Ernesto Bustamante

A tumorigenic anchorage-dependent cell line (H-91) was established in culture from an azo-dye-induced rat ascites hepatoma. When grown in a glucose-containing medium the cells exhibit high rates of lactic acid production characteristic of rapidly growing tumor cells. However, when glucose is replaced with galactose the cells grow equally well but exhibit only moderately elevated rates of lactic acid production. The molecular basis for this observation cannot be attributed to differences in permeability because initial rates of glucose and galactose entry into hepatoma cells are identical. Rather, the activity of hexokinase (ATP:D-hexose 6-phosphotransferase, EC 2.7.1.1) is found to be high in hepatoma cells, about 20-fold higher than that of control and regenerating rat liver. Moreover, tumor hexokinase activity is not inhibited by low concentrations (<0.6 mM) of the reaction product glucose 6-phosphate. Additionally, 50% of the hexokinase activity of hepatoma cells is found associated with the mitochondrial fraction. This fraction is 3-fold enriched in hexokinase activity relative to the homogenate and 4-fold enriched relative to the nuclear and postmitochondrial fractions. Tumor mitochondrial hexokinase appears to be coupled directly to oxidative phosphorylation, because addition of glucose to respiring hepatoma mitochondria (after a burst of ATP synthesis) results in stimulation of respiration. In contrast, glucose has no effect on the respiration of mitochondria from control and regenerating liver. These results suggest that the high glycolytic capacity of H-91 hepatoma cells is due, at least in part, to an elevated form of hexokinase concentrated in the mitochondrial fraction of the cell.

Rat liver cytoplasm (postnuclear supernatant) has a low aerobic glycolytic rate in the presence of added glucose, ATP, ADP, Pi, and NAD+, whereas cytoplasm from Ehrlich ascites tumor cells exhibit a high aerobic glycolytic rate which is typical of rapidly proliferating tumor cells. Tumor mitochondria, unlike liver mitochondria, contain bound hexokinase which constitutes about 70% of the total cellular hexokinase activity. The high aerobic glycolytic rate of Ehrlich tumor cytoplasm is reduced markedly if the mitochondria are removed and can be restored almost completely upon addition of the hexokinase-containing tumor mitochondria to tumor cytosol (postmitochondrial supernatant). Addition of tumor mitochondria to liver cytosol can enhance its glycolytic rate to levels approaching those of tumor cytoplasm, whereas added liver mitochondria are without effect on the already low glycolytic rate of liver cytosol. Addition of tumor mitochondria to tumor cytosol increases its glycolytic rate to the level of tumor cytoplasm, as mentioned above, but liver mitochondria added to tumor cytosol actually depress its glycolytic rate to the level of liver cytosol. The stimulatory effect of tumor mitochondria on liver cytosol can be ascribed to its associated hexokinase activity since hexokinase specifically removed from mitochondria of tumor cells can also enhance the glycolytic rate of liver cytosol. The depressing effect of added liver mitochondria on tumor cytosol glycolysis suggests that liver mitochondria can compete more effectively than tumor mitochondria for a common intermediate and/or cofactor. Examination of 12 different tumor cell lines revealed that only those which reached maximum size in 1 month or less, and which have elevated glycolytic activities, had detectable mitochondrially associated hexokinase activity. The studies reported here describe resolution and reconstitution of tumor cytoplasm, supplementation of cytosol with intact mitochondria or mitochondrial hexokinase, and a survey of mitochondrial hexokinase content in various tumors, and provide strong evidence for the view (Bustamante, E., and Pedersen, P. L. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 3735-3739) that a form of hexokinase with a propensity for mitochondrial binding plays a key role in the high aerobic glycolysis of cancer cells.

The highly glucolytic hepatoma cell line H-91 is characterized by a high hexokinase activity to rat liver; 50% of this activity is associated with the mitochondrial fraction [Bustamante, E., & Pederson, P.L. (1977) Proc. Natl. Acad. Sci. U.S.A. 74, 3735--3739]. Treatment of mitochondria from this cell line with adenosine 5'=triphosphate (ATP) or glucose 6-phosphate solubilizes bound hexokinase activity. Solubilization of the enzyme by ATP results in a six- to sevenfold purification. Free ATP, unchelated by Mg ions, induces the release of the enzyme from the membrane, whereas the MgATP complex is ineffective. Ethylenediaminetetraacetic acid (EDTA) fails to release mitochondrial hexokinase indicating that the enzyme is not attached to the membrane by divalent cations. Energization of mitochondria is not required for ATP to induce solubilization of bound hexokinase. This is evidenced by (a) the ability of the nonhydrolyzable ATP analogue adenylyl imidodiphosphate to solubilize the enzyme, (b) the inability of uncouplers and inhibitors of oxidative phosphorylation to either solubilize or prevent the release of mitochondrial hexokinase, and (c) the inability of atractyloside to solubilize or prevent the release of bound hexokinase. The bound and the ATP-solubilized forms of mitochondrial hexokinase from H-91 hepatoma cells are kinetically different. When membrane bound, the enzyme has a significantly higher apparent affinity (Km = 0.25 mM) for its substrate MgATP than when solubilized (Km = 1.2 mM). Free ATP acts as a competitive inhibitor of mitochondrial hexokinase. Both the membrane-bound and the solubilized forms of mitochondrial hexokinase have about the same apparent affinity for glucose (Km = 56 and 83 microM, respectively). The experiments reported here provide the first description of the properties and the nature of binding of mitochondrial hexokinase from a tumor cell line growing in tissue culture.