Main

Ketone bodies are important vectors of energy transfer(1). Ketogenesis primarily occurs in liver. The relative capacities of tissues for ketone body utilization is thought to be determined by their levels of ketolytic enzymes. Ketosis is a normal physiologic response to fasting and other stresses by which fat-derived energy is provided to extrahepatic tissues, particularly brain, which has no other non-glucose-derived source of energy. Ketoacidosis, a potentially life-threatening condition, can develop in uncontrolled diabetes and in some organic acidemias, including those resulting from a primary deficiency of ketone body using enzymes. We studied the tissue distribution of three enzymes that mediate ketone body utilization. The hereditary deficiency of each of these enzymes has been associated with ketosis in humans(1–13).

Figure 1 summarizes current views of ketone body metabolism. In liver and in other tissues that use fatty acids for energy, acetoacetyl-CoA and acetyl-CoA are produced mainly through β-oxidation. In liver, ketogenesis occurs via the HMG-CoA pathway, resulting in the export of acetoacetate for use in other tissues. An enzyme of β oxidation, T2(EC 2.3.1.9) catalyzes the interconversion of acetyl-CoA and acetoacetyl-CoA and is essential for efficient ketogenesis from fatty acids. In extrahepatic tissues, the mitochondrial enzyme SCOT (EC 2.8.3.5) activates acetoacetate to acetoacetyl-CoA, which is converted by T2 into acetyl-CoA for use in energy production. Hence, T2 is active in both ketogenesis and ketolysis(14), whereas SCOT is active only in ketolysis(15). The third enzyme examined in this report, CT (EC 2.3.1.9), produces acetoacetyl-CoA for the synthesis of cholesterol and other isoprenoid compounds in both liver and extrahepatic tissues. For simplicity the cytoplasm is shown only for extrahepatic cells.

Figure 1
figure 1

Diagram of ketone body metabolism, showing the roles of T2, CT, and SCOT. For simplicity the cytoplasm is shown only for extrahepatic cells, and the biochemical reactions show only the path of the carbon skeletons of the ketone bodies. Abbreviations: AcAc, acetoacetate;AcAcCoA, acetoacetyl-CoA; AcCoA, acetyl-CoA; AS, acetoacetyl-CoA synthetase; FA-CoA, fatty acyl-CoA;HMG-CoA, 3-hydroxy-3-methylglutaryl-CoA; HL, HMG-CoA lyase; mHS, mitochondrial HMG-CoA synthase.

Full size image

Activity levels of these three ketolytic enzymes have been reported in rat tissues(14, 15). However, studies in humans have been limited to enzyme assay in a small number of tissues(5, 16). We have obtained human cDNAs and have raised antibodies to human and/or rodent polypeptides for T2(17, 18), and recently to human CT(19, 20) and SCOT(21, 22). To better understand human ketone body utilization, we studied the distributions of T2, CT, and SCOT proteins and mRNA in different human tissues.

METHODS

Patients. We examined autopsy specimens from three patients who died of diseases which have no known relationship to ketone body metabolism. Patient 1 (Hunter syndrome) died of pneumonia at 13 y of age; hepatosplenomegaly and cardiomegaly were present. Patient 2 (Alagille syndrome) died suddenly, at 10 y of age, of massive bleeding from esophageal varices; hepatic cirrhosis was found at autopsy. Patient 3 (GM1 gangliosidosis) died suddenly at 4 y of age of respiratory arrest. Mild hepatosplenomegaly and bronchopneumonia were present. In all cases, tissues were frozen within 10 h after death and stored at -80 °C until use.

Northern blot analysis. We used a commercially available filter containing 2 μg of poly(A+) RNA of multiple human tissues (Clontech Laboratories, Inc.). The filter was sequentially probed and stripped according to the manufacturer's protocol, using radiolabeled full-length T2, SCOT, and CT cDNAs, in that order.

Immunoblot analysis. Small pieces of human tissues were freeze-thawed and homogenized in 10 volumes of 50 mM sodium phosphate (pH 8.0), 0.1% Triton X-100, then centrifuged at 10 000 × g for 10 min. Protein concentration was determined in the supernatant by the method of Lowry(23) using BSA as a standard. Between 0.5 and 45μg of protein were used for SDS-PAGE and immunoblot analysis(24, 25), using the ProtoBlot Western blot AP Systems (Promega, Madison, WI). In this analysis, we used previously described rabbit polyclonal antibodies: anti-rat T2(18), anti-human CT(19, 20), and anti-recombinant human SCOT(22). As first antibody for immunoblotting, we used either a mixture of the anti-human CT antibody and anti-rat T2 antibody or a mixture of the T2 antibody and anti-human SCOT antibody. For estimation of the amounts of T2 and CT in tissues, known amounts of purified human T2 and CT were included as a standard; for SCOT, we used the recombinant human SCOT protein for which we developed the antibody(22).

To determine the amounts of T2, SCOT, and CT in different tissues, a graded series of known amounts of purified enzymes and different amounts of tissue homogenates were applied to the same immunoblot membrane. Peak areas of these bands were compared by densitometry using an ACD-18 automatic computing densitometer (Gelman). For example, for hepatic T2, 2.5, 5, 10, and 20 ng of purified T2 and 0.5, 1, 2, and 4 μg of liver protein were compared on the same gel.

RESULTS

Northern blot analysis. mRNAs for CT, T2, and SCOT were detected as single bands of about 1.7, 1.7, and 4 kb, respectively (Fig. 2). CT mRNA was most abundant in liver and was also easily detectable in high levels in brain, lung, and kidney. T2 mRNA was abundant in skeletal muscle, liver, heart, and kidney. On long exposure, CT and T2 mRNAs were detected in all tissues tested. The SCOT mRNA level was high in heart, brain, skeletal muscle, and kidney, but as expected was not detected in liver, even after prolonged exposure.

Figure 2
figure 2

Northern blot analysis of poly(A+) RNA from multiple human tissues was hybridized with radiolabeled full-length human CT, T2, or SCOT cDNAs, as indicated. The origins of the mRNA were as follows:H, heart; B, brain; Pl, placenta; Lu, lung; Li, liver; M, skeletal muscle; K, kidney;Pa, pancreas.

Full size image

Immunoblot analysis. Figure 3 shows a typical immunoblot result in which 15 μg of tissue proteins of case 1 were studied using a mixture of anti-rat T2 and anti-human CT antibodies (Fig. 3A), and also a mixture of anti-rat T2 and anti-human SCOT antibodies (Fig. 3B).Figure 4 summarizes the T2, SCOT, and CT content of the various tissues. The tissue distributions of CT, T2, and SCOT were similar among the three cases, although there was some intersubject variation. Notably, T2 and CT were somewhat lower in the liver of patient 2 than in the other patients, which may reflect the presence of cirrhosis.

Figure 3
figure 3

Immunoblot analysis of various tissues from case 1. Tissue samples corresponding to 15 μg of protein were analyzed. Lanes:Li, liver; H, heart; Bl, cerebrum; B2, cerebellum; K, kidney; A, adrenal gland; M, skeletal muscle; Pa, pancreas; Lu, lung; F, fibroblasts; Ly; lymphocytes. Fibroblasts and lymphocytes were from a normal control unrelated to case 1. (A) A mixture of anti-human CT antibody and anti-rat T2 antibody was used as the first antibody. Lanes 1, 2, and 3, respectively, contain 5, 10, and 15 ng each of purified human T2 and human CT. The positions of the CT and T2 subunits are indicated by arrows. (B) A mixture of anti-[human SCOT] antibody and anti-[rat T2] antibody was used as the first antibody. Lanes 1, 2, and 3 contain 5, 10, and 15 ng of purified human T2 and 0.5, 1, and 2 ng of a purified recombinant human SCOT protein, respectively. The positions of T2 and SCOT subunits are indicated by arrows.

Full size image
Figure 4
figure 4

Tissue content of CT, T2, and SCOT as calculated from quantitation of immunoblots. Tissue samples from cases 1-3 were analyzed, as were three normal control fibroblast and lymphocyte lines. Protein content is expressed as μg/mg tissue protein.

Full size image

Immunoreactive CT and T2 were detected in all tissues examined. For CT, liver had the highest concentration, followed by adrenal glands, kidney, brain, and lung, each of which were ≈4-fold less than that of liver. Skeletal and heart muscle had the lowest amounts of CT. T2 was most abundant in liver, but was easily detected in kidney, heart, adrenal glands, and skeletal muscle.

SCOT was detected in all tissues except liver and was most abundant in myocardium. Brain, kidney, and adrenal glands also had substantial amounts of SCOT. With two exceptions the tissue distributions of T2 and SCOT were similar. The most striking difference is in liver in which T2 is greatest and SCOT is undetectable. Second, the SCOT/T2 ratio was higher in brain than in other tissues. A similar pattern is seen at the mRNA level (Fig. 2).

DISCUSSION

This is the first reported survey of the protein and mRNA levels of T2, SCOT, and CT in human tissues. Although the tissues came from patients with diverse diseases, the tissue distribution of immunoreactive T2, SCOT, and CT was similar in all patients, suggesting that neither the presence of these diseases nor frozen storage had a marked effect upon tissue peptide levels.

These data will be particularly useful in the case of thiolases because at least four different thiolases can use acetoacetyl-CoA as substrate: T2, CT, mitochondrial medium-chain 3-ketoacyl-CoA thiolase (T1), and peroxisomal 3-ketoacyl-CoA thiolase(18). These enzymes cannot be distinguished from one another by activity assay but are antigenically distinct. The data presented here are therefore useful indices of the capacity and pathway of ketone body use in human tissues.

Regarding SCOT, a technical consideration is that the antibody was raised to a denatured human SCOT polypeptide isolated from a bacterial expression system(22). We cannot eliminate the possibility that the antibodies raised against the recombinant protein may react differently with recombinant SCOT in comparison with native SCOT from human tissues, leading to a systematic error in our estimates of the absolute amount of SCOT polypeptide in human tissues. However, this should not influence estimates of the relative amounts present in different tissues.

For all three enzymes, the levels of mRNA and polypeptide were proportional in all tissues studied, suggesting that transcriptional regulation may be important for these enzymes. An exception was skeletal muscle, in which SCOT and T2 mRNAs were more intense than expected for the amount of their specific protein observed, perhaps due to the high concentration of structural proteins present in skeletal muscle. Although skeletal muscle expresses only moderate amounts of SCOT and T2, because of the large muscle mass of humans, it is a major consumer of ketone bodies. One half of T2-deficient patients have mild muscle weakness during metabolic decompensation(2, 3), which we speculate may be due to in part to an accumulation of T2 substrates in muscle.

The principal discordance between T2 and SCOT levels was observed in the liver, in which large amounts of T2, but no detectable SCOT, were present. This supports current views of ketone body metabolism; because SCOT activates ketone bodies for use as energy substrates in mitochondria, its level is thought to reflect the ketolytic capacity of tissues. Liver mitochondria generate large amounts of ketone bodies but do not consume them, and the suppression of SCOT expression in liver can be viewed as a mechanism to avoid futile cycling. In contrast to SCOT, T2 is active both in ketogenesis and ketolysis (Fig. 1).

A discrepancy between SCOT and T2 levels was also seen in brain, in which levels of SCOT mRNA and immunoreactive protein were higher with respect to T2 than in other tissues. We speculate that this may reflect the low capacity of brain to oxidize fatty acids(26), a process which is active in the other tissues studied and in which T2 but not SCOT is involved. To obtain lipid-derived energy, the brain relies nearly entirely upon ketolysis, which requires both SCOT and T2. Ketone bodies furnish the only substantial non-glucose-derived energy supply to mammalian brain and are important at times of glucopenia(28, 29); during fasting, ketone bodies can provide up to 60% of brain energy in humans(27).

Developmental changes in the enzymes of ketone body metabolism may also be pertinent to our results. For instance, the activity of SCOT is greater in 3-10-y-old subjects than in infants (0-2 y) or adults aged 37-69 y(16). In the same report, acetoacetyl-CoA thiolase activity was highest in adults(16). The high levels of SCOT with respect to T2 in our subjects aged 4-13 y may be explained in part by age-specific changes in SCOT and T2. In rat brains, the levels of T2 and CT activities were also reported to change in relation to the age; CT level is highest at birth and falls slowly to the adult level, and T2 level rises rapidly during suckling and declines to the adult level after weaning(30). Further studies will be required to examine whether such developmental or physiologic changes occur in human brain.

In other tissues, the relative amounts of T2 and SCOT are similar. SCOT was present in large amounts in heart, as expected from data in rats(15) and by our preliminary Northern blot results(21), consistent with the heart being a major consumer of ketone bodies in humans as well as rodents. Of note, congestive cardiomyopathy has occurred in at least one patient with hereditary T2 deficiency(31), perhaps because of secondary low free CoA levels, which are known to reduce cardiac function(32). Cardiomegaly was also noted in two of nine SCOT-deficient patients reported to date(5, 6). In human kidney, SCOT and T2 levels were considerably less than that in the in heart. This contrasts with rat kidney, in which SCOT and T2 levels are nearly equal to those of heart(14, 15) and for which ketone bodies are a major energy source(33).

CT is important for the biosynthesis of cholesterol and fatty acids(15); its role in energy metabolism is generally felt to be minor. The levels of CT mRNA and polypeptide were higher in tissues known to synthesize cholesterol (liver > brain, lung, adrenal) than in skeletal muscle or myocardium. CT was also present in large amounts in the kidney. The high levels CT in human liver is consistent with compartmentalization of acetoacetate metabolism within human hepatocytes, with ketogenesis in mitochondria, and ketolysis in cytoplasm, to obtain acetyl-CoA for cholesterol synthesis. Of note, rat hepatocytes also contain the enzyme for cytoplasmic activation of acetoacetate, cytoplasmic acetoacetyl-CoA synthetase(34). The level of CT in adrenal gland with respect to liver is lower in humans than was reported for the rat(15), in which adrenal levels are equal to those in liver. In general, the tissue distribution of CT reflects the capacity for cholesterol synthesis rather than that of ketolysis. In this context, it is difficult to explain why the two patients reported to date to have CT deficiency(12, 13) had ketosis as well as delayed mental development. It will be of great interest to evaluate ketolysis and also hormone status of such patients identified in the future.