and apocytochrome.

Synthesis of heme and cytochrome P-450 have heen studied in adult rat hepatocytes in primary monolayer culture. Incubation of cells with b-aminolevulinic acid increases both cellular heme and cytochrome P-450 relative to that found in cells incubated under control conditions. Formation of heme is proportional to the concentration of d-aminolevulinic acid in the culture medium and is not saturable. By contrast, formation of cytochrome P-450 is saturable; excess intracellular heme appears as a new absorption band at 420 nm in the carbon monoxide-reduced difference spectrum. We have studied the effect of cobalt on heme and hemoprotein formation in this ce11 system. The metal blocks formation of cytochrome P-450 but fails to affect heme synthesis. In contrast to previous findings with isolated mitochondria, no cobalt protoporphyrin formed in hepatocytes cultured in the presente of the metal. In studies of rats in uiuo, it was confirmed that cobalt acts to reduce the amount of ["C Jheme in the liver after administration of b-amino['"C]levulinic acid. However, the present findings suggest that this effect of cobalt represents accelerated breakdown of newly labeled hepatic heme rather than inhibition of synthesis. We conclude that cobalt interferes with formation of cytochrome P-450 not by direct inhibition of heme synthesis but most likely by blocking the association of heme and apocytochrome.
A group of microsomal cytochromes, known collectively as cytochrome P-450, are inducible in the liver by administration of various drugs, carcinogens, and environmental contaminants. Despite the importance of cytochrome P-450 in metabolism of these substances, regulation of this hemoprotein remains poorly understood. For example, it is not clear whether stimulated synthesis of heme or of apoprotein is the initial event of induction. Although previous work had suggested that availability of heme is rate-determining for cytochrome formation (l), evidente against this postulate has been adduced (2), and recent findings are consistent with the concept that synthesis of apocytochrome is the limiting factor (3,4). In most of these studies, inhibitors of heme synthesis were used to dissociate the metabolism of heme and apocytochrome. Cobalt is particularly attractive in this regard, because available information suggests that the metal blocks heme synthesis selectively either by competing with iron for insertion into protoporphyrin (5), or possibly by decreasing enzymatic formation of heme precursors (6). We have used adult rat hepatocytes in primary monolayer culture (7,8) to investigate the regulation of heme and cytochrome formation and the effect of cobalt on these processes. The culture system differs from those involving fetal rat (9) or embryonic chick liver (10) in a number of important ways. Because the cells are derived directly from adult animals, they contain easily measurable amounts of cytochrome P-450 and in culture perform numerous functions of intact rat liver (11). The cells differ from intact rat liver in that, during the first 24 hours of incubation in standard medium, the concentration of cytochrome P-450 falls with respect to that measured in uiuo. This change appears to reflect metabolic adaptation of the cells to the conditions of culture, since it can be modified by alteration of the culture medium (12). In particular, addition of &aminolevulinic acid to the medium increases the rate of endogenous heme synthesis and the concentration of cytochrome P-450 in the cultured hepatocytes (12). This response to d-aminolevulinic acid permits controlled manipulation of the formation of both heme and cytochrome P-450. Therefore, by this approach, the relationship between heme synthesis and cytochrome formation in the liver can be studied by examining the effect of cobalt on these processes. In agreement with the results of previous studies in uiuo, we found that formation of cytochrome P-450 is blocked by cobalt. However, we failed to detect any direct inhibitory effect of the metal on heme synthesis per se and conclude that it acts at a step in the formation of cytochrome P-450  (7). Cells were obtained from rat liver perfused in situ with buffered collagenase solution. The isolated hepatic parenchymal cells were separated cleanly from debris and non-parenchymal cells by centrifugation and were placed in plastic tissue culture dishes (60 mm) in a serum-free defined medium. With plating of cells at optimal density (5 x lo6 tells/60-mm plate) and in a volume of 3.0 ml, a viable monolayer of hepatic parenchymal cells forms after 12 to 18 hours of incubation.
The medium used in the present experiments, termed L-16, was slightly modified from that described previously (7), in that phenol red, pyruvate, galactose, and Amphotericin B were omitted. Substances to be added to the media were dissolved directly, and the solution was sterilized by filtration.

Experimental Approach
All experiments with cultured adult rat hepatocytes were performed during the initial 24 hours after the cells had been plated. This period was selected for study because the rapid changes in cytochrome P-450, as well as the response of the cells to modified culture media, occur largely during this time. Thus, in a given experiment, cells from a single rat liver were apportioned according to protocol and plated either in control or in modified media.
After 20 to 24 hours of incubation, all cultures were terminated, and the cells were analyzed according to the procedures described below.

Assay Procedures
After removal of the medium and washes with phosphate-buffered saline, pH 7.4. monolayer cells from two or more plates were combined, disrupted by sonication. and centrifuged, for preparation of a 10,000 x g supernatant or of microsomes, as described previously (7). Determination of Cytochrome P-450--Suspensions of microsomes or 10,000 x g supernatant in 0.1 M potassium phosphate buffer were reduced with a few crystals of dithionite, and were placed in 1 ml of self-masked microcells (Helma Cell Inc., Jamaica, N.Y.). The cytochrome P-450 concentration was then determined by the method of Omura and Sato (13) from the carbon monoxide (CO) difference spectrum assuming a millimolar extinction coefficient of E150.180 nm = 91 cm-' mM-'. Results were expressed as nanomoles per mg of protein.

Measurements
were made on a DW-2 split beam recording spectrophotometer (American Instrument Co., Silver Springs, Md.). The concentration of heme in the CO complex absorbing at 420 nm was calculated using E,20.,sOnm = 110 cm-' mM-' (13). The concentration of cytochrome P-450 in samples of material from one to two culture plates were in excellent agreement with lo-fold more concentrated preparations from cultured cells or from normal rat liver.
Determination of C'vtochrome b,-The method of Omura and Sato was used (14). The reduced minus oxidized difference spectrum in microsomal suspension was determined at room temperature by addition of 50 nmol of NADH in 10 ~1 of 0.1 M potassium phosphate, pH 7.4, to the sample cuvette and 10 ~1 of buffer to the reference cuvette after a base-line had been previously established.
Assay of Aminopyrine N-Demethylase-The microassay used was a slight modification of that described by Poland and Nebert (15 In brief, rats weighing 190 to 210 g were injected intraperitoneally with 5 rCi of d-amino[5-"Cllevulinic acid in isotonic saline (0.9% NaCl solution) to selectively label hepatic heme (18). Exhaled '"CO was measured by a system described previously (19), in which "CO, is removed first by a series of traps. The '"CO which remains is then converted by heated hopcalite to '"COz and trapped in 4 ml of ethanolamine/methoxyethanol (l/l). A l-ml aliquot of this solution is added to 4 ml of absolute methanol, 1 ml of methoxyethanol, and 12 ml of Liquifluor (New England Nuclear Corp., Boston, Mass.). Radioactivity is quantitated by liquid scintillation spectrometry with an efficiency of 92%. To facilitate comparison of control and treated animals, rats were studied in pairs in two CO trains operated in parallel from the same vacuum source. Recovery of "CO in this system is quantitative. Quantitation of ["C]Heme in Cytochromes P-450-The heme of rat liver was labeled in viuo with &amino[5-"Cllevulinic acid (as above, for the studies of '"CO excretion).
After time intervals determined by experimental protocol (see "Results"), rats were killed and the liver perfused with 0.25 M sucrose and excised. Microsomes were prepared and treated with subtilisin, as described by Comai and Gaylor (20). The treated microsomes exhibited virtually total loss of cytochrome b,, with negligible conversion of cytochrome P-450 to P-420. The specific activity ["lheme of cytochrome P-450 was determined directly from the treated microsomes.
Protein-The method of Lowry et al. (21) was used with crystalline bovine albumin as the standard.

Microsomal Function in Adult
Rat Hepatocyte Culture-The evidence for sustained viability and the morphologic appearance of hepatocytes in this culture system have been presented in detail elsewhere (7, 22). Among microsomeassociated processes, the most striking finding is an early decrease in the cellular level of cytochrome P-450. This change is evident, by comparison with the liver in Lo, after 4 hours in culture and after 20 hours is pronounced, with the level of cytochrome P-450 dropping to 10 to 20% of its concentration in freshly isolated hepatocytes. The selectivity of this change has been assessed by survey of other microsomal enzymes. Glucase-6-phosphatase activity is unchanged while the content of cytochrome b, decreases moderately (Table I) over the first 24 hours of culture. The activity of various drug-metabolizing enzymes is variably affected, suggesting that the amount of cytochrome P-450 in the microsomes may be limiting for some of these activities but not for others. These results are reported in detail elsewhere.* No evidence for the loss of a specific type of cytochrome P-450 (e.g. cytochrome P-450, as distinguished from cytochrome P-448) was found, in that the dithionitereduced CO difference spectra of the hemoproteins in intact liver, in fresh cells, and in 24.hour cultures are identical. Specifically, no shift of the peak absorbance to shorter wavelengths in the cultured cells was detected. Furthermore, at no point during the first 24 hours of culture while cytochrome P-450 was falling rapidly, did the reduced CO difference spectrum reveal material absorbing at 420 nm.
Effect of &Aminolevulinic Acid on Heme Synthesis and on Cytochrome P-450 in Cultured Hepatocytes-Addition of 6-; aminolevulinic acid to the culture medium increased the rate of heme formation by the cells in proportion to the concentration of &aminolevulinic acid, as judged both by spectrophotometric studies (Figs. 1 and 2) and by incorporation of 6. amino['"C]levulinic acid into cellular heme (Table II). When cells are plated in medium containing 6-aminolevulinic acid, cytochrome P-450 also increases significantly after 24 hours, as compared to the concentration of this hemoprotein in cells incubated in control medium (Table I). The increase in cytochrome P-450 is dependent upon the concentration of added d-aminolevulinic acid, up to 8 x 10m5 M. At concentrations greater than this, the level of cytochrome P-450 fails to change, although the heme content of the cells continues to increase. The additional heme formed is proportional to the height of a peak at 420 nm in the reduced CO difference spectrum (Figs. 1 and 2). Although the absolute height of the 420 nm peak varied among batches of cells incubated with a given concentration of Saminolevulinic acid (cf. Figs. 1 and 3), the "threshold" concentration of &aminolevulinic acid for production of the 420 nm peak was relatively constant. The increase in cytochrome P-450 concentration, resulting from addition of &aminolevulinic acid to the medium, was reflected in increased activity of the drug-metabolizing system in vitro (Table III), which indicates that the observed increment in hemoprotein represents functionally active material. Addition of 6-aminolevulinic acid modified the composition of microsomal hemoproteins selectively, since the level of cytochrome b, was unaltered under these conditions ( Table I).
Effect of Cobalt on &Aminolevulinic Acid-mediated Increase in Cytochrome P-450 in Cultured Hepatocytes-The decrease in cytochrome P-450 exhibited by cells cultured in control medium was unaffected by the presence of cobalt, so that concentrations of the hemoprotein at 24 hours of incubation were similar in cobalt-treated and control cultures (Fig. 3). However, cobalt blocked the increase in cytochrome P-450 associated with addition of d-aminolevulinic acid to the culture medium (Fig. 3), and the effect of cobalt was concentrationdependent (Table IV). The metal had no effect on incorpora-' P. S. Guzelian, manuscript in preparation.
tion of [Ylleucine into protein measured either in the total cellular material or in the microsomal fraction, and it had no effect on the morphology of the cultured hepatocytes. These data suggest that the actions of cobalt in the liver are relatively specific. In contrast to its effect on the concentration of cytochrome P-450, cobalt failed to inhibit the conversion of b-aminolevulinic acid to heme. As shown in Fig. 3 and in Table  IV, the absorbance at 420 nm from cells incubated with b-aminolevulinic acid and cobalt, is increased in comparison to cells incubated with &aminolevulinic acid alone. The incorporation of radioactive d-aminolevulinic acid into heme similarly is either unchanged or increased in cells incubated with cobalt (Table II).
Additional studies were carried out to exclude the possibility that the spectrophotometric changes and metabolism of 6amino["C]levulinic acid in cells incubated with cobalt may represent formation of cobalt protoporphyrin.
Hepatic mitoc'rondria, isolated either from intact liver or from cultured hepatocytes, were incubated with cobalt and protoporphyrin. Under these conditions cobalt protoporphyrin was formed as has been described previously (23) ( Table V). However, when the mitochondrial preparations were examined for cytochrome P-450, the reduced CO difference spectrum revealed no absorption peak between 420 and 490 nm. Thus, the spectral changes in cells incubated with b-aminolevulinic acid and cobalt are unlikely to represent cobalt protoporphyrin.
This was further established by incubation of hepatocytes with d-aminolevulinic acid and cobalt or with d-aminolevulinic acid alone. Extract (10,000 x g supernatant) from the cobalt-treated culture was placed in the sample cuvette of the split beam spectrophotometer; a comparable amount of extract from a control culture was placed in the reference cell, and the oxidized difference spectrum was examined. In contrast to the studies with isolated mitochondria, no peak at 438 nm appeared under these conditions. Finally, we examined directly the production of heme by hepatocytes cultured in the presence or absence of cobalt and &aminolevulinic acid, by measuring the incorporation of 5SFe into heme. Use of 55Fe circumvents the possibility that cobalt protophorphyrin if formed, could be included in the heme isolation procedure. As shown in Table VI, production of ["Felheme increased 4-to 'i-fold in the presence of baminolevulinic acid. This is comparable to the spectrophotometric increment in cellular heme, calculated from the change in cytochrome P-450 plus the 420 nm complex (Table IV). Addition of cobalt (10m5 M) to the system had no effect on the synthesis of [""Felheme. These observations establish that cobalt neither inhibits heme synthesis in cultured hepatocytes nor causes production of cobalt protoporphyrin. Thus, its ability to block formation of cytochrome P-450 (Fig. 3) appears to involve a process other than heme synthesis per se.
Effect of Cobalt on Hepatic Heme Synthesis in Vivo-For purposes of examining the evidence for inhibition of heme synthesis in oiuo by cobalt, we adopted the experimental protocol of Tephly et al. (5). When animals were treated with cobalt or saline and, 6 hours later, given a pulse of 6amino[Y]levulinic acid, the amount of labeled heme in the liver was found to be reduced in the cobalt-treated animals, relative to the control (Table VII). These data confirm previous findings (6, 7). To determine whether this effect of cobalt involves decreased synthesis or increased degradation of newly labeled heme, we treated rats with cobalt as above, administered a pulse of &amino[5-"Cllevulinic acid and measured the rate of heme breakdown in terms of "CO production (18). As is that increased breakdown accounts fully for the reduced amount of ['%]heme in the liver of cobalt-treated animals. Differences in precursor pool size are unlikely to account for these results, because the effect of cobalt treatment on ['"Clheme breakdown and on hepatic [Y]heme content are reciprocal. These data obtained in intact animals are consistent with the observations in cell culture, that cobalt has little, if any, direct effect on the conversion of &aminolevulinic acid to heme in the liver.

DISCUSSION
Adult rat parenchymal cells in culture respond to the addition of &aminolevulinic acid by increasing the synthesis of both heme and cytochrome P-450. By contrast, hepatocytes in uiuo react to administration of &aminolevulinic acid by increasing only the synthesis of heme (2). The reason for this discrepancy remains unexplained; nevertheless, the present hepatocyte culture system provides a way to examine the relationship between the formation of heme and cytochrome P-450 in intact liver cells. With increasing amounts of 6aminolevulinic acid (up to 8 x 10m5 M) added to the culture medium at the time of cell plating, the concentration of cytochrome P-450 after 24 hours of incubation is increased proportionately.
which suggests that available heme is utilized preferentially in the formation of this hemoprotein. At concentrations of 6aminolevulinic acid greater than 8 x 10m5 M, the level of cytochrome P-450 ceases to increase; instead, a new absorption band at 420 nm appears (Fig. l), which increases in proportion to the concentration of &aminolevulinic acid in the medium, i.e. to the rate of heme synthesis in the cells. The formation of the 420 nm complex was not saturable at the highest level of b-aminolevulinic acid introduced in these studies (1Om3 M). At relatively high concentrations of d-aminolevulinic acid in the medium, the amount of 420 nm material represents a quantity Cells prepared from the livers of seven rats (Experiment 1) or three hours in presence of 6-aminoleoulinic acid rats (Experiment 2) were plated and incubated for 24 hours in either control or 6-aminolevulinic acid (ALA)-supplemented medium. The pooled monolayers were scraped and homogenized. Measurements of cytochrome P-450 concentration and enzyme activity were made on 10,000 x g supernatant as described under "Experimental Procedure" in the material from two or three plates, respectively. Results are expressed as per cent of the values in incubated control plates which were as follows for Experiments 1 and 2, respectively: cytochrome P-450, 0.022 and 0.026 nmol/mg of protein; aminopyrine N-demethylase (AP), 0.163 and 0.230 nmol of HCHO formed/min/mg of protein; aniline hydroxylase (ANL), 0.061 and 0.089 nmol/min/mg of protein.  Cell cultures were plated in L-16 medium in the presence or absence of 6-aminolevulinic acid (ALA) (1 x 10.' M) or cobaltous chloride. After 24 hours of incubation, the cells from two plates were combined, 10,000 x g supernatant was prepared, and the dithionite-reduced CO spectrum was recorded (see "Experimental Procedure").     (3,4,26 formed hepatic heme was not excluded. Since it is known that breakdown products of labeled heme (bilirubin or carbon monoxide) normally appear as early as 30 min after injection of &amino["C]levulinic acid (28), it seemed important to reexamine the studies of ["Clheme incorporation in cobalttreated and control animals, with particular attention to degradation of newly labeled heme. When such studies were carried out, it was found that degradation of newly synthesized heme was strikingly accelerated in cobalt-treated animals (Fig.  4). Indeed, accelerated degradation may account entirely for the decreased content of ["Clheme in the liver of treated animals. These findings in uiuo corroborate the studies in culture, with the conclusion that cobalt has little direct effect on synthesis of heme in the liver.
Acknowledgments-We are indebted to Lydia Hammaker for many helpful discussions and for assistance with some of the studies and to Doctors Rudi Schmid and Almira Correia for a critical review of the manuscript. Ruth Wayne McPherson and Linda Kendrick provided very able technical help.
An unexpected finding in these studies was that incubation of hepatocytes with cobalt failed to reduce the basal level of cytochrome P-450 over and above the reduction associated with adaptation of cells to culture. By contrast, administration of cobalt in vivo causes a marked drop in the level of cytochrome P-450 in the liver (5). A possible explanation for this discrepancy is that the conditions of cell culture may perturb hemoprotein metabolism in a manner similar to that produced by cobalt in uivo. If this is correct, the effects of cobalt in the hepatocytes would not be additive with the culture-related perturbation under standard conditions of incubation.
However, when the availability of endogenous heme for synthesis of cytochrome P-450 was increased by addition of &aminolevulinic acid, the effect of cobalt in preventing new formation of cytochrome P-450 became apparent.
The present observations suggest a new concept for the effect of cobalt on the synthesis of cytochrome P-450. Not only does the metal fail to block heme synthesis, it also has no inhibitory effect on the synthesis of microsomal protein, either in cell culture or in uivo (5), which suggests that synthesis of apocytochrome is unaffected by cobalt. Therefore, cobalt may act on a step in the formation of cytochrome P-450 which follows synthesis of the heme moiety, as for example, by blocking the translocation of heme from mitochondria to endoplasmic reticulum or by interfering with the association of heme and its specific receptor in the membrane (apocytochrome P-450). In hepatocytes cultured in the presence of &aminolevulinic acid and cobalt, the reciprocal relationship of cytochrome P-450 and of the 420 nm complex (heme bound "nonspecifically" to microsomes) suggests that cobalt prevents the normal association of heme and apocytochrome P-450, displacing cytochrome P-450 heme. This mechanism would imply that cobalt administered in vivo diverts heme destined for assembly with apocytochrome P-450, presumably into degradative channels. Consistent with this postulate is the fact that cobalt has been shown to stimulate hepatic heme oxygenase activity in the liver (29), and in the present studies, to accelerate heme breakdown.