The intracellular site of action of insulin: the mitochondrial Krebs cycle

Rapid radiochemical filter paper assay for determination of hexokinase activity and affinity for glucose-6-phosphate

Glucose phosphorylation by hexokinase (HK) is a rate-limiting step in glucose metabolism. Regulation of HK includes feedback inhibition by its product glucose-6-phosphate (G6P) and mitochondria binding. HK affinity for G6P is difficult to measure because its natural product (G6P) inhibits enzyme activity. HK phosphorylates several hexoses, and we have taken advantage of the fact that 2-deoxyglucose (2-DG)-6-phosphate does not inhibit HK activity. By this, we have developed a new method for rapid radiochemical analysis of HK activity with 2-DG as a substrate, which allows control of the concentrations of G6P to investigate HK affinity for inhibition by G6P. We verified that 2-DG serves as a substrate for the HK reaction with linear time and concentration dependency as well as expected maximal velocity and KM. This is the first simple assay that evaluates feedback inhibition of HK by its product G6P and provides a unique technique for future research evaluating the regulation of glucose phosphorylation under various physiological conditions.

NEW & NOTEWORTHY Traditionally, hexokinase activity has been analyzed spectrophotometrically in which the product formation of glucose-6-phosphate (G6P) is analyzed by an indirect reaction coupled to NADPH formation during conversion of G6P to 6-P gluconolactone. By nature, this assay prevents measurements of hexokinase (HK) affinity for inhibition by G6P. We have developed a rapid radiochemical filter paper assay to study HK affinity for G6P by use of radiolabeled 2-deoxyglucose as substrate to study physiological regulation of HK affinity for G6P-induced inhibition.

Insulin-status-dependent modulation of FoF1 ATPase activity in rat kidney mitochondria

The early and late effects of alloxan-diabetes and insulin treatment on kinetic properties of mitochondrial FoF1 ATPase were examined. Diabetic state resulted in significant decrease in the activity while insulin treatment caused hyper-stimulation. In control animals the enzyme activity resolved in three kinetic components. In diabetic condition only component I and II were present. With insulin treatment component III was restored but component II was abolished. Diabetic state and insulin treatment had varied effects on Km values of the three components, whereas the Vmax values were generally on the higher side. Evaluation of the AppKcat/Km values revealed that diabetic state resulted in increased catalytic efficiency; insulin treatment brought back these values to normality. Temperature kinetics studies indicated that the phase transition temperature decreased significantly in the diabetic and insulin-treated diabetic animals. The energy of activation in low temperature range increased in the diabetic animals. Insulin treatment corrected the Arrhenius pattern at early stage of diabetes; at late stage the pattern was reversed. The results are suggestive of subtle insulin-status-dependent alterations in membrane structure - function relationships.

Insulin-status-dependent modulation of FoF1-ATPase activity in rat liver mitochondria

Early and late effects of alloxan diabetes and insulin treatment on mitochondrial membrane structure and function were evaluated by studying the kinetic properties of mitochondrial membrane marker enzyme FoF1-ATPase and its modulation by membrane lipid/phospholipid composition and membrane fluidity. Under all experimental conditions the enzyme displayed three kinetically distinguishable components. In 1 wk-old diabetic animals the enzyme activity was unchanged; however, K(m) and V(max) of component I increased and K(m) of component II decreased. Insulin treatment resulted in lowering of K(m) and V(max) of components II and Ill. One-mon diabetic state resulted in decreased enzyme activity, whereas insulin treatment caused hyperstimulation. K(m) of components I and II decreased together with decreased V(max) of all the components. Insulin treatment restored the K(m) and V(max) values. In late-stage diabetes the catalytic efficiency of components I and II increased; insulin treatment had drastic adverse effect. Binding pattern of ATP was unchanged under all experimental conditions. Diabetic state resulted in progressive decrease in energy of activation in the low temperature range (E(L)). Insulin treatment lowered the energy of activation in the high temperature range (E(H)) without correcting the E(L) values. The phase transition temperatures increased in diabetic state and were not corrected by insulin treatment. Long-term diabetes lowered the total phospholipid content and elevated the cholesterol content; insulin treatment had partial restorative effect. The membrane fluidity decreased in general in diabetic condition and was not corrected by insulin treatment at late stage. Regression analysis studies suggest that specific phospholipid classes and/or their ratios may play a role in modulation of the enzyme activity.

Insulin status differentially affects energy transduction in cerebral mitochondria from male and female rats

Effects of STZ diabetes and treatment with insulin on cerebral mitochondrial metabolism in the male and female rats were examined. Diabetic state resulted in generalized decrease in the state 3 respiration rates in the males with practically all the substrates except glutamate where the opposite effect was seen. Diabetic state had no adverse effect on the respiratory activity in the females. Insulin treatment had no restorative effect in the males. By contrast in the females, adverse effects were noted. The cytochromes contents decreased in STZ diabetes with the effect being more pronounced in the males; treatment with 1 unit of insulin restored the cytochromes contents. STZ diabetes also resulted in decreased dehydrogenases activities with the effect being more pronounced in the females: insulin treatment resulted in hyper-stimulation of glutamate dehydrogenase and succinate DCIP reductase activities; restoration of malate dehydrogenase activity was only partial. The results point out that STZ diabetes and insulin treatments differentially affect cerebral mitochondrial energy metabolism in the male and female rats.

Detachment of the glycolytic enzymes, phosphofructokinase and aldolase, from cytoskeleton of melanoma cells, induced by local anesthetics

Cancer cells are characterized by a high rate of glycolysis, which is their primary energy source. An important mechanism that controls glycolysis is the reversible binding of glycolytic enzymes to cytoskeleton. We report here that the local anesthetics, lidocaine and bupivacaine, induced a dose-dependent detachment of the glycolytic enzymes, phosphofructokinase (EC 2.7.1.11) and aldolase (EC 4.1.2.13), from cytoskeleton of B16 melanoma cells. The detachment of glycolytic enzymes from cytoskeleton would reduce the provision of local ATP, in the vicinity of cytoskeleton-membrane and would also affect cytoskeleton structure. We show here that the local anesthetics decreased the viability of melanoma cells. The detachment of the glycolytic enzymes from cytoskeleton, induced by the drugs, preceded melanoma cell death, which indicates that this is an early effect and not a result of cell death. Bupivacaine was more potent than lidocaine both on the glycolytic enzymes and on cell viability. The present results suggest that local anesthetics, and especially bupivacaine, are promising drugs for the treatment of melanoma.

Taxol (paclitaxel) induces a detachment of phosphofructokinase from cytoskeleton of melanoma cells and decreases the levels of glucose 1,6-bisphosphate, fructose 1,6-bisphosphate and ATP

Glucose utilization through glycolysis, which is the primary energy source in cancer cells, is known to be controlled by allosteric regulators, as well as by reversible binding of glycolytic enzymes to cytoskeleton. Here we report of a novel mechanism of action of taxol (paclitaxel; Baccatin III N-benzyl-beta-phenylisoserine ester), the anti-microtubule agent with remarkable anticancer activity. We show that taxol affects both levels of regulation of glycolysis in melanoma cells; it decreases the levels of glucose 1,6-bisphosphate and fructose 1,6-bisphosphate, the two allosteric stimulatory signal molecules of glycolysis, and also causes a detachment of phosphofructokinase (ATP: D-fructose-6-phosphate 1-phosphotransferase, EC 2.7.1.11), the rate-limiting enzyme of glycolysis, from the cytoskeleton of B16 melanoma cells. These effects of taxol were dose-dependent, and preceded the decrease in ATP levels and cell viability. Thus, taxol not only inhibits the essential dynamic processes of microtubule network, but also reduces glycolysis, through the novel mechanisms described here.

Insulin stimulates binding of phosphofructokinase to cytoskeleton and increases glucose 1,6-bisphosphate levels in NIH-3T3 fibroblasts, which is prevented by calmodulin antagonists

We report here a novel mechanism of insulin action in cultures of NIH-3T3 fibroblasts. Our experiments revealed that in these cells, insulin induced a rapid and transient increase in cytoskeleton-bound phosphofructokinase (EC 2.7.1.11), the rate-limiting enzyme in glycolysis, with a corresponding decrease in soluble (cytosolic) activity. Insulin also induced a slower increase in the levels of glucose 1,6-bisphosphate, the potent activator of cytosolic glycolysis. Both the rapid and the slower stimulatory actions of insulin were prevented by treatment with structurally different calmodulin antagonists, which strongly suggest that calmodulin is involved in these effects of insulin. The present and our previous experiments in muscle suggest that rapid, Ca2+-calmodulin-mediated increase in the binding of glycolytic enzymes to cytoskeleton, as well as the slower increase in glucose 1,6-bisphosphate, may be a general mechanism, in different cells, in signal transduction of insulin.

Hexokinase binding to mitochondria: a basis for proliferative energy metabolism

Current thought is that proliferating cells undergo a shift from oxidative to glycolytic metabolism, where the energy requirements of the rapidly dividing cell are provided by ATP from glycolysis. Drawing on the hexokinase-mitochondrial acceptor theory of insulin action, this article presents evidence suggesting that the increased binding of hexokinase to porin on mitochondria of cancer cells not only accelerates glycolysis by providing hexokinase with better access to ATP, but also stimulates the TCA cycle by providing the mitochondrion with ADP that acts as an acceptor for phosphoryl groups. Furthermore, this acceleration of the TCA cycle stimulates protein synthesis via two mechanisms: first, by increasing ATP production, and second, by provision of certain amino acids required for protein synthesis, since the amino acids glutamate, alanine, and aspartate are either reduction products or partially oxidized products of the intermediates of glycolysis and the TCA cycle. The utilization of oxygen in the course of the TCA cycle turnover is relatively diminished even though TCA cycle intermediates are being consumed. With partial oxidation of TCA cycle intermediates into amino acids, there is necessarily a reduction in formation of CO2 from pyruvate, seen as a relative diminution in utilization of oxygen in relation to carbon utilization. This has been assumed to be an inhibition of oxygen uptake and therefore a diminution of TCA cycle activity. Therefore a switch from oxidative metabolism to glycolytic metabolism has been assumed (the Crabtree effect). By stimulating both ATP production and protein synthesis for the rapidly dividing cell, the binding of hexokinase to mitochondrial porin lies at the core of proliferative energy metabolism. This article further reviews literature on the binding of the isozymes of hexokinase to porin, and on the evolution of insulin, proposing that intracellular insulin-like proteins directly bind hexokinase to mitochondrial porin.

Clotrimazole and bifonazole detach hexokinase from mitochondria of melanoma cells

Cancer cells are characterized by a high rate of glycolysis. Hexokinase (ATP: D-hexose 6-phosphotransferase, EC 2.7.1.1), the only glycolytic enzyme which binds to mitochondria, is exceptionally high in cancer cells, and believed to play a key role in regulating cell energy metabolism and cancer cell growth rate. We have previously found that clotrimazole (1-(alpha-2-chlorotrityl)imidazole) and bifonazole (1-(alpha-biphenyl-4-ylbenzyl)imidazole), the antifungal azole derivatives, which were recently recognized as calmodulin antagonists, are calmodulin antagonists which most effectively reduce glycolysis and ATP level in B16 melanoma cells. They act through allosteric regulation and detachment of glycolytic enzymes from cytoskeleton. Here we report of a novel, additional, mechanism of action of these drugs. We show that they induce a dose-dependent detachment of hexokinase from mitochondria of B16 melanoma cells. This effect preceded the decrease in cell viability. These results suggest that clotrimazole and bifonazole may be promising drugs in treatment of melanoma.

Hexokinase binding to mitochondria: a basis for proliferative energy metabolism

Current thought is that proliferating cells undergo a shift from oxidative to glycolytic metabolism, where the energy requirements of the rapidly dividing cell are provided by ATP from glycolysis. Drawing on the hexokinase-mitochondrial acceptor theory of insulin action, this article presents evidence suggesting that the increased binding of hexokinase to porin on mitochondria of cancer cells not only accelerates glycolysis by providing hexokinase with better access to ATP, but also stimulates the TCA cycle by providing the mitochondrion with ADP that acts as an acceptor for phosphoryl groups. Furthermore, this acceleration of the TCA cycle stimulates protein synthesis via two mechanisms: first, by increasing ATP production, and second, by provision of certain amino acids required for protein synthesis, since the amino acids glutamate, alanine, and aspartate are either reduction products or partially oxidized products of the intermediates of glycolysis and the TCA cycle. The utilization of oxygen in the course of the TCA cycle turnover is relatively diminished even though TCA cycle intermediates are being consumed. With partial oxidation of TCA cycle intermediates into amino acids, there is necessarily a reduction in formation of CO2 from pyruvate, seen as a relative diminution in utilization of oxygen in relation to carbon utilization. This has been assumed to be an inhibition of oxygen uptake and therefore a diminution of TCA cycle activity. Therefore a switch from oxidative metabolism to glycolytic metabolism has been assumed (the Crabtree effect). By stimulating both ATP production and protein synthesis for the rapidly dividing cell, the binding of hexokinase to mitochondrial porin lies at the core of proliferative energy metabolism. This article further reviews literature on the binding of the isozymes of hexokinase to porin, and on the evolution of insulin, proposing that intracellular insulin-like proteins directly bind hexokinase to mitochondrial porin.

Detachment of glycolytic enzymes from cytoskeleton of melanoma cells induced by calmodulin antagonists

Glycolysis, which is the primary energy source in cancer cells, is known to be controlled by allosteric regulators, as well as by reversible binding of glycolytic enzymes to cytoskeleton. We have previously found that different calmodulin antagonists decrease the levels of allosteric activators of glycolysis, and reduce ATP content and cell viability in B16 melanoma cells. Here we report of a novel, additional, mechanism of action of calmodulin antagonists in melanoma cells. We show that these drugs cause a detachment of the glycolytic enzymes, phosphofructokinase (ATP: D-fructose-6-phosphate 1-phosphotransferase, EC 2.7.1.11) and aldolase (D-fructose-1,6-bisphosphate D-glyceraldehyde-3-phosphate-lyase, EC 4.1.2.13), from cytoskeleton of B16 melanoma cells. This effect was dose- and time-dependent, and preceded the decrease in cell viability. The detachment of glycolytic enzymes from cytoskeleton would reduce the provision of local ATP, in the vicinity of the cytoskeleton-membrane and would affect cytoskeleton structure. Since the cytoskeleton is being recognized as an important modulator of cell function, proliferation, differentiation and neoplasia, detachment of the glycolytic enzymes from cytoskeleton induced by calmodulin antagonists, as well as their reported inhibitory action on cell proliferation, make these drugs most promising agents in treatment of cancer.

Serotonin decreases cytoskeletal and cytosolic glycolytic enzymes and the levels of ATP and glucose 1,6-bisphosphate in skin, which is prevented by the calmodulin antagonists thioridazine and clotrimazole

Serotonin (5-hydroxytryptamine) is believed to play a pathogenic role in skin damage and various skin abnormalities; however, its mechanism of action remains unknown. We show here that intradermal injection of serotonin in rats induced a marked reduction in the activities of the glycolytic enzymes, phosphofructokinase (EC 2.7.1.11) and aldolase (EC 4.1.2.13), in both the cytoskeletal and cytosolic fractions from skin. Serotonin also decreased the levels of glucose 1,6-bisphosphate in skin, the powerful regulator of glucose metabolism. These serotonin-induced changes were accompanied by a marked decrease in ATP content in skin. All these pathological changes induced by serotonin were prevented by treatment with two structurally different calmodulin antagonists: thioridazine, an antipsychotic phenothiazine, or clotrimazole, from the group of the antifungal azole derivatives that were recently recognized as calmodulin antagonists. The present results suggest that calmodulin antagonists may be effective drugs in the treatment of skin damage under various pathological conditions and diseases in which serotonin levels are increased.

Calmodulin antagonists decrease glucose 1,6-bisphosphate, fructose 1,6-bisphosphate, ATP and viability of melanoma cells

Glycolysis is known to be the primary energy source in cancer cells. We investigated here the effect of four different calmodulin antagonists: thioridazine (10-[2-(1-methyl-2-piperidyl) ethyl]-2-methylthiophenothiazine), CGS 9343B (1,3-dihydro-1-[1-[(4-methyl-4H,6H-pyrrolo[1,2-a] [4,1]-benzoxazepin-4-yl)methyl]-4-piperidinyl]-2 H-benzimidazol-2-one (1:1) maleate), clotrimazole (1-(alpha-2-chlorotrityl)imidazole) and bifonazole (1-(alpha-biphenyl-4-ylbenzyl)imidazole), on the levels of glucose 1,6-bisphosphate and fructose 1,6-bisphosphate, the two stimulatory signal molecules of glycolysis, and on ATP content and cell viability in B16 melanoma cells. We found that all four substances significantly reduced the levels of glucose 1,6-bisphosphate, fructose 1,6-bisphosphate and ATP, in a dose- and time-dependent manner. Cell viability was reduced in a close correlation with the fall in ATP. The decrease in glucose 1,6-bisphosphate and fructose 1,6-bisphosphate did not result from the cytotoxic effects of the calmodulin antagonists, since their content was already reduced before any cytotoxic effect was observed. These findings suggest that the fall in the levels of the two signal molecules of glycolysis, induced by the calmodulin antagonists, causes a reduction in glycolysis and ATP levels, which eventually leads to cell death. Since cell proliferation was also reported to be inhibited by calmodulin antagonists, these substances are most promising agents in treatment of cancer by inhibiting both cell proliferation and the glycolytic supply of ATP required for cell growth.

Specific growth rate as a parameter for tracing growth-limiting substances in animal cell cultures

The specific growth rate (mu) reaches its maximum very early during culture (at 20 h), but declines again thereafter so that no exponential growth phase occurs in batch cultures of hybridoma cells. This growth rate limitation depends neither on exhaustion of any of the macro-nutrients, nor on inhibition by metabolic byproducts (Ljunggren and Häggström, 1994). Intermittent additions of serum, after 20 and 40 h of culture, resulted in exponential growth throughout the growth phase. Insulin was found to be the main component responsible for the growth rate increasing effect. Intermittent additions of serum or insulin to a dual substrate (glucose and glutamine) limited fed batch culture increased the growth rate also here, and the results indicate the existence of a minimum growth rate (about 0.02 h-1) at a threshold glutamine level (0.005 mM). Serum and insulin additions markedly enhanced the glucose consumption and lactate formation rates, a metabolic effect that was not coupled to the increase in mu. The reduced concentrations of glucose and glutamine in substrate limited fed batch cultures suppressed substrate consumption rates and byproduct formation (lactate, ammonium, alanine, other amino acids) even in the serum and insulin stimulated cultures and rendered the energy metabolism much more efficient than in batch culture. Further, the serum and insulin stimulated cells growing in substrate limited fed batch culture produced almost 4-times more antibodies, from the same amount of nutrients as supplied to the batch grown cells.

The insulin receptor--single function and dual effect

Net effects of insulin on glucose entry, metabolism and other cellular processes have been well documented over the past 30-40 years. Although it is known that insulin binds to a specific cell membrane receptor protein which undergoes autophosphorylation and tyrosine kinase activation, the individual reactions following receptor activation that cause the metabolic changes remain unknown. It is well documented that the isolated insulin receptor has a high degree of basal autophosphorylation capacity and externally directed tyrosine kinase. There is also evidence that some in vivo autophosphorylation can take place in the total absence of insulin. If receptor activity does exist in the absence of insulin, then receptor function needs to be reanalyzed. It will be proposed here that the insulin binding membrane protein functions mainly to inhibit glucose transport under low physiological levels of insulin. Evidence of basal receptor enzymatic activity in the absence of insulin supports this theory. Under metabolically sufficient conditions, enough insulin receptors are functionally active to interact with the glucose transport system in an inhibitory manner, providing membrane control of internal glucose metabolism. Insulin acts by aggregating this inhibitory system. If inhibitory insulin receptors are aggregated following insulin elevation, their inhibitory action is prevented and glucose transport increases. This increase in transport will be in direct proportion to the temporal inhibitory level of the receptor and to the area of the cell membrane cleared of their inhibitory effect. When insulin receptor protein is confined to small areas of the cell membrane through aggregation, any potential inhibitory function is negated and glucose entry increases dramatically. This is the classical insulin effect. Both of these concepts were suggested 37 years earlier. Randle & Smith (1957, Biochem. Biophys. Acta 25, 442; 1958, Biochem. J. 70, 490) proposed that the internal supply of energy rich compounds limited glucose entry and that the effect of insulin was to inhibit this process which was inhibiting glucose entry. The present report provides a mechanism for this.

Insulin, growth factors, and cancer cell energy metabolism: an hypothesis on oncogene action

The hexokinase-mitochondrial acceptor theory provides a model of insulin action which unifies the metabolic effects of this hormone and suggests that these result from insulin's stimulatory effect on mitochondrial ATP synthesis. There are similarities between these changes in cells exposed to insulin and in mitochondria of transformed cell lines and cancer cells, where an increased binding of hexokinase to mitochondria is observed. This phenomenon may play a key role in the high rates of glycolysis sustained by cancer cells under aerobic conditions. Structural and functional evidence support the hypothesis that certain growth factors and oncogenes act through stimulation of oxidative phosphorylation via promoting HK binding to mitochondria.

Ammonia inhibits insulin stimulation of the Krebs cycle: further insight into mechanism of hepatic coma

Oxidation of [2,3-14C]succinate in the intramitochondrial Krebs cycle was used as a probe to investigate the effect of ammonia on protein incorporation and Krebs cycle oxidation of succinate carbons in isolated rat hepatocytes. At low concentrations of ammonium chloride (0.1 to 0.5 mM) a slight increase in 14CO2 formation from [2,3-14C]succinate was observed, however, the stimulatory effect of insulin was significantly reduced. Insulin failed to cause any stimulation of succinate carbons incorporation into hepatocyte protein in the presence of ammonium chloride. Addition of ammonium chloride also depressed the movement of tracer carbons into the gluconeogenesis pathway. The activity of the amphibolic amino acid pool was significantly enhanced by ammonia. The data presented in this paper lend strong support to the Krebs-cycle depletion theory of hepatic coma. They also suggest that reduced mitochondrial Krebs cycle activity caused by increased amphibolic depletion of substrates results in loss of insulin sensitivity in ammonia toxicity.

Extracellular phosphate requirement for insulin action on isolated rat hepatocytes

Isolated rat hepatocytes were prepared in KHB buffer, pH 7.4; were centrifuged and washed twice in KHB buffer containing various amounts of phosphate and calcium; and were incubated at 30 degrees in the presence of tracer [2,3-14C]succinate and a 0.5 mM concentration of each of the 20 natural amino acids. Hepatocytes washed and incubated in KHB buffer containing less than 0.1 mM phosphate failed to show any insulin stimulation of [2,3-14C]succinate oxidation or protein incorporation of tracer carbons. The absence or presence of extracellular phosphate did not alter the specific activity of 32P-adenine nucleotides; they remained the same in the presence or absence of insulin. The maximal insulin stimulatory effect on succinate oxidation and tracer incorporation into protein was observed in the presence of 1.18 mM phosphate and 1.9 mM calcium ion. The lack of external phosphate did not prevent the stimulation of succinate oxidation by either glucagon on epinephrine, whereas removal of calcium from the medium abolished their hormonal effects. The lack of medium calcium also prevented the insulin stimulation of succinate oxidation and protein synthesis. Our data indicate that a diminished insulin responsiveness in hypophosphatemic patients may be due to the insensitivity of mitochondria to insulin in the hypophosphatemic state.

Differential effects of insulin, epinephrine, and glucagon on rat hepatocyte mitochondrial activity

Oxidation of [2,3-14C]succinate carbons in the mitochondrial Krebs cycle was used as a probe to investigate the effects of insulin, epinephrine, glucagon, and 2,4-dinitrophenol (2,4-DNP) on isolated rat hepatocytes. Epinephrine, glucagon, and 2,4-DNP had a far greater stimulatory effect on 14CO2 formation from [2,3-14C]succinate than insulin. Unlike insulin, epinephrine and glucagon had no significant effect on the anabolic utilization of succinate carbons for protein synthesis. Our results suggest that although epinephrine, glucagon, and 2,4-DNP enhance the movement of tracer carbons through the Krebs cycle, only insulin is capable of enhancing amphibolite utilization for protein synthesis.

Red blood cell insulin receptors in health and disease

CONTENTS

Structure and characteristics of erythrocyte insulin receptor. Red blood cell age and insulin receptors. Insulin receptors in human disease states. Obesity. Chronic renal failure. Acanthosis nigricans. Miscellaneous disease states. Insulin receptors in children. Insulin receptors in women during pregnancy. Insulin binding and other hormones. Comparison of biosynthetic insulin, pancreatic human insulin and porcine insulin binding to erythrocytes. Effect of exercise on insulin binding to red blood cells of normal human volunteers. Miscellaneous insulin binding studies. Insulin internalization and degradation. Insulin and erythrocyte metabolism. Summary and conclusion.