Neuronal Energy Deprivation

There are three major conditions, or encephalopathies, that could lead to neuronal energy deprivation: Hypoglycemia, or glucose deficiency in the blood; anoxia (hypoxia), or O2 deficiency in the inspired air; and ischemia, or deficiency in blood supply, which, in essence, combines the first two deficiencies. All three encephalopathies are characterized by an inadequate supply of one of the two major energy substrates, or both. Since glucose and O2 are the two major energy substrates of the brain, it is easy to understand how their deficiencies will affect the production of cerebral ATP and all the energy-demanding processes. Nevertheless, there are many other encephalopathies of intrinsic origin, such as deficiencies in one enzyme or another of the energy metabolic pathways, and seizures. Others may arise from extrinsic factors, such as vitamin deficiencies, toxicity of heavy metals, and viral infections. This chapter covers only the three major encephalopathies, i.e., hypoglycemia, anoxia (hypoxia), and ischemia.

There are two groups of researchers who study degenerative diseases of the brain. Basic scientists, who attempt to better understand the workings of cerebral energy metabolism and its regulation, and clinical scientists, who delve into the mysteries of these disorders, in an attempt to alleviate their terrible toll. Both groups depend on each other's discoveries to bring them closer to their goals. The basic scientist's approach is the one taken here, keeping the clinical aspects of the major encephalopathies to a necessary minimum. For scientists to study any disorder and its consequences, an experimental model which either mimics or closely resembles this disorder, is a must. An in vivo model usually employs an experimental animal; an in vitro one employs any number of subcellular, cellular, tissue, or organ preparations.

Once established, a good experimental model affords a better understanding of the sequence of events that lead to the final outcome of the disorder under study. The model also opens an opportunity for experimentation with potential treatments or preventive measures that could minimize or eliminate the damaging outcome of the disorder. This is when basic scientists and clinicians closely collaborate, namely, in the discovery, development, and later, the clinical test of these treatments for their potential neuroprotective value.

3.1. Hypoglycemia and Ischemia/Hypoxia

Although the phenomena of hypoglycemia, ischemia, and hypoxia have been dealt with separately in many publications, the decision to combine the three under one heading here is based chiefly on information published over the past two decades, which indicated many overlaps and parallels between these three encephalopathies. The main reason they have been traditionally separated has to do more with the distinguished metabolic vehicles that lead to the observed outcome, rather than with the outcome itself.

Hypoglycemia may occur in diabetes through insulin overdose, liver disease, and in many enzyme deficiencies, such as glucose-6-phosphatase, fructose-1,6-biphosphatase, phosphoenol-pyruvate carboxykinase, pyruvate carboxylase, or glycogen synthetase. Each of these deficiencies could affect normal function of the brain, and lead, if not corrected, to neurological deficits.

The most common outcome of the interruption of glucose supply to the brain is hypoglycemia-induced coma, which ensues when glucose blood levels drop to 8-9 mg/100 mL, or about 10% of the normal isoglycemic glucose level of 90 mg/100 mL

(66). In contrast, arterial O2 content, mean blood pressure, cerebral respiratory quotient, and cerebral blood flow, all remain unchanged. However, cerebral O2 consumption declines in hypoglycemic coma (56% of normal), along with a dramatic decline in cerebral glucose consumption (66). These declines can be produced by an overdose of insulin, hepatic insufficiency, or by the inhibition of glucose metabolism with sufficiently high dose of 2DG (67-69). Nevertheless, one should be aware that symptomatic hypoglycemia may occur at blood glucose levels that are significantly different from those mentioned above. Thus, in the diabetic, a decrease in blood glucose level from 300 to 100 mg/100 mL, upon insulin administration, may cause symptomatic hypoglycemia (70). Yet, in a better-controlled diabetes, symptomatic hypoglycemia may not occur until the blood glucose levels reach 50 mg/100 mL. In normal, young humans after glucose tolerance test, blood glucose levels may fall to 30-40 mg/100 mL, without the appearance of any symptoms (70).

Intraperitoneal injection of insulin is the most common method employed by investigators for the induction of hypoglycemia. With the right dose of insulin (50-100 U/kg), hypoglycemic coma ensues within 1-2 h following insulin injection in rats or mice (71). In larger animals, such as the dog, higher doses of insulin (200-300 U/kg) are required to achieve coma. For metabolic studies, a small animal is the model of choice. For studies in which cerebral blood flow is involved, a larger animal model is required.

As mentioned earlier, hypoglycemic symptoms can be induced with a pharmaco-logic dose (400 mg/kg) of 2DG (67-69). When hypoglycemia is induced with 2DG, blood glucose levels actually increase, rather than decrease, which is an indication that symptomatic hypoglycemia is not the result of low levels of blood glucose, but rather the result of a reduction in glucose metabolism or inadequate supply of glucose to its transporters, because of their occupation by 2DG.

Thus, although both methods abolish glucose metabolism, the consequences of these deprivations are markedly different. Insulin enhances glucose metabolism to the point at which the monosaccharide levels are low enough to trigger the utilization of alternative energy substrates, such as glycogen and ketones. In contrast, 2DG competes with glucose for both the glucose transporter and the enzyme, hexokinase, which neither reduces the glucose levels nor triggers the utilization of alternative substrates. Apparently, the effect of 2DG would be more swift and severe than that of insulin injection. Nevertheless, many of the gross consequences of hypoglycemia, such as histologic, electroencephalographic, and behavioral changes induced by both methods, should be similar.

Histologically, cerebral cortex neurons of rats exposed to insulin-induced hypogly-cemia exhibit changes dependent on the duration of the hypoglycemic period beyond the onset of isoelectric EEG. A short, 30-min hypoglycemia produces mostly shrinkage and condensation of both cytoplasm and nuclei, but rats that are exposed to a longer, 60-min hypoglycemia exhibit significantly greater number of damaged cells (71). Moreover, hypoglycemia produces cell swelling and the appearance of empty vacuoles. Swollen neurons are more abundant after the longer hypoglycemic period than after the shorter one. Although cell damage observed in other encephlopathies is frequently irreversible, the histologic changes that are recorded in hypoglycemic rats can be reversed by glucose supplement. Within 4 h after the onset of glucose supplementation by iv injection, more than 95% of the neurons have normal appearance (71). The phenomenon of swollen neurons could be explained by the derailment of the energy-dependent ion homeostasis. Shrinkage of cells, as acutely as it occurs in hypoglycemia, remains an unclear phenomenon, especially its reversibility with glucose supplementation.

Slow-wave EEG occurs when glucose blood levels fall to 2 mM. When glucose levels fall to 1 mM or lower, EEG becomes isoelectric (71). Brain tissue in vitro responds in a similar way when glucose levels in the incubation medium are reduced. Evoked population spikes in the CA1 region of hippocampal slices cannot be sustained when glucose concentration falls below 1-2 mM (31). In vivo, cerebral metabolic rate of glucose does not decrease until plasma glucose levels fall below 2.5 mM (72). At these levels, the rate of glucose transport into the brain decreases to values 2-3x slower than normal, so that glucose concentration is insufficient to saturate hexokinase (71,73). Rats, like humans, exhibit hypoglycemia-related behavioral changes that range from normal to comatose, in close correlation with the blood glucose levels (71). Coma usually ensues when blood glucose levels fall to 1 mM or less. In some cases, however, hypoglycemia may induce seizures and seizure-like EEG, in which event, some of the changes, including the rate of decline in the blood glucose levels, and hence the onset of coma, may occur faster than expected.

Despite all these changes, reports on variations in cerebral blood flow range from an increase to no change during hypoglycemia-induced coma (71). Similarly, depending on the depth of hypoglycemia and accompanying events, such as seizures, O2 consumption can increase, decrease, or stay unchanged (71).

One of the typical events that accompany hypoglycemia is hypothermia. By all accounts, this phenomenon is directly dependent on the reduction in glucose availability, rather than any of the possible causes of hypoglycemia. The fact that 2DG-induced hypoglycemia is accompanied by hypothermia (67,69) is a strong indication that 2DG, although increasing blood glucose levels, also produces functional hypoglycemia (71). Thus, the postulate has been advanced that, when glucose becomes unavailable somehow, through specific temperature-regulating receptors, it induces the observed hypoglycemic hypothermia. The hypothermic effect of hypoglycemia can be reversed by high doses of fructose (5x the dose of 2DG), when administered intraventricularly, along with 2DG (71,74). High doses of glucose should have similar effect to that of fructose, and, based on several recent in vitro studies, lactate, and possibly pyruvate, should also be able to overcome 2DG-induced hypothermia.

Since glucose is the main energy substrate in the brain, it is obvious that any deprivation of this fuel can lead to a shutdown of the very pathways of its metabolism. Thus, both the glycolytic pathway and the mitochondrial TCA are first to be affected by hypoglycemia. The levels of pyruvate, lactate, and glycogen fall significantly, once glucose blood levels fall to 2.5 mM or lower (71). Similarly, the levels of the TCA intermediates also decrease in response to hypoglycemia (71). Of the different metabolic effects of hypoglycemia, none is more important than the decline in brain ATP levels and, to a lesser degree, phosphocreatine levels. Notwithstanding, of the different brain regions, the cerebellum does not exhibit such falls in ATP and phosphocreatine levels, even late into the isoelectric period (71). Although this unique ability of the cerebellum to withstand metabolic strain has been documented (71) using other models of metabolic stress, the mechanism governing this ability is unknown.

All the hypoglycemic symptoms, functional and behavioral, which are induced by an overdose of insulin, can be completely reversed by an administration of glucose.

Hypoxia is defined as low O2 content or tension or O2 deficiency in the inspired air. The extreme case of hypoxia is anoxia, the complete absence of O2. Ischemia is defined as deficiency in blood supply caused by functional constriction or actual obstruction of a blood vessel (75). Clearly, an organ that suffers from a deficiency in blood supply is also afflicted with hypoxia. Hence, many of the symptoms associated with hypoxia and ischemia are identical. Furthermore, the cellular processes and mechanisms that lead to posthypoxic and postischemic damage are identical. Thus, the term "cerebral ischemia" is being used throughout this subheading, with the understanding that it also refers to cerebral hypoxia.

There are probably more books, review articles, and research papers written about cerebral ischemia than any other encephalopathy. Focal cerebral ischemia (stroke) is the third leading cause of death in the United States and other industrialized countries. Stroke survivors usually suffer permanent and debilitating brain damage. The huge social and economic cost of this encephalopathy, on one hand, and the potential benefits from a future protective modality, on the other, continue to stimulate the research in this field.

The complete dependency of the CNS on glucose and O2 for the generation of sufficient ATP supplies makes the brain the most vulnerable of all tissues to ischemia. Pure hypoxia is rarely encountered in humans. Pure hypoxia is experienced only in cases of pulmonary insufficiency, such as in pulmonary emphysema, in anemic patients, or in normal individuals who are acutely exposed to altitudes above 2400 m. Obviously, cerebral ischemia of identical duration should be more devastating than hypoxia, or even anoxia, because the ischemic brain suffers not only from oxygen deficiency, but also from shortage of glucose supplies. It is the compounded effect that makes cerebral ischemia so devastating. As is demonstrated below, and has been the case with hypoglycemia, both the brain in vivo, and any of its in vitro preparations, can tolerate either hypoxia or hypoglycemia alone longer than when the two are combined. In other words, the presence of one deficiency sensitizes the brain to the effect of the other. Thus, ischemia of a given duration is always more damaging than either hypoxia or hypoglycemia of a similar duration.

The most common events causing cerebral ischemia are stroke, cardiac arrest, and head injury or trauma. Stroke or focal ischemia occurs when blood supply is interrupted to a specific region of the brain; global ischemia occurs when blood supply to the entire brain is interrupted, as in cardiac arrest. Furthermore, both focal and global ischemia, depending on their severity and duration, may be complete (a total absence of blood flow) or incomplete (a drastic reduction in blood flow). The severity and the duration of these insults determine whether a reversible or irreversible cell injury occurs. An irreversible injury eventually leads to cell death (infarction).

Since there are brain regions that are more sensitive to ischemic injury than others, they are referred to as "selectively vulnerable" regions of neurons. In many of these regions, the neuronal death does not occur during or immediately after the ischemic insult. Frequently, neuronal death is delayed for 3-7 d. In focal ischemia, the vascula-ture surrounding the ischemic umbra, known as the penumbra, may still provide some collateral, although compromised, circulation, which, if treated early enough, could rescue the penumbric neurons. Of the sensitive brain regions, the most vulnerable neurons are located in the CA1 sector of the hippocampus. Neurons in certain portions of the caudate nucleus, and in layers 3, 5, and 6 of the neocortex, are also selectively vulnerable. Understanding the mechanism that leads to this vulnerability is a major goal of many research laboratories around the world.

Several recently published monographs on cerebral ischemia (76-78) describe both in vivo and in vitro models of this encephalopathy. Many of the animal models of yesteryear, such as monkey, dog, cat, rabbit, and pig, are only rarely used today, mostly because of their very high cost. Most of the animal models of the past two decades make use of rodents. These models proved to be both cost-effective and reliable. Moreover, many of the recent advances in this field of research appear to be directly applicable to humans.

The focus of this chapter is on the effect of energy deprivation on neurons, but much can be learned from the enormous number of studies conducted over the last three decades on the cellular mechanisms leading to ischemic neuronal death. The putative roles of lactic acid, Glu, Ca2+ influx, and many other factors, in causing the delayed demise of ischemic neurons, have been discussed exhaustively elsewhere and are not the focus of this chapter. However, the hypotheses that were erected to explain this neuronal demise led to many elegant studies that reveal important details about cerebral energy metabolism and the energy-dependent cellular processes that could go awry when ATP is diminished. Of these processes, only those that are directly affected by failure of energy metabolism upon ischemia, hypoxia, and hypoglycemia are of interest here.

Of the many developments characterizing the past two decades, with which cerebral ischemic research is concerned, the employment of in vitro systems, such as the hip-pocampal slice preparation and cell cultures, has had the greatest impact. These systems can be manipulated, controlled, and analyzed in ways that in vivo systems cannot tolerate. Because most recent knowledge on the effects of energy deprivation on neuronal tissue has emerged from in vitro studies, the last portion of this chapter deals mainly with their findings.

A great advantage of the hippocampal slice preparation, compared to other in vitro systems, such as cell cultures, is the ability of the investigator to use electrophysiologic means to record neuronal function indistinguishable from in vivo records. Moreover, the recording can be made from the most vulnerable sector to ischemia/hypoxia in the hippocampus, the CA1 region. Within 2-3 min of changing the atmosphere that slices are exposed to, from 95% O2/5% CO2 to 95% N2/5% CO2, the evoked population spike (neuronal function) amplitude falls to 0 mV. The ability of hippocampal slices to recover neuronal function, upon return to normal reoxygenation, depends on the duration of the hypoxic period, and on the concentration of glucose in the artificial cere-brospinal fluid. The higher the glucose concentration, the longer the hypoxic period from which neuronal function can recover (Fig. 5).

Reduction in glucose concentration does not visibly affect neuronal function in hippocampal slices, until it falls below 1.5 mM. However, when glucose deprivation is combined with hypoxia (in vitro ischemia), a reduction in glucose concentration, from 10 to 5 mM, significantly decreases the ability of hippocampal neuronal function to recover posthypoxia (Fig. 5). In contrast, increasing glucose concentration, from 10 to

Fig. 5. Rates of recovery of neuronal function (orthodromically evoked CA1 population spike) in rat hippocampal slices after 5, 12, or 30 min of hypoxia and 30 min reoxygenation in the presence of 5, 10, or 20 mM glucose, respectively. The higher was the glucose concentration in the perfusion medium, the longer was the duration of hypoxia slices could tolerate.

Fig. 5. Rates of recovery of neuronal function (orthodromically evoked CA1 population spike) in rat hippocampal slices after 5, 12, or 30 min of hypoxia and 30 min reoxygenation in the presence of 5, 10, or 20 mM glucose, respectively. The higher was the glucose concentration in the perfusion medium, the longer was the duration of hypoxia slices could tolerate.

20 mM, significantly prolongs the hypoxic period from which hippocampal neuronal function could recover (Fig. 5). Thus, the higher the concentration of glucose during O2 deprivation in vitro, the better the recovery of neuronal function posthypoxia. This relationship is in disagreement with in vivo findings, in which hyperglycemia has been shown to aggravate neuronal ischemic damage, a phenomenon known as the "glucose paradox" (79,80). These findings led to the formulation of the "lactic acidosis" hypothesis of cerebral ischemic damage (50,81). The glucose paradox, and other apparent contradictions between results obtained in vitro and in vivo, mostly in the early and mid-1980s, impeded the acceptance of the hippocampal slice preparation as a valuable tool in the study of brain energy metabolism and brain encephalopathies.

A hallmark of ischemic/hypoxic conditions in all tissues, including brain, is the increase in lactate production above the baseline levels under normoxia. When hippocampal slices are exposed to O2 deprivation, lactate production quickly increased 5-6-fold (44). As mentioned earlier, lactate can support normal neuronal function in hippocampal slices as the sole energy substrate. Moreover, lactate also has been shown to be a preferred energy substrate over glucose in excised sympathetic chick ganglia (40-42) and in rat cerebellar slices (34). Lactate is preferable to glucose when recovery from hypoxia is concerned (44). Moreover, lactate has been shown to be a mandatory aerobic energy substrate for recovery of neuronal function posthypoxia in hippocampal slices (44,56,82).

To demonstrate this role of lactate in the recovery of neuronal function posthypoxia, two experimental protocols were used. The first manipulated the ability of the glucose analog, 2DG, to block lactate production during hypoxia via inhibition of glycolysis (Fig. 6). Normally, 100% of slices perfused with 20 mM glucose recover their neuronal

Fig. 6. A schematic illustration of the experimental paradigms used in establishing the role of anaerobically produced lactate in the posthypoxic recovery of neuronal function in rat hip-pocampal slices. Shown are five different experimental paradigms (A-E), the various compositions of artificial cerebrospinal fluid that perfused the slices during each paradigm, and the different gas mixtures (bubbles) that were passed over the slices during these paradigms. In each paradigm, slices were first equilibrated with a medium containing 20 mM glucose and an oxygenated atmosphere, for at least 30 min (pretreatment). After a treatment period, slices were allowed to recover for 30 min under pretreatment conditions, before measuring recovery of neuronal function (bars on the right side of each paradigm's panel). In paradigm A, slices were exposed to 13-min hypoxia (gray shading), as 20 mM glucose-containing medium was replaced with 20 mM 2DG-containing medium. In paradigms B-D, the hypoxic period started 5, 8, or 10 min prior to the onset of 2DG perfusion, and continued an additional 13 min, for a total of 18, 21, or 23 min of hypoxia, respectively. The longer the hypoxic period prior to the blockade of glycolysis with 2DG, the more lactate produced and the higher the recovery rate of neuronal function posthypoxia. Thus, in what appears as a paradox, slices that were exposed to a total of 23 min hypoxia exhibited a significantly higher rate of neuronal function recovery than slices exposed to only 13 min hypoxia. Paradigm E is similar to paradigm D, except that, for the first 13 min of hypoxia, slices were perfused with 2DG, and, for the last 10 min, slices were perfused with glucose-containing medium. Slices treated according to paradigm E were unable to produce energy and lactate glycolitically during the first 13 min of hypoxia and, hence, the lack of recovery of neuronal function. Histograms on the right are mean values +SD. *Significantly different from paradigm A (p < 0.0005).

Fig. 7. The effect of 0.5 mM 4-CIN, a monocarboxylate transporter inhibitor, on the evoked population spike (neuronal function) amplitude in two slices, over time. One slice (open circles) was perfused with artificial cerebrospinal fluid containing 10 mM glucose, while the other was perfused with the same fluid containing 20 mM lactate. When 4-CIN was added to the perfusion medium of both slices, only the one that was perfused with the medium containing lactate exhibited diminution of neuronal function, because the inhibitor blocked the entry of lactate into neurons. The inhibitor had no effect on the entry of glucose into neurons.

Fig. 7. The effect of 0.5 mM 4-CIN, a monocarboxylate transporter inhibitor, on the evoked population spike (neuronal function) amplitude in two slices, over time. One slice (open circles) was perfused with artificial cerebrospinal fluid containing 10 mM glucose, while the other was perfused with the same fluid containing 20 mM lactate. When 4-CIN was added to the perfusion medium of both slices, only the one that was perfused with the medium containing lactate exhibited diminution of neuronal function, because the inhibitor blocked the entry of lactate into neurons. The inhibitor had no effect on the entry of glucose into neurons.

function after 30-min reoxygenation from 23-min hypoxia. Control slices continued to be perfused with 20 mM glucose during the first 10 min of hypoxia, at which time the perfusion was switched to 20 mM 2DG for the last 13 min of hypoxia: 80% of these slices exhibited recovery of synaptic function posthypoxia (Fig. 6, paradigm D). Experimental slices were also exposed to 23-min hypoxia; however, 20 mM 2DG was perfused during the first 13 min of hypoxia, followed by 20 mM glucose for the last 10 min of hypoxia. None of the experimental slices recovered neuronal function posthypoxia (Fig. 6, paradigm E). Although control slices were able to produce lactate during the first 10 min of hypoxia, experimental slices were unable to do so, since glycolysis was inhibited from the onset of hypoxia, thus preventing lactate production. Results shown in Fig. 6 clearly demonstrate the importance of lactate in the posthypoxic recovery of neuronal function. The more time slices were allowed to produce lactate anaerobically before the inhibition of glycolysis with 2DG, the higher the percentage of slices that recovered neuronal function posthypoxia. Hence, more slices that were exposed to a total of 23 min hypoxia recovered their neuronal function than did slices that were exposed to 13-, 18-, or 21-min hypoxia (Fig. 6, paradigms A-D). The most plausible explanation for this phenomenon is the higher levels of lactate found in tissue slices exposed to 23-min hypoxia than the levels in slices exposed to shorter hypoxic periods.

The second experimental protocol employed the specific neuronal lactate transporter inhibitor, 4-CIN (56,83,84). As shown in Fig. 7, this compound is able to block lactate-

Fig. 8. The effect of 0.5 mM 4-CIN on rat hippocampal slices' ability to recover their neuronal function posthypoxia (histogram in upper panel), and on their levels of lactate and glucose. A significantly lower percentage of slices recovered neuronal function posthypoxia in the presence of 4-CIN. This poor recovery rate probably resulted from inhibition of neuronal lactate utilization by the monocarboxylate transporter inhibitor. The significantly higher levels of lactate that were present during reoxygenation in 4-CIN-treated slices indicate that neuronal lactate utilization was inhibited. Bars represents means ± SD. *Signifi-cantly different from control slices, p < 0.0005; **p < 0.05.

Fig. 8. The effect of 0.5 mM 4-CIN on rat hippocampal slices' ability to recover their neuronal function posthypoxia (histogram in upper panel), and on their levels of lactate and glucose. A significantly lower percentage of slices recovered neuronal function posthypoxia in the presence of 4-CIN. This poor recovery rate probably resulted from inhibition of neuronal lactate utilization by the monocarboxylate transporter inhibitor. The significantly higher levels of lactate that were present during reoxygenation in 4-CIN-treated slices indicate that neuronal lactate utilization was inhibited. Bars represents means ± SD. *Signifi-cantly different from control slices, p < 0.0005; **p < 0.05.

supported neuronal function in rat hippocampal slices (see also ref. 56). When perfused with 10 mM glucose, 78% of control slices recovered neuronal function after 10-min hypoxia; only 15% of experimental slices recovered neuronal function after being exposed to 10-min hypoxia in the presence of 0.5 mM 4-CIN (Fig. 8). Both groups of slices produced high levels of lactate during O2 deprivation, although experimental slices exhibited a significantly slower decline in those levels during reoxy-genation. This slower decline could result from blockade of neuronal lactate utilization by 4-CIN (Fig. 8).

The results of the above experiments also explain the ability of elevated glucose concentration, if supplemented before hypoxia, to afford neuroprotection against hypoxic damage in hippocampal slices (Figs. 5 and 6). This neuroprotection is probably the result of two separate processes: First, the increase in glucose availability allows the tissue to maintain glycolytic flux for a longer period of time, for the support of ion homeostasis; second, the longer glycolytic flux produces a significant increase during hypoxia in the lactate level, which in turn is available for aerobic utilization during the initial stages of reoxygenation, when both ATP and glucose levels are very low. The ability of the specific neuronal monocarboxylate transporter inhibitor, 4-CIN, to block posthypoxia recovery of neuronal function, provides further evidence that aerobic lactate utilization is crucial for such recovery.

Although these in vitro results befit the general understanding of both the aerobic and anaerobic energy metabolism pathways and their consequences (including the "more glucose, better tolerance to hypoxia" concept), most investigators in the field of cerebral ischemia find them hard to accept fully. The reason for this skepticism has much to do with the heavily heralded in vivo phenomenon, in which preischemic hyperglycemia significantly exacerbates delayed neuronal damage, as measured 4-7 d postischemia. This glucose paradox phenomenon of cerebral ischemia was first reported almost 25 yr (79), and has been reproduced numerous times in different in vivo models of ischemia. It became the cornerstone of the lactic acidosis hypothesis, which attributes the bulk of the delayed neuronal damage observed 4-7 d postischemia to the increase in lactic acid levels in the brain and the resultant acidosis (50).

The glucose paradox of cerebral ischemia also appears to affirm the main premise of the lactic acidosis hypothesis: more glucose = more lactic acid = more delayed neuronal damage. But does it? If acidosis (pH 6.8-6.5) is detrimental to neurons exposed to O2 deprivation or to O2-glucose deprivation, the same trend should be seen in vitro. However, experiments, both in brain slices and in neuronal cultures, showed that aci-dosis not only did not exacerbate hypoxic or ischemic neuronal damage in vitro, but in essence was actually neuroprotective (85-87). Again, this contradiction between the in vitro and in vivo outcomes was used to criticize the in vitro approach to the study of cerebral ischemia. Nevertheless, the agreement of in vitro results on cerebral ischemia with in vivo findings, over the past two decades, significantly surpasses the disagreement.

If one hypothesizes lactate to be a major energy substrate upon reoxygenation after hypoxia, rather than to be a detrimental factor, perhaps one could uncover supportive in vivo data. Of the vast number of in vivo studies over the past three decades, several included measurements of tissue levels of energy substrates and metabolites, such as glucose, lactate, ATP, and adenylate energy charge, before, during, and after an episode of cerebral ischemia. Such data from three key studies (88-90) are depicted in Fig. 9. The trend illustrated by these three studies is typical: Normal preischemic brain levels of glucose (1.7-3.0 ^mol/g) and lactate (1.0-1.8 ^mol/g) were drastically changed following 5-10 min of anoxia or ischemia. Although glucose levels fell (to 00.5 ^mol/g), those of lactate rose sharply (to 12-20 ^mol/g). After only 15 min of reperfusion/reoxygenation, glucose in the brain climbed to levels significantly higher than those existing preischemia (88-90). Even after 90 min of reperfusion, brain glucose level remained over 200% of preischemic level (88). Despite the dramatic increase

Brain Reperfusion Images
Fig. 9. Brain tissue levels (^mol/g) of glucose (white bars) and lactate (gray bars) at normoxia (N), at the end of ischemia/anoxia (I), and after reperfusion/reoxygenation (R), as reported in three separate and independent studies (88-90).

in lactate concentration after 10 min of anoxia, the rate of its efflux from the brain to the venous blood remained unchanged from the very low control levels (approx 0.2 ^mol/g/min) before anoxia (89).

These findings indicate that the brain tightly conserves the lactate it produces, and agree with the results described above for hippocampal slices (44,56,82). The post-ischemic/postanoxic increase in brain glucose was accompanied by a rapid decrease in lactate levels. The investigators themselves termed this decrease "lactate utilization" (89). Energy charge (EC, [ATP + 1/2ADP/{ATP + ADP + AMP}]) was also measured before anoxia (0.89), after 10 min of anoxia (0.39), and after 15 min of reperfusion/reoxygenation (0.92) (89). Similar results were reported in the other two studies (88,90).

A plausible explanation for a postischemic/postanoxic decline in brain lactate levels, and a concomitant increase in glucose levels, could be as follows. Upon reoxygenation, lactate is utilized, aerobically, as an energy substrate, but glucose remains mostly unused, resulting in its accumulation above control levels. This aerobic lactate utilization is sufficient to rapidly restore preanoxic ATP levels and energy charge values.

Although it would be prohibitive to review the vast body of data here, the most revealing information can be extracted from a study (81) that was aimed at investigat ing the role of lactic acidosis in ischemic brain damage under hyperglycemic conditions. But it is the outcome during hypoglycemic conditions that deserves closer attention. Rats fasted for 16-24 h, and exposed to 30 min of severe incomplete ischemia, showed more than a 17-fold increase in brain lactate (from 0.88 to 15.5 ^mol/g). Both glucose and glycogen levels fell essentially to zero (from 2.78 and 2.84 ^mol/g, respectively) during the same period; ATP levels fell by 95% from the control level. Recirculation for 90 min brought about a significant decrease in brain lactate levels, from 15.5 to 3.19 ^mol/g (a decrease of 79%). Glucose levels rose concomitantly to 106% of control, although the rats were not supplemented with glucose during recirculation. ATP rose to 81% of its control level, and 54% of the glycogen pools was restored. Even more striking, however, were the results when another group of fasting rats was infused with glucose (2 mL 50% glucose solution) during the first 10 min of recirculation. After 30 min of recirculation, the lactate level dropped 58% (from 15.5 ^mol/g at the end of 30 min of ischemia, to 6.44 ^mol/g), declining by 83% after 90 min recirculation. In contrast, glucose levels, which rose only to 3.58 ^mol/g in the control, nonischemic animals, rose dramatically in the ischemic rats, to 15.7 ^mol/g (439% increase, compared to control) at 30 min recirculation. Even after 90 min recirculation, glucose levels were still 380% of the control level. Again, this 20-yr-old study indicates clearly that, during the first hour or so of postischemia, the brain metabolizes lactate aerobically, while glucose remains unused.

In a recent preliminary study, using a rat model of cardiac-arrest-induced transient global cerebral ischemia (TGI), three important results were obtained (91). First, the glucose paradox, the phenomenon that was heralded as the proof for the detrimental role of lactic acidosis in delayed neuronal damage, can be brought about by hyperglycemia (2 g/kg glucose loading), when induced up to 1 h pre-TGI. Rats that were treated in this way exhibited a significant increase in delayed neuronal damage post-TGI. Second, loading glucose 2-4 h pre-TGI, while rendering the rats hyperglycemic at the time of the ischemic insult, did not exacerbate the delayed neuronal damage post-TGI. Rather, these rats showed significantly less neuronal damage than that measured in control, isoglycemic rats. Third, brain lactate levels were equally high in both hyperglycemic groups, thus refuting the argument that increased lactate level during an ischemic insult is the reason for exacerbation of neuronal damage.

Delicious Diabetic Recipes

Delicious Diabetic Recipes

This brilliant guide will teach you how to cook all those delicious recipes for people who have diabetes.

Get My Free Ebook

Post a comment