| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Department of Medicine, King’s College School of Medicine and Dentistry (M.L.E., K.M., J.L., A.P., I.C.P.C., S.A.A.), London SE59PJ, United Kingdom; and the Department of Physiology and Pharmacology, Queen’s Medical Center, University of Nottingham (I.M.), Nottingham NG7 2UH, United Kingdom
Address all correspondence and requests for reprints to: Dr. M. L. Evans, Department of Medicine, King’s College School of Medicine and Dentistry, London SE59PJ, United Kingdom.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
During hypoglycemia, ILH raised circulating levels of nonesterified fatty acids, glycerol, and ß-hydroxybutyrate, although the latter did not rise until after the onset of counterregulation. With ILH, epinephrine responses [area under the curve (AUC), 127.9 ± 31.7 vs. 175.1 ± 27.4 nmol/L·180 min; P = 0.03] and GH responses (AUC, 260 ± 91 vs. 1009 ± 150, P < 0.01) were reduced and delayed (glucose thresholds, 2.8 ± 0.04 vs. 3.0 ± 0.1 mmol/L; P = 0.04), with a trend toward reduced cortisol responses. Similarly, hypoglycemic symptom scores were diminished during ILH (AUC, 647 ± 162 vs. 1222 ± 874; P = 0.03). However, there was no significant effect on the deterioration in four-choice reaction time, one measure of cognitive deterioration [glucose thresholds, 2.6 ± 0.1 vs. 2.7 ± 0.1 mmol/L, ILH vs. control (P = 0.75); AUC, 1420 ± 710 vs. 2250 ± 1080 ms/min (P = 0.59)]. During euglycemic clamping with Intralipid/heparin infusion studies, there was no rise in hormones, four-choice reaction time, or symptoms other than hunger and tiredness.
Both nonesterified fatty acids and glycerol can penetrate the mammalian brain and be metabolized. Raised levels were able to reduce neurohumoral responses to hypoglycemia, but could not protect cognitive function. This suggests that regional differences exist in human brain metabolism between glucose-sensing and cognitive areas of brain, which may be important in the understanding of the mechanisms of glucose sensing and in the genesis of hypoglycemia unawareness in insulin-dependent diabetes.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The mammalian brain is known to contain transport mechanisms and enzyme systems that allow for the metabolism of a range of other potential nonglucose substrates (11, 12, 13, 14, 15, 16). In particular, specific blood-brain barrier transport mechanisms exist for nonesterified fatty acids (NEFA) (11, 17), and the mammalian brain also contains the necessary enzymes for metabolizing any transported fatty acids (13, 14, 15). We have therefore investigated whether the human brain can use lipid substrates to support cerebral metabolism and function by raising circulating levels of potential nonglucose lipid substrates in normal volunteers during stepped hypoglycemia while measuring the effects on counterregulatory hormone responses, symptoms, and cognitive function.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Nine subjects underwent stepped hyperinsulinemic hypoglycemic studies on two occasions in random order. On one occasion only Intralipid (20%) and heparin were infused, as described below (ILH), with the other hypoglycemic clamp acting as a control study (CON). Five subjects also underwent a hyperinsulinemic euglycemic study (EU), with infusion of Intralipid (20%) and heparin. All studies were performed at least 2 weeks apart. Subjects were kept blinded as to the exact study conditions at all times.
Subjects fasted for 10 h before each study. Two iv cannulas were placed in the nondominant arm using intradermal lidocaine as a skin anesthetic. One cannula was put in the antecubital vein for infusion of insulin, glucose, Intralipid, and heparin. The other was placed retrogradely in a distal wrist or hand vein, and that hand rested inside a warmed hot box (55–60 C) to arterialize venous blood (18). This cannula was used for sampling arterialized blood for measurement of plasma glucose, metabolites, and counterregulatory hormones.
Cognitive function was assessed by four-choice reaction time (19) using a lap-top computer in which up to 500 targets are presented over a 5-min period in 1 of 4 quadrants on a lap-top screen, and subjects press the corresponding key on a specially adapted keyboard. The average time and accuracy of response are recorded. The relative brevity of the test allows for repeated measures through the course of the studies. Subjects were trained before the study commenced, performing at least 4 measurements until stable results were obtained. The technique is validated as a simple and reproducible cognitive test that is sensitive to hypoglycemia (20) and has been shown to be supported by lactate and ß-hydroxybutyrate in the presence of hypoglycemia (5, 7).
Not less than 30 min after cannulas insertion, a primed continuous infusion of 1.5 mU/kg·min regular insulin (human Actrapid, Novo Nordisk Pharmaceuticals, Crawley, UK) was started. Arterialized plasma glucose was measured in duplicate at 5-min intervals by the bedside using a glucose oxidase technique (Yellow Springs glucose analyzer, Yellow Springs Instruments, Yellow Springs, OH). Plasma glucose was controlled by adjusting a simultaneous variable infusion of 20% dextrose (Clintec Nutrition, Slough, UK). Plasma glucose levels were maintained at 5 mmol/L for 40 min before being sequentially reduced to 3.8, 3.4, 2.8, and 2.4 mmol/L. Each step consisted of 40 min (except for the nadir at 2.4 mmol/L, which was for 20 min only), with the target glucose being reached during the first 15–20 min and then maintained for the duration of the step. After the nadir, the plasma glucose level was restored to 5 mmol/L and maintained at euglycemia during the final 40 min.
During one of each pair of hypoglycemic studies and during the euglycemic studies, infusions of 20% Intralipid (90 mL/h; Pharmacia, Milton Keynes, UK) and sodium heparin (0.1 U/kg·min) were given with the insulin infusion to raise levels of nonglucose lipid substrates. Intralipid (20%) is an oil in water emulsion containing 100 g fractionated soybean oil, 6 g fractionated egg phospholipid, and 11 g glycerol in 500 mL.
Additional arterialized blood samples were collected for measurement of catecholamines, cortisol, GH, glucagon, insulin, C peptide, triglycerides, NEFA, glycerol, BOB, lactate, urea, and electrolytes. Every 20 min, blood pressure and pulse were recorded, and subjects were asked to complete a standard symptom questionnaire. Subjects were asked to grade a standard list of 15 symptoms from 0 (absent) to 6 (very severe). The symptoms asked about were sweating, warmth, palpitations, tingling, anxiety, trembling, hunger, blurred vision, tiredness/drowsiness, confusion, weakness, headache, difficulty in speaking, dizziness, and irritability. Autonomic symptom scores were constructed from the first seven symptoms, and neuroglycopenic scores were developed from the last eight symptoms. On each occasion, after completion of the questionnaire, four-choice reaction time was measured.
The protocol for the euglycemic studies was identical, except that the infusion of Intralipid and heparin was started 40 min after the start of the insulin infusion and continued for a further 220 min, with arterialized plasma glucose maintained at 5 mmol/L throughout.
Plasma catecholamines (epinephrine and norepinephrine) were measured using high performance liquid chromatography (21). Cortisol (22), GH (23), free insulin (24), and C peptide were measured using RIAs. Glycerol was measured using an enzymatic fluorometric method (25), and NEFA (Wako kit from Alpha Laboratories, Eastleigh, UK) and BOB (Randox, Crumlin, Ireland) were measured using enzymatic methods. Paired samples were run in single assays. The inter- and intraassay variabilities were less than 10% for all assays.
Data analysis
All results are presented as the mean ± SE unless otherwise indicated. For analyzing parameters that were repeatedly measured over time (hormones, symptomatic responses, and four-choice reaction time), the incremental area under the curve (AUC) during hypoglycemia (from the end of the 40-min run-in period to the end of the hypoglycemic nadir) was calculated as a summary measure (26).
The onset of hypoglycemic responses was described by defining thresholds. For counterregulatory hormone responses, this was the plasma glucose concentration at which hormone levels became statistically greater than baseline values. Subjects did not experience symptoms during the baseline, so we could not use the same technique for symptom responses. Instead, we used the widely used method of defining the threshold as the time at which symptom scores rose by 2 points or more (27). Similarly for four-choice reaction time, we used a validated method of defining thresholds as the time at which response time consistently exceeded, or accuracy rates fell below stable baseline measurements ± twice the coefficient of variation for the test (28). In addition to this well validated technique for describing thresholds, we also analyzed the AUC for the incremental increase in the four-choice reaction time.
Where subjects failed to show a significant change as defined, the lowest glucose level achieved during the study was used for statistical comparisons. Results were analyzed using SPSS for Windows 6.1 (SPSS, Woking, UK). Repeated measures over time were summarized by using AUC, as described above, and by ANOVA corrected for repeated measures. To compare paired hypoglycemic studies, paired Student’s t testing was performed for normally distributed data, or Wilcoxon matched pairs signed ranks test was performed for nonnormally distributed data.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In response to the insulin infusion, free plasma insulin rose
rapidly to a steady state level of 83.6 ± 7.1 mU/L. Arterialized
plasma glucose was maintained at 5.01 ± 0.03 mmol/L. NEFA,
glycerol, triglycerides, and BOB all fell during the initial 40 min of
insulin and glucose infusion, but then rose, reaching peak values of
1.64 ± 0.23 mmol/L, 759.3 ± 60.7 µmol/L, 3.94 ±
0.24 mmol/L, and 326 ± 94 µmol/L at 140, 220, 220, and 220 min,
respectively. Lactate levels rose continuously throughout the study to
a peak of 1.51 ± 0.05 mmol/L (Table 1
).
|
).
|
Hypoglycemia
Glucose, insulin, C peptide, and substrates. Free insulin
levels were elevated similarly in both ILH and CON studies (82 ±
5 and 76 ± 6 mU/L), and C peptide levels were equally suppressed
(nadir, 0.12 ± 0.07 and 0.07 ± 0.01 nmol/L). Arterialized
plasma glucose profiles did not differ between the two studies (Fig. 1; nadir, 2.59 ± 0.07
vs. 2.55 ± 0.07 mmol/L, ILH vs. CON;
P = 0.58).
|
). In contrast, ILH infusion resulted in
elevated NEFA (peak, 2.24 ± 0.37 vs. 0.17 ± 0.22
mmol/L; P = 0.02), glycerol (peak, 1820 ± 315
vs. 22 ± 5 µmol/L; P < 0.001), and
BOB (peak, 257 ± 64 vs. 11 ± 4 µmol/L;
P = 0.01). During ILH, BOB first became significantly
elevated above control study levels only at the hypoglycemic nadir
(244 ± 53 vs. 18 ± 5 µmol/L; P
= 0.005). This elevation in BOB, therefore, only became significant
after counterregulatory neurohumoral responses were initiated (see
results below). Triglycerides also rose with ILH (0.97 ± 0.16
vs. 3.91 ± 0.52 mmol/L, baseline vs. peak;
P < 0.001), but cholesterol levels remained stable
throughout. Lactate levels rose by a similar degree during hypoglycemia
in ILH and CON studies (peak, 1.40 ± 0.11 vs.
1.47 ± 0.10 mmol/L; P = 0.49).
|
). However, the response was
significantly less during ILH (Table 3; AUC, 127.9 ± 31.7
vs. 175.1 ± 27.4 nmol/L·180 min; P =
0.03). GH responses were diminished and delayed by ILH [Table 3 and Fig. 3; AUC, 260 ± 91
vs. 1009 ± 150 (P < 0.01); glucose
thresholds, 2.8 ± 0.04 vs. 3.0 ± 0.1 mmol/L
(P = 0.04)]. A trend for reduced cortisol responses
did not achieve statistical significance (Table 3 and Fig. 3;
-926 ± 1933 vs. 7369 ± 3055;
P = 0.09). Noradrenaline and glucagon responses to
hypoglycemia were not altered by ILH (Table 3).
|
|
; AUC, 647 ± 162
vs. 1222 ± 874 score points/180 min; P
= 0.028). A trend for reduced autonomic symptoms failed to reach
significance, as did glycemic thresholds for symptom generation (Table 4).
|
|
). Reaction time and/or accuracy
of reaction deteriorated at plasma glucose thresholds of 2.6 ±
0.1 mmol/L with ILH and 2.7 ± 0.1 mmol/L during CON studies
(P = 0.75). Similarly, there was no significant
difference in the AUC (2250 ± 1080 vs. 1420 ±
710 ms/min; P = 0.59).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
At euglycemia we found no effect of the experimental infusion of Intralipid and heparin on four-choice reaction time or on catecholamines. The suppressive effects of NEFA on GH secretion are well documented (33), and the fall in cortisol we observed was consistent with the normal diurnal rhythm of this hormone. The only symptoms to register were hunger and tiredness, which we have interpreted as a relatively nonspecific effect of prolonged studies in fasting individuals.
During hypoglycemia, ILH reduced counterregulatory hormones, as reflected by epinephrine, GH, and, to a lesser degree, cortisol responses. The stimulus for generation of these responses is a fall in metabolic rate in the putative cerebral glucose sensing area(s) (10, 34). It seems likely that the ILH infusion is exerting these effects by maintaining metabolism in this sensing area(s). These effects are analogous to those seen with infusions of lactate or BOB (5, 6, 7). No changes were seen in noradrenaline or glucagon responses. The former represents a relatively insensitive measurement of synaptic transmission and increased sympathetic outflow during hypoglycemia. Glucagon may be at least partly under the control of local pancreatic mechanisms, and contrasting effects have been found previously with other alternate fuel studies, suggesting that glucagon responses may be resistant to the provision of nonglucose fuels during hypoglycemia (35).
Total symptoms were also reduced by ILH, as were neuroglycopenic symptoms. The difference in autonomic symptoms failed to reach statistical significance. The explanation for this and also for the lack of a significant difference in thresholds for symptoms may again be the relatively high reporting of the common, relatively nonspecific symptoms of hunger and tiredness, which will reduce the power of our study to detect differences. When we reexamined autonomic symptom thresholds without hunger, the differences in thresholds approached statistical significance (2.8 ± 0.1 vs. 3.1 ± 0.2 mmol/L, ILH vs. CON; P = 0.06). The origin of neuroglycopenic symptoms is assumed to be brain glucopenia, although the exact mechanisms are not yet defined. The disparity between our findings for neuroglycopenic symptoms and cognitive function discussed below suggests that these functions originate in different brain regions.
Which of the lipid substrates is likely to be responsible for these actions? Triglycerides (36) are not transported across the blood-brain barrier. Although no specific carrier for glycerol has been firmly identified, it does penetrate into the cerebrospinal fluid when given iv (37). In contrast, specific blood-brain barrier transport systems for NEFA have been well documented in mammals (11, 17). Enzyme systems for ß-oxidation (38) and gluconeogenesis (13, 14, 15, 39) have been described, so that delivered NEFA and/or glycerol may be oxidized directly to provide energy or converted into glucose or other intermediary metabolites within the brain.
BOB is transported by the monocarboxylic acid transporter, which also transports lactate across the blood-brain barrier (40). Similarly, the enzyme systems necessary for metabolism of ketone bodies have been identified in human brain (16). This reduction in neurohumoral and symptomatic responses to hypoglycemia is therefore most likely due to either the elevation of NEFA or glycerol or to the modest elevation in BOB seen with ILH (or to a combination of substrates). However, the increased circulating BOB levels were lower than levels that have been shown to be effective in diminishing counterregulatory hormone responses in earlier work. Previous studies using BOB have demonstrated effects with concentrations of 580 (7) and 1900 µmol/L (6) respectively, compared to a peak level in our study of 257 µmol/L. Furthermore, this modest rise in BOB was not apparent, at least statistically, until after the onset of neurohumoral responses. Finally, previous studies in which BOB has been shown to have an effect have shown an equivalent alteration of thresholds for neurohumoral responses and cognitive dysfunction (6, 7). We, therefore, believe that it is unlikely that BOB is responsible for the reduction in counterregulatory responses. Lactate cannot explain our findings, as levels rose equally during hypoglycemia during ILH and CON, and only achieved peak levels less than half of those shown to delay counterregulatory responses in earlier studies (5, 6). This leaves NEFA and/or glycerol as the most likely candidates.
We found no effect of ILH on the glycemic thresholds for deterioration in cognitive function. Examining the incremental changes in reaction time might suggests a trend toward a difference in the presence of ILH, albeit not statistically significant. However, any effect is small and differs from the marked differences seen with other alternate substrates, such as BOB or lactate infusions, and from the effects of ILH on symptomatic and neurohumoral responses to hypoglycemia.
Four-choice reaction time measures only one aspect of cognitive function. We used this sole test because of the advantages that its brevity affords, allowing repeated administration and the sensitive determination of thresholds for cognitive dysfunction during stepped hypoglycemia. Four-choice reaction testing has been the best characterized of the plethora of available cognitive tests and has previously been shown to be not only reproducible and robust when applied as a single measure, but also sensitive to hypoglycemia (20). It is possible that some other aspects of cognition may be more readily supported by the lipid substrates. This does not, however, affect the conclusion that the areas of brain involved in an important cognitive function cannot use NEFA or glycerol as alternatives to glucose during hypoglycemia in the same way in which they can use lactate or BOB. Previous studies on alternate substrates in hypoglycemia have, in fact, shown similar results whether using four-choice reaction time as a single cognitive test or a battery of cognitive tests (5, 6, 7).
The results suggest that there may be regional differences in the brain’s capacity to use alternate substrates during hypoglycemia. The glucose sensor(s) is able to adapt and sustain metabolism in the presence of nonglucose lipid substrates, as shown by the delayed and diminished symptomatic and hormonal responses, whereas cognitive areas of the brain cannot do so significantly. These differences could conceivably be because of differing substrate delivery (regional differences in blood-brain barrier transport or blood flow) or because of a variation in enzyme distribution and thus metabolism. In recent years considerable heterogeneity between regions has been demonstrated in cerebral metabolism for both glucose and nonglucose substrates (41, 42).
Patients with type 1 diabetes mellitus who have hypoglycemia unawareness have a greatly increased risk of severe hypoglycemia, with the attendant medical, psychological, and social costs (43, 44). A greater understanding of regional differences in metabolism between different brain may help increase understanding about the mechanisms of genesis of hypoglycemia unawareness. This, in turn, may help move us toward clinical strategies that will protect the brain during hypoglycemia and diminish the problem of severe hypoglycemia in patients with diabetes.
Received March 19, 1997.
Revised October 9, 1997.
Revised May 8, 1998.
Accepted May 18, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  L. Jacobson, T. Ansari, and O. P. McGuinness Counterregulatory deficits occur within 24 h of a single hypoglycemic episode in conscious, unrestrained, chronically cannulated mice Am J Physiol Endocrinol Metab, April 1, 2006; 290(4): E678 - E684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Wang, C. Cruciani-Guglielmacci, S. Migrenne, C. Magnan, V. E. Cotero, and V. H. Routh Effects of Oleic Acid on Distinct Populations of Neurons in the Hypothalamic Arcuate Nucleus Are Dependent on Extracellular Glucose Levels J Neurophysiol, March 1, 2006; 95(3): 1491 - 1498. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. K. Dalsgaard, S. Ogoh, E. A. Dawson, C. C. Yoshiga, B. Quistorff, and N. H. Secher Cerebral carbohydrate cost of physical exertion in humans Am J Physiol Regulatory Integrative Comp Physiol, September 1, 2004; 287(3): R534 - R540. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Hansen, J. F. J. Kun, J. E. Schultz, and E. Beitz A Single, Bi-functional Aquaglyceroporin in Blood-stage Plasmodium falciparum Malaria Parasites J. Biol. Chem., February 8, 2002; 277(7): 4874 - 4882. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
