Phosphorylase: a new isozyme in rat hepatic tumors and fetal liver.

A third set of phosphorylase a and b isozymes, distinguishable kinetically and immunologically from liver and muscle forms, is present in various rat hepatomas, and is also present, together with the adult liver form, in fetal rat liver. This is one of several striking examples of suppression of isozymes of adult liver coupled with the appearance of fetal isozymes in hepatomas.

Mitochondrial polyriboadenylate polymerase: relative lack of activity in hepatomas.

An enzyme that polymerizes adenylate residues from adenosine triphosphate was prepared from rat liver mitochondria and compared to similar preparations from the mitochondria of three hepatomas. Enzyme activity in the hepatomas was only 1 to 2 percent of that in normal liver.

Loss of the cholesterol feedback system in the intact hepatoma-bearing rat.

By means of the desmosterol suppression technique described in the previous paper, the influence of hepatomas on sterol metabolism has been studied in the intact rat. The major finding of this study is that all hepatoma-bearing rats demonstrate a consistent in vivo loss of the cholesterol feedback system that is characteristic of normal liver. The results also demonstrate that such tumors retain only minor amounts of the sterol they synthesize, releasing over 90% of such endogenous sterol into the circulation. Finally, the in vivo loss of cholesterol feedback control was found to occur in at least two minimal deviation hepatomas and in one highly malignant adenocarcinoma of hepatic origin. These findings indicate that even tumors that are capable of only very limited cholesterol synthesis in vitro, can contribute significant quantities of sterol to the bloodstream. It is concluded that as a result of their lack of normal cholesterol feedback control, hepatomas may represent a physiologically important source of sterols in the tumor-bearing animal, and that the absence of feedback control of sterol synthesis may provide a means of detecting the presence of such tumors in the intact animal.

Ultrastructure of a mammosomatotrophic and a nonfunctional transplantable pituitary tumor induced in rats by 2,4,6-trimethylaniline.

A pituitary tumor was induced in a female inbred BUF rat on an 18-month diet containing 1.1 mmole 2,4,6-trimethylaniline/kg. In the sixth transfer, this tumor developed into two lines of transplantable tumors with different characteristics. Here these two lines were studied by light and electron microscopy; histologically, both tumors were well-differentiated pituitary carcinomas. In the ultrathin sections, the neoplastic cells were separated by a wide intercellular space and covered by numerous microvilli. In the mammosomatotrophic tumor (7315a) the neoplastic cells contained big, electron-dense, ovoid or globular secretory granules (560-1,700 nm in size) that were similar to the prolactin granules of mammotrophs. However, in these cells were also small, pale, uniform, secretory granules that were along the plasma membrane, measured 160 nm in diameter, and resembled the ACTH-containing granules of corticotrophs. The neoplastic cells in tumor line 7315i possessed secretory granules comparable to the granules of the ACTH-secreting cells. The differentiation of the ergastoplasm was abnormal. The tumor cells contained an endoplasmic reticulum similar to mammotrophs and somatotrophs but dissimilar to ACTH-secreting cells. These investigations suggested that the production of several hormones in transplantable pituitary tumors resulted from the multisecretory differentiation of one neoplastic pituitary cell.

Hormonal induction of tyrosine aminotransferase activity in host liver and hepatoma no. 7777 of normal and cofactor-depleted animals.

Induction of tyrosine aminotransferase (TAT) (EC 2.6.1.5) by hydrocortisone was studied during cofactor (pyridoxal phosphate) depletion in hepatoma-bearing BUF strain female rats. Pairs of rats were matched for weight and age and one from each pair was fed ad libitum a diet lacking pyridoxine; the other (referred to as "pair-fed") was given the same diet supplemented with the vitamin, with the amount restricted to that consumed by the matched animal on the deficient diet. All animals were inoculated with Morris hepatoma no. 7777 cell after 21 days on the respective diets. TAT specific activity was determined weekly in host liver and hepatoma, in the presence and absence of cofactor, before and after the administration of hydrocortisone. Free and bound pyridoxal phosphate was estimated enzymatically. The average weight of hepatomas from pair-fed animals was 1.5-fold to twofold greater than that of hepatomas from animals on deficient diets. TAT activity of hepatomas was two times greater than that of host liver, and lack of dietary pyridoxine was without effect. Hormonal induction of enzymatic activity was maximal after the first week of tumor growth and subsequently reached minimal values. In pair-fed animals, tumor TAT was approximately 60% saturated with cofactor. In vitamin-deficient animals, only 6% of the tumor enzyme was saturated with the cofactor. The percent saturation of host liver TAT varied, with minimal values found in the vitamin-deficient animals. Hepatic and tumor pyridoxal phosphate content of pair-fed animals was unusually high (10 mug/g); in vitamin-deficient animals, only the coenzyme content of hepatomas was high (7.0 mug/g). The results showed that presence of the tumor altered the a) specific activity level of TAT and tissue content of cofactor, b) pattern of hormonal induction of the enzyme, and c) effects of the absence of dietary pyridoxine on TAT induction observed in animals without tumors.

Chromosome banding patterns of two transplantable pituitary tumors induced in rats by 2,4,6-trimethylaniline.

The chromosomal constitution of a mammosomatotrophic and an inactive transplantable pituitary tumor induced in rats by 2,4,6-trimethylaniline was studied by the Giemsa-banding techniques. The inactive tumor line 7315i evolved from the active tumor line 7315a in one of the early transfers. A comparison was also made with chromosomal patterns of a transplantable rat hepatoma (7316A) induced by the same chemical carcinogen. Pituitary tumor line 7315a had a pseudotriploid complement with 63 chromosomes, whereas the inactive line 7315i was hypodiploid, with a dominant stemline of 36 chromosomes. The stemline chromosome number of the hepatoma was in the hypotetraploid region. Giemsa banding revealed that all chromosomes of the normal rat complement were present in both pituitary tumors. Four abnormal chromosomes were detected in the active tumor line and three in the inactive line. Both tumor lines contained a single minute chromosome. One common marker, a deleted chromosome No.1, was found in both pituitary tumors. This common marker chromosome was, however, not present in the hepatoma. The stemline karyotype of the hepatoma contained seven different markers, but none of them was identical to the abnormal chromosomes of the pituitary neoplasms. The findings suggested that the abnormal karyotypes of lines 7315a and 7315i could reflect the multisecretory activities of these neoplastic pituitary cells but that chromosomal localization of the secretory defect awaits exact genetic mapping of the rat chromosomes.

Chromosome banding patterns and breakpoints of three transplantable hepatomas induced in rats by aromatic amines.

The chromosome constitution of transplantable hepatomas H-35tc1, 7316A, and 8994 induced in ACI male, BUF female, and BUF male rats, respectively, by monocyclic or polycyclic aromatic amines was studied by Giemsa banding techniques. Hepatoma H-35tc1 was hyperdiploid, with a dominant stemline of 46 chromosomes. The stemline of the heterogeneous 7316A was in the hypotetraploid region (75-8,). Hepatoma 8994 had a near-triploid complement; most metaphase cells and chromosome numbers from 62 to 68. Thirty marker chromosomes were detected. Nineteen rearanged chromosomes were in hepatoma 8994, whereas only 8-10 markers could be found in the 80 +/- 3 chromosome complement of hepatoma 7316A. Two constant common markers were noted: mar2, a chromosome No. 11 with translocation short arm in H-35tc1 and 8994, and mar10 (mar10a), a chromosome No. 1 with a duplicated segment at breakpoint q54 in hepatomas 7316A and 8994. An analysis of the distribution of 232 breakpoints in the rat karyotype, 26 identified in the hepatomas and 206 collected from the literature, revealed a statistically significant excess of breaks in chromosomes No. 1, 2, 3, and 10 (P less than 0.001). The X- and Y-chromosomes showed a considerably lower number of breaks than anticipated (P less than 0.01). Even after a history of continuous transplantation for 10 years, these 3 hepatomas retained intact sex chromosomes that corresponded to the phenotype of the primary host. Preservation of the sex chromosomes in the hepatomas may be attributed to a lower susceptibility of the sex chromosomes than of the autosomes to breakage.

Response of microbodies in Morris hepatoma 9618A to clofibrate.

Regulation of the formation of microbodies in Morris hepatoma 9618A was studied by examination of the response of the organelles to clofibrate. The fine structures of microbodies in the hepatoma cells closely resembled those in hepatocytes of normal adult rats. In clofibrate-treated rats, the tumor cells showed a slight increase in the size of microbodies and in catalase activity; however, the tumor microbodies did not increase in number. In contrast, in adult clofibrate-treated rats and rats on the day of birth whose mothers received clofibrate during the gestation period, the hepatocytes showed microbodies that were greater in both number and size, and the catalase activity in the liver was definitely elevated.

Nuclear magnetic resonance studies of cancer. VI. Relationship among spin-lattice relaxation times, growth rate, and water content of Morris hepatomas.

Relaxation time(s) (T1), growth rates, and the water content of six Morris hepatomas, several murine tumors, and a selection of normal tissues indicated that malignant tumors did not always exhibit longer T1 values than any normal tissue, as previously suspected. This overlap raised the possibility of confusion between normal and malignant tissues studied by this method. Tissue T1 values depended primarily on the hydration of the tissue and correlated well with water content determinations. A rough correlation of T1 and water content with tumor growth rate was also observed.

Depressed growth of Morris hepatomas in altitude- and heat-stressed but not in cold-stressed buffalo rats.

Female inbred BUF rats bearing Morris hepatomas 5123C, 5123D, 7795, and 7800 bilaterally in the femoral musculature were exposed for 3 weeks to ether 4,500-m simulated altitude or sea level or to an ambient temperature of either 7, 23, or 33 degrees C. Rats were given inoculations 12 days before these exposures. Tumor size, body weight, food consumption, and body temperatures were measured weekly in these treated rats and in normal rats. At time of killing, tumor mass, DNA synthesis (by [3H]thymidine incorporation), and respiration (by conversion of [1,4-14C]succinic acid to 14CO2) were measured in each of the 4 hepatoma lines, in the livers of normal and host rats, and in regenerated livers 10 days post 70% hepatectomy. Growths of all 4 tumors and regenerated livers were significantly impaired in rats stressed by exposure to altitude and heat but not to cold. Neither DNA synthesis nor respiration was altered in the hepatomas and livers by any environmental stress. The environmentally stressed rats gained weight at a slower rate and consumed less food than did their controls, but no differences were found in these variables for tumor-bearing and non-tumor-bearing rats. However, whereas the ratio of body weight gain to food consumed was reduced under the three stressful environments, that of tumor weight gain to food consumed was not altered by any environment. Host survivorship was not influenced by any of these effects.

Glutathione and gamma glutamyl transpeptidase in rat liver during chemical carcinogenesis.

Continued administration of several hepatocarcinogens led to an increase in the concentration of glutathione (GSH) in the livers of intact, but not of hypophysectomized or adrenalectomized rats. The concentration of GSH remained high untill the development of hyperplastic nodules. Subsequently, the concentration of GSH dropped to the normal level or below. A single dose of 3'-methyl-4-dimethylaminoazobenzene (3'-Me-DAB) produced an increase of GSH which, within a certain range, depended upon the amount of the carcinogen. In well differentiated, slowly growing hepatomas, the concentration of GSH approached the level in normal adult rat liver. On the other hand, in nondifferentiated and rapidly growing hepatomas, GSH was only 30-40% of that in normal liver. The activity of gamma-glutamyl transpeptidase (GTase) increased within 24-48 hours after a single large dose of 3'-Me-DAB. Continued feeding of 3'-Me-DAB led to an exponential increase of GTase. During hepatocarcinogenesis, the level of GTase activity corresponded to the degree and size of pathologic changes produced in rat liver. Chloramphenicol partially inhibited the increase of GTase induced by 2-acetylaminofluorene. Pretreatment with 3-methylcholanthrene partially inhibited the increase of GTase that had been induced by a single dose of 3'-Me-DAB. Puromycin partially inhibited the increase of GTase induced by several doses of dimethylnitrosamine. These observations indicated a close connection between the activation of GTase and chemical carcinogenesis in rat liver. Measurements of GTase activity in 12 Morris hepatomas supported this conclusion; their GTase levels were greatly elevated compared with that in normal adult rat liver.

Effects of dietary vitamin B6 on the in vitro inactivation of rat tyrosine aminotransferase in host liver and Morris hepatomas.

Control rats or rats bearing Morris hepatoma 5123C (intact), 5123C (adrenalectomized), 7794A, 7800, 8999, 9121, or 9618A were fed a purified diet either deficient or adequate for vitamin B6. The concentration of pyridoxal phosphate in the plasma, host livers, and hepatomas was determined, as well as the in vitro rate of inactivation of induced tyrosine aminotransferase in homogenates of host livers and hepatomas. The results demonstrated the presence of a cysteine-independent inactivating system for tyrosine aminotransferase in hepatomas 5123C (adrenalectomized), 7800, 8999, and 9121. Only in hepatoma 9121 was there a dramatic influence of the dietary vitamin B6 on the rate of cysteine-independent inactivation. A cysteine-dependent inactivating system for the enzyme was present in all host livers and hepatomas. The rate of this in vitro inactivation for both host livers and hepatomas apparently was a function of the concentration of pyridoxal phosphate, but inactivation of tyrosine aminotransferase occurred at a significantly lower concentration of pyridoxal phosphate in the hepatomas than in the host livers.

Regulation of the adenylate cyclase system in transplantable hepatomas.

Adenylate cyclase systems were examined in purified membrane preparations from normal rat liver and several Morris hepatomas with differing growth rates. All tumor membrane preparations had lower relative specific activities than did liver preparations. Liver adenylate cyclase was stimulated by fluoride, glucagon and guanyl-5'-yl imidodiphosphate [Gpp(NH)p]. Membranes from two slow-growing hepatomas (hepatomas 20 and 21) contained adenylate cyclase activities which are also stimulated by each of these three modulators. Membrane adenylate cyclases from several fast-growing hepatomas (hepatomas 3924A, 7777, 5123tc, and 9618A2) were marginally stimulated by glucagon but were readily stimulated by fluoride and Gpp(NH)p. Examination of the highly specific binding of 125I-glucagon to the various membrane preparations revealed much less binding in all the tumor membranes than in liver membranes. More detailed kinetic examination of membranes prepared from liver, slow-growing hepatoma 21 (which had reasonable binding to and stimulation by glucagon), and fast-growing hepatoma 3924A (which had marginal binding to and stimulation by glucagon) revealed major differences in rates of cyclic adenosine 3':5'-monophosphate production in the absence and presence of glucagon, Gpp(NH)p, and glucagon plus Gpp(NH)p and in the combined alteration of magnesium:adenosine 5'-triphosphate ratio and temperatures. The different kinetic characteristics in the hepatoma adenylate cyclase systems may be due to different structural characteristics of the tumor membranes or may be due to altered hormonal receptors, catalytic units, or receptor-catalytic unit interrelationships within the tumor membrane.

Superoxide dismutase and superoxide radical in Morris hepatomas.

Total superoxide dismutase (SOD) and manganese superoxide dismutase (Mn SOD) specific activities were measured in tissue homogenates and in isolated mitochondria from normal rat liver and three Morris hepatomas of different growth rates. Total SOD and Mn SOD specific activities were decreased in all tumor homogenates when compared to normal liver; the lowest activity was associated with the fastest growing tumor. These results are consistent with the hypothesis that total Mn SOD specific activity is decreased in all tumors. The Mn SOD specific activity was similar to the total SOD specific activity of isolated mitochondria, indicating that mitochondrial SOD is almost entirely manganese containing. This activity was decreased in the fast- and medium-growth-rate hepatomas but was slightly increased in the tumor with the slowest growth rate when compared to liver. The normal or higher than normal mitochondrial Mn SOD specific activity indicates that decreased mitochondrial SOD specific activity is not a characteristic of all tumors. Superoxide radical (O2-.) formation was measured in submitochondrial particles obtained by sonication of isolated mitochondria and subsequent washings to remove the SOD. The difficulty encountered in reducing the SOD activity suggests that at least part of the mitochondrial SOD might be associated with the mitochondrial membrane. In liver submitochondrial particles, O2-. was formed only when succinate and antimycin A were used together, as substrate and inhibitor of the electron transport chain, respectively. In the hepatomas studied for O2-. production (slow- and fast-growth rates), the formation of the radical was detected in the presence of succinate even when no inhibitor was present. Antimycin A stimulated the production of O2-. in normal rat liver and slow-growth-rate tumor, but not in fast-growth-rate tumor submitochondrial particles. Reduced nicotinamide adenine dinucleotide did not lead to the production of O2-. by normal liver or hepatoma submitochondrial particles. Mitochondrial membrane damage was seen in micrographs of the medium- and fast-growing hepatomas. This could be a consequence of low mitochondrial SOD concomitant with a flux of superoxide, if the radical is produced in vivo by these mitochondria.

Differences in total mitochondrial proteins and proteins synthesized by mitochondria from rat liver and Morris hepatomas 9618A, 5123C, and 5123tc.

Mitochondria were isolated from a slow-growing (9618A) and two intermediate-to-fast-growing (5123C, 5123tc) Morris hepatomas and host livers. The mitochondrial proteins were solubilized and fractionated on sodium dodecyl sulfate:polyacrylamide slab gels. One Coomassie blue-stained band was absent or reduced in amount in all tumors relative to host livers. In addition, a major mitochondrial enzyme present in normal liver, carbamyl phosphate synthetase, was missing or greatly reduced in the slow-growing, highly differentiated hepatoma 9618A, a tumor that is considered to be similar to normal liver in many biochemical and morphological respects. Incubation of mitochondria with [35S]methionine and a suitable amino acid incorporation system resulted in labeling of specific mitochondrial proteins. Autoradiography of the slab gels disclosed four prominently labeled fractions and a number of minor fractions. Preparations from hepatoma 5123tc demonstrated two labeled bands that were absent or greatly reduced in host liver. Host liver preparations displayed a minor band that was absent or greatly reduced in hepatoma 5123C. However, no single change in labeling pattern was common to all three tumors, suggesting the absence of a causal relationship between carcinogenesis and mutations in mitochondrial DNA.

Behavior of transaldolase (EC 2.2.1.2) and transketolase (EC 2.2.1.1) Activities in normal, neoplastic, differentiating, and regenerating liver.

The objective of this investigation was to throw light on the biological behavior and metabolic regulation of hepatic enzymes of the nonoxidative branch of the pentose phosphate pathway. The activities of transaldolase (EC 2.2.1.2) and trasketolase (EC 2.2.1.1) Were compared in biological conditions that involve modulation of gene expression such as in starvation, in differentiation, after partial hepatectomy, and in a spectrum of hepatomas of different growth rates. The enzyme activities were determined under optimal kinetic conditions by spectrophotometric methods in the 100,000 X g supernatant fluids prepared from tissue homogenates. The kinetic properties of transaldolase and transketolase were similar in normal liver and in rapidly growing hepatoma 3924A. For transaldolase, apparent Km values of 0.13 mM (normal liver) and 0.17 mM (hepatoma) were observed for erythrose 4-phosphate and of 0.30 to 0.35 mM for fructose 6-phosphate. The pH optima in liver and hepatoma were at approximately 6.9 to 7.2. For the transketolase substrates, ribose 5-phosphate and xylulose 5-phosphate, the apparent Km values were 0.3 and 0.5 mM, respectively, in both liver and hepatoma. A broad pH optimum around 7.6 was observed in both tissues. In organ distribution studies, enzyme activities were measured in liver, intestinal mucosa, thymus, kidney, spleen, brain, adipose tissue, lung, heart, and skeletal muscle. Taking the specific activity of liver as 100%, transaldolase activity was the highest in intestinal mucosa (316%) and in thymus (219%); it was the lowest in heart (53%) and in skeletal muscle (21%). Transketolase activity was highest in kidney (155%) and lowest in heart (26%) and skeletal muscle (23%). Starvation decreased transaldolase and transketolase activities in 6 days to 69 and 74%, respectively, of those of the liver of the normal, fed rat. This was in the same range as the decrease in the protein concentration (66%y. In the liver tumors, transaldolase activity was increased 1.5- to 3.4-fold over the activities observed in normal control rat liver. Transketolase activity showed no relationship to tumor proliferation rate. In the regenerating liver at 24 hr after partial hepatectomy, the activity of both pentose phosphate pathway enzymes was in the same range as that of the sham-operated controls. In differentiation at the postnatal age of 5, 12, 23, and 32 days, hepatic transaldolase activities were 33, 44, 55, and 72%, respectively, of the activities observed in the 60-day-old, adult male rat. During the same period, transketolase activ-ties were 18, 21, 26, and 55% of the activities observed in liver of adult rat. The demonstration of increased transaldolase activity in hepatomas, irrespective of the degree of tumor malignancy, differentiation, or growth rate, suggests that the reprogramming of gene expression in malignant transformation is linked with an increase in the expression of this pentose phosphate pathway enzyme...

Acetoacetate coenzyme A transferase activity in rat hepatomas.

The presence of succinyl-coenzyme A:acetoacetate CoA transferase (CoA transferase) (EC 2.8.3.5), an initiator of ketone body utilization in nonhepatic tissue, was examined in liver from normal, partly hepatectomized, neonatal, and tumor-bearing rats, as well as in a series of transplantable rat hepatomas ranging widely in growth rate. While levels of CoA transferase are extremely low in normal, host, and regenerating liver, considerable amounts of activity are detectable in neonatal liver and in the hepatomas. In fact, the content of CoA transferase in the series of Morris hepatomas increases progressively with increase in tumor-growth rate. The fastest-growing tumor studied (7288Ctc) contains about the same amount of CoA transferase activity as rat skeletal muscle (i.e., an activity of about 0.1 mumole of acetoacetate used per min per g tissue). These results clearly indicate that the faster-growing hepatomas have adequate capacity to utilize ketone bodies in bioenergetic or biosynthetic activities. Furthermore, the enzymes from normal and hepatoma 7288Ctc tissues are quite similar with respect to (a) size of about 10(5) daltons, (b) reaction mechanism requiring formation of an enzyme:CoA intermediate (from ping-pong kinetic data), and (c) various kinetic parameters (such as Michaelis constants, product competitive inhibition constants, and acetoacetate substrate inhibition). The enzymes from rat skeletal muscle and Morris hepatoma 7288Ctc have the same isoelectric point (7.6), which differs from that for the rat heart enzyme (6.8).

Tumor-selective inhibition of the incorporation of 3H-labeled amino acids into protein by cyanate.

Sodium cyanate at a dose level of 125 or 250 mg/kg i.p. caused an inhibition of incorporation of 3H-labeled amino acids into cytoplasmic and nuclear proteins of the rapidly growing hepatoma 7777 and the slowly growing hepatoma 9618A. There was no inhibitory effect on 3H-labeled amino acid incorporation into protein in the livers of rats bearing these tumors. Studies on the effects of sodium cyanate on incorporation of 3H-labeled amino acids into total acid-insoluble material indicated that a greater than 85% inhibition could be achieved in hepatoma 5123C, hepatoma 9618A2, and the MK3 kidney tumor with either little or no effect in host liver, kidneys, brain, skeletal muscle, intestinal mucosa, and regenerating liver after partial hepatectomy.

Nuclear protein changes in rat hepatomas correlating with growth rate.

The nature of nuclear proteins that are soluble in 8 M urea-50 mM phosphate, pH 7.6, was compared in rat liver and Morris hepatomas, Isoelectric focusing, using carrier ampholytes for a pH gradient of 3.5 to 10, indicated that with increasing growth rate of the hepatomas there was a progressive tendency for a decrease in nonhistone nuclear proteins with isoelectric points in the range 7.5 to 8.9 and an increase in the range 5.1 to 6.7. Studies on the influence of time on the pH gradient revealed that a nonuniform drift provided a better resolution of the pH range 7.5 to 8.9 at 7 hr than at 24 hr, while the latter time for electrofocusing gave an improved resolution of the pH range 5.1 to 6.7 Polyarcylamide gel electrophoresis in a urea-acetic acid system showed that 8 M urea-50 mM phosphate; pH 7.6 extracted a small part of the histones from nuclei of both liver and hepatomas. There was less extraction of histones from the hepatoma nuclei, especially in two rapidly growing hepatomas with the most notable difference being seen in the lysine-rich H1 histone. The results suggested that in addition to qualitative or quantitative changes in nonhistone nuclear proteins in liver cancer there are alterations in the binding of histones to chromatin.

Positive correlation between calmodulin content and hepatoma growth rates.

Calmodulin contents of normal rat liver, host liver [bearing hepatoma 5123t.c.(h)], regenerating liver, and Morris hepatomas 7800, 5123t.c.(h), and 7794A were determined by phosphodiesterase assay and by radioimmunoassay. The calmodulin levels determined by both assays were significantly increased in three hepatomas when compared to the corresponding values of normal liver. The order of increase in calmodulin content was as follows: normal liver = host liver less than 7794A (slow growth rate) less than 5123t.c.(h) (intermediate growth rate) less than 7800 (fast growth rate). In regenerating liver (24 hr after partial hepatectomy), the calmodulin content was not different from that of normal liver. In good agreement with the literature, the calmodulin values measured by the phosphodiesterase assay were always lower than those determined by radioimmunoassay. Calcium and magnesium contents were measured by atomic absorption spectrophotometry in acid digests of these tissues. Both cation contents were significantly increased in the three hepatomas studied when compared to the corresponding values of normal liver; the extent of increase for calcium content (120 to 240%) was much greater than that for magnesium (30 to 40%). The order of increase for both cations was as follows: normal liver = host liver less than 5123t.c.(h) less than 7794A less than 7800. Therefore, there does not appear to be any correlation between the cation contents and hepatoma growth rates. In regenerating liver, magnesium content was about 14% higher than that of normal liver. In summary, the results indicate that only the increase of calmodulin appears to correlate positively with the growth rate of these tumors. This correlation suggests that calmodulin may be involved in tumor cell growth regulation.