This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Feneberg, R. Articles by Schaefer, F. Search for Related Content PubMed PubMed Citation Articles by Feneberg, R. Articles by Schaefer, F.

Synchronous Fluctuations of Blood Insulin and Lactate Concentrations in Humans¹

Divisions of Pediatric (R.F., M.S., O.M., F.S.) and Clinical Nephrology (E.R.), Ruperto-Carolus University, 69120 Heidelberg, Germany; and Department of Internal Medicine (J.D.V.), University of Virginia Health Sciences Center and National Science Foundation Center for Biological Timing, Charlottesville, Virginia 22908

Address all correspondence and requests for reprints to: Franz Schaefer, M.D., Pediatric Nephrology Division, Im Neuenheimer Feld 150, 69120 Heidelberg, Germany. E-mail: franz_schaefer{at}ukl.uniheidelberg.de

	Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Oscillatory organization is a universal mode of signal transduction in living organisms. In vitro studies suggest spontaneous pulsatile fluctuations of intracellular energy metabolism. It is possible that, in vivo, some of these processes are synchronized by the pulsatile release of insulin. We assessed a potential coupling among plasma insulin, glucose, and lactate concentrations, by frequent blood sampling for 24 h in 11 healthy volunteers. Insulin sensitivity was assessed using the euglycemic hyperinsulinemic clamp technique. Lactate concentrations exhibited pulsatile fluctuations at an average interval of 84 ± 11 min, whereas sodium and pH were nonpulsatile. The lactate concentration pulses closely corresponded to insulin oscillations, which occurred with a periodicity of 86 ± 11 min. Blood glucose also fluctuated during daytime at an interval of 89 ± 32 min. During nighttime, the frequency and amplitude of glucose oscillations were lower. The daytime profiles showed significant temporal coupling and pattern synchrony among insulin, lactate, and glucose. Only the close temporal relationship between insulin and lactate release persisted during nighttime. The temporal coupling and pattern synchrony between insulin and lactate were correlated inversely with insulin sensitivity, and positively with the degree of abdominal obesity. Our results suggest that: 1) the concentration of lactate, an indicator of cellular energy metabolism, fluctuates periodically in vivo; 2) the lactate concentrations fluctuate in synchrony with insulin pulses; and 3) such coupling is more pronounced in obese, insulin-resistant individuals.

	Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References

 
THE ISLETS of Langerhans secrete insulin in a pulsatile fashion in animals (1, 2, 3) and humans (4). Two dominant frequencies of pulsatile insulin release have been identified: high-frequency pulses occur at a periodicity of 8–16 min (4); low-frequency, high-amplitude oscillations, with a period length of approximately 90 min, are superimposed (5, 6, 7). The slower insulin pulses are an intrinsic property of the ß-cell and persist during continuous infusion of glucose (8, 9). Nevertheless, they are completely entrainable by intermittent enteral uptake of nutrients, particularly glucose (10), and by oscillatory glucose infusion (8). The physiological function of these slow insulin oscillations is largely unknown (11).

The rate of glycolysis of intact cells also oscillates (12, 13), with phosphofructokinase apparently acting as the primary oscillophore (14, 15). Oscillations of energy metabolism may be the mechanism underlying spontaneous pulsatility of insulin secretion from the pancreatic islets (16, 17). We reasoned that, in turn, pulsatile insulin release might entrain synchronous fluctuations of energy metabolism at the whole-body level. To evaluate this hypothesis, we examined volunteers and assessed whether the concentrations of blood lactate, one indicator of intracellular energy metabolism, fluctuate periodically and whether the variations of lactate concentrations are coupled with those of plasma insulin and/or glucose.

	Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Experimental Subjects

Eleven healthy male volunteers [body mass index (BMI), 19.7–26.7 kg/m²; ages, 24–69 yr] participated in the study. The anthropometric and endocrine characteristics are given in Table 1. Relevant diseases were excluded by history, physical examination, and routine laboratory investigation. There was no known diabetes mellitus in first-degree relatives of study subjects. No medication was taken for at least 1 month before the study. The subjects were advised to maintain food intake and physical activity constant (and keep records thereof) and to abstain from alcoholic beverages and caffeine for at least 2 weeks before the study. All participants were nonsmokers.

Study protocol. All studies were performed after an overnight fast. At 0730 h, a cannula was inserted into a cubital vein for blood sampling. From 0800 h (day 1) to 0800 h (day 2), 1.3-mL blood specimens were withdrawn at 15-min intervals without tourniquet use.

Three meals were prepared on the basis of individual food preferences. The meals were standardized with respect to carbohydrate, protein, and fat content and were consumed at 0800 h, 1300 h, and 1900 h. Each meal contained 45% carbohydrates, 20% protein, and 35% fat, delivering 36 ± 7 kcal/kg BW·day. The subjects were allowed to walk in a quiet environment and went to bed at 2300 h. Sleeping was confirmed by inspection and interview in the morning. Subsequently, starting at 0800 h on the second day, a 2-h hyperinsulinemic euglycemic clamp study was performed. For this procedure, a second iv line was inserted immediately before the clamp study.

The study protocol was approved by the Ethics Committee of the University of Heidelberg. All participants gave their voluntary written informed consent before the study.

Analytical methods. One aliquot of the blood sample was used for measurement of electrolytes and pH immediately after sampling. Another aliquot was deproteinized with perchloric acid and stored at 4 C for measurement of blood lactate. The residual volume was stored on ice and centrifuged after clotting. The serum was stored at -80 C. All measurements were performed within 2 weeks after the study.

Serum insulin concentrations were determined using the Insulin ELISA (enzyme-linked immunosorbent assay; Dako Diagnostika, Hamburg, Germany), a highly sensitive two-site immunospecific ELISA with two monoclonal murine antibodies. The detection limit of the assay is 0.5 mU/L. There is no cross-reactivity with human C-peptide and virtually none with intact human proinsulin (0.3%), 32–33 split proinsulin (0.3%), and des-31–32 split proinsulin (0.5%). The antibodies cross-react with 65–66 split proinsulin (45%) and des-64–65 split proinsulin (66%). The concentrations of the latter two split products in healthy humans are negligible (18). The interassay coefficient of variation was 8.4%. The intraassay coefficient of variation was 6.3%, 5.0%, and 4.9% at 0–5, 5–10, and >10 mU/L, respectively. Each sample was assayed in duplicate.

Whole-blood lactate was measured by enzymatic-amperometric determination using the EBIO plus analyzer (Eppendorf, Hamburg, Germany). The assay protocol was modified to increase sensitivity to a detection limit of 0.25 mmol/L lactate. The intraassay coefficient of variation was less than 3% at 2 mmol/L.

For the 24-h profiles, blood glucose was measured using the hexokinase method on a Hitachi 705 autoanalyzer (Boehringer Mannheim, Mannheim, Germany). During the euglycemic clamp studies, glucose concentrations were determined by the glucokinase method, using the Glucose Analyzer II (Beckman Coulter, Inc. Instruments, Munich, Germany). Sodium and pH were measured using an ion-selective electrode device (Fresenius Ionometer EF-H, Fresenius, Bad Homburg, Germany).

Euglycemic hyperinsulinemic clamp studies. The euglycemic hyperinsulinemic clamp studies were performed as described in detail by DeFronzo et al. (19), with slight modifications. A priming bolus of insulin (100 mU/m² per min, H-Insulin, Hoechst, Frankfurt, Germany) was infused through a venous line for 2 min. The insulin infusion rate was decreased stepwise every 2 min, from 90 to 80 and 60 mU/m²/min, to reach a constant maintenance rate of 40 mU/m²/min. Plasma insulin levels were raised to 69.1 ± 20 mU/L, i.e. about 10 times the basal level. Five minutes after the start of the insulin infusion, a 20% glucose infusion (Glucosteril 20%, Fresenius) was started through the same venous line. Arterialized plasma glucose levels were determined at 5-min intervals throughout the duration of the insulin infusion. Starting at a glucose dose of 2.5 mg/kg·min, the infusion rate was subsequently adjusted continuously to keep the blood glucose concentration within ±10% of the baseline level. The mean intrastudy coefficient of variation of plasma glucose concentrations was 9%.

Data analysis.

Peak detection.The Cluster Analysis program was used as a waveform-independent technique to detect significant peaks in each plasma analyte profile (20). Nadir test cluster sizes were set to 2, and peak test cluster sizes were set to 1. Minimum t statistics of 1.96 were used as a stringency criteria both for upstrokes and downstrokes. These settings detect significant fluctuations, with 2.1% false-positive and <10% false-negative detection errors, provided that the signal-to-noise ratio in the profile is 1.3.

Each insulin concentration value was assigned an SD value based on the concentration-dependent power function of assay variance against sample concentration derived from all duplicate measurements in an individual profile. The same procedure was applied for the lactate and glucose measurements (determined in singlets) using an assay variance curve derived from 500 duplicate measurements obtained in quality control series and individual blood lactate profiles assessed in duplicate. Constant CVs, derived from daily quality control measurements, were assigned to the singlet measurements of sodium (3%) and pH (3%).

Coincidence analysis.Possible associations between the blood concentration patterns of pairs of analytes were assessed by two statistical approaches. First, the temporal coupling between the peaks, detected by Cluster Analysis in any two concentration-vs.-time series in an individual, was analyzed using the program ANCOPULS (21, 22). Using the hypergeometric probability distribution (21), this program estimates the probability that the number of coincident peaks in two series may be caused by random associations only. The cumulative probability that at least the observed number of peak coincidences in the concatenated series is caused by chance alone equals the probability of falsely rejecting the null hypothesis of purely random coupling of the peaks in the two time series (22). Second, cross-correlation and cross-approximate entropy (Cross-ApEn) analyses were employed to evaluate synchrony between relevantly paired time series. Both of these methodologies are independent of pulse analysis and, hence, are important to complement pulse analysis. Cross-correlation analysis was applied to determine any temporal association in the blood concentrations of pairs of analytes. The cross-correlation coefficients between the concentrations of a pair of analytes at individual lags of -150 to +150 min at 15-min intervals were transformed to standardized z-scores according to the equation:

where r is the cross-correlation coefficient; n, the number of data points in each 15-min series; and k, the number of lag units. For each lag, the Kolmogorov-Smirnov statistic was applied, to assess whether the distribution of resultant z-scores in the study population departed significantly from a mean of zero with unit normal SD (22). The resultant P value denotes the probability that the set of cross-correlation r values at that lag is caused by chance correlations alone.

Approximate entropy (ApEn) statistics.The scale- and model-independent ApEn statistic provides a measure of regularity (orderliness) of fluctuations in a given hormone time series (23, 24). This statistic evaluates the negative logarithm of the probability that any given particular data pattern length of m consecutive points will be repeated within a tolerance or distance r on next incremental comparison. Here, m was set at 1, and r at 0.2 (20%) of the respective substance SD, to normalize ApEn against different absolute substrate concentrations (25). ApEn values typically lie between zero (perfectly ordered) and 2–3 (highly random) in this circumstance.

Additionally, to quantify asynchrony (conditional irregularity), we used Cross-ApEn. Cross-ApEn denotes the conditional regularity of two series, such that significant synchrony reflects a tendency of patterns in one series to be reproduced in the other. Cross-ApEn can be employed to compare sequences from two distinct (yet intertwined) variables in a network, herein applied to the joint insulin, lactate, and glucose time-series. The precise definition is thematically similar to that for ApEn. For this study, we applied Cross-ApEn, with m = 1 and r = 0.2, to standardized time-series data; i.e. for each subject, we applied Cross-ApEn (1, 0.2) to the {u x (l),ⁱⁱ v x (l)} series, where u x (l) = [u(i) - mean u]/SD u and v x (i) = [v(i) - mean v]/SD v. This standardization, in conjunction with the choice of m and r, ensures good replicability properties for Cross-ApEn for the data lengths studied (26).

Calculation of insulin sensitivity.The glucose disposal rate (insulin-mediated glucose uptake, M value), the ratio of the M value to the insulin plasma concentrations (M/I ratio), and the metabolic clearance rate of insulin (MCR) were calculated for the last 100 min of the hyperinsulinemic euglycemic clamp studies, at 20-min intervals, as described previously (1).

Statistics. For analytical purposes, the 24-h observation period was subdivided into a daytime and a nighttime period. The daytime period started with the first blood sample and the start of breakfast at 0800 h. The daytime period was further subdivided into pre- and postprandial periods. The postprandial state was defined by the broad waves of insulin release that regularly occurred after each of the three meals. The end of the postprandial insulin waves was defined as the first of two consecutive plasma insulin values that were within 1 SD of the mean baseline insulin concentration, as calculated by Cluster Analysis. The start of the nighttime period was given by the end of the postprandial period after dinner.

Data are given as mean ± SD, unless indicated otherwise. Statistical contrasts, Pearson’s correlation coefficients, and z-scores were assumed to be significant at P < 0.05. Analogously, an error probability of less than 0.05 was defined to reflect nonrandom coincidence of discrete peaks.

	Results

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Insulin

Visual inspection of the plasma insulin concentration vs. time profiles revealed three major epochs of increased insulin release after meals, which persisted 126 ± 40 min after breakfast, 78 ± 40 min after lunch, and 198 ± 54 min after dinner, respectively. Although the postprandial increase in plasma insulin was usually composed of one or two high-amplitude bursts, regular low-amplitude peaks were also detected during the preprandial periods. Daytime plasma insulin pulses occurred at an average interval of 77 min; the pulse frequency was higher in the postprandial periods, compared with the preprandial periods (0.89 ± 0.26 h^-1vs. 0.52 ± 0.24 h^-1, P < 0.01). During the nighttime fasting period (10.2 ± 1.5 h), the frequency of insulin concentration pulses (0.52 ± 0.16 h^-1) was significantly lower (P < 0.0005) than during the postprandial daytime periods but was similar to that seen during the daytime preprandial periods. The fractional amplitudes of the insulin pulses were more than two times higher during the postprandial than during the preprandial period (769 ± 1104% vs. 300 ± 364%, P < 0.0001, nadir basal = 100%) but did not differ between the daytime preprandial period and the nighttime period (190 ± 62%).

Lactate

Significant fluctuations of blood lactate concentrations also were identified in each subject. The daytime lactate pulses occurred at intervals of approximately 70 min, without a significant difference between the pre- and postprandial phases (0.70 ± 0.21 vs. 0.70 ± 0.18 h^-1). During nighttime, lactate bursts were less frequent than during daytime (0.59 ± 0.21 h^-1, not significant, Table 2). The fractional amplitudes of the lactate pulses were higher in the postprandial (168 ± 56%) than in the preprandial (140 ± 36%, P < 0.01) or nighttime periods (138 ± 13%, P < 0.01). However, even during nighttime, the lactate peak amplitudes exceeded the individual mean baseline concentrations by an average of 132 ± 11% and the intraassay CV by 1000%.

Significant fluctuations of glucose were noted in each subject. The frequency of the glucose oscillations was significantly lower than that of insulin and lactate, both during the preprandial period (0.47 ± 0.19 h^-1; glucose vs. insulin: P < 0.01, glucose vs. lactate: P < 0.02) and during nighttime (0.30 ± 0.14 h^-1; glucose vs. insulin: P < 0.01, glucose vs. lactate: P < 0.005). The frequency of postprandial glucose oscillations (0.79 ± 0.18 h^-1) was similar to that of insulin and lactate. The mean postprandial glucose pulse amplitude was 149 ± 41%, compared with 139 ± 61% in the preprandial period and 112 ± 5% in the nighttime period.

The amplitudes of the glucose oscillations, relative to the preceding baseline, were smaller for glucose than for insulin (P < 0.0001) and lactate (P < 0.0001) (Table 2).

Sodium and pH

Sodium concentrations and pH were measured in all samples as internal controls, to exclude artifacts caused by factors such as sampling errors and variations in the state of hydration. Results showed constancy for these parameters in all study subjects. The mean venous pH was 7.41 ± 0.03. Sodium concentration was 139 ± 1 mmol/L. Variations in the blood pH (<2%) and sodium concentrations (<5%) did not exceed the error of measurement. No significant peaks were detected by Cluster Analysis in any of the time series.

Coincidence analysis

Simultaneous occurrence of insulin, lactate, and glucose concentration peaks was suggested by visual inspection (Figs. 1 and 2). This impression was confirmed by the statistical analysis of coincidence.

View larger version (21K): [in this window] [in a new window]   Figure 1. Circadian profiles of insulin, lactate, and glucose concentrations. Example of an individual study. Triangles () indicate significant pulses. Simultaneous occurrence of insulin and lactate (L) and insulin and glucose (G) pulses is indicated above the triangles. The time of meals is denoted below the insulin profile.

View larger version (19K): [in this window] [in a new window]   Figure 2. Circadian profiles of insulin, lactate, and glucose concentrations. Example of an individual study during the nighttime period. Triangles () indicate significant pulses. Simultaneous occurrence of insulin and lactate (L) and insulin and glucose (G) pulses is indicated above the triangles.

Cross-correlation analysis was used to evaluate nonrandom temporal associations between pairs of concentration vs. time profiles. The results of the cross-correlation analysis are illustrated in Fig. 3. During daytime, significant temporal associations were detected between insulin and glucose (P < 0.01), insulin and lactate (P < 0.01), and glucose and lactate concentrations (P < 0.01). The synchrony of insulin and lactate fluctuations during daytime was maximal at lags of 0 and 15 min, suggesting that changes in lactate closely followed changes in plasma insulin concentrations. Changes in blood glucose concentrations preceded those of insulin by 0–15 min. Accordingly, lactate concentrations followed glucose concentrations at an average lag of 30 min. During the nighttime observation period, the coupling between insulin and glucose profiles and between glucose and lactate was lost. In contrast, the synchrony of insulin and lactate oscillations persisted during nighttime (P < 0.05 at lag 0 min).

View larger version (33K): [in this window] [in a new window]   Figure 3. Insulin, lactate, and glucose-correlation analysis. Cross-correlation coefficients are rendered as standardized z-scores (see Subjects and Methods).

The regularity (or orderliness) of insulin and lactate concentration changes was significantly (P < 0.01) higher than that of glucose concentration changes for the total observation period. The regularity of fluctuations of each analyte was significantly (P < 0.05) enhanced during nighttime, compared with daytime (see Table 2).

As summarized in Table 3, we observed no day/night differences for any hormone/metabolite pair, by way of nonparametric paired comparisons. However, insulin/lactate were significantly more synchronous (P < 0.0154) than glucose/lactate or than glucose/insulin (P = 0.0366, by ANOVA). In addition, the ratio of random to observed Cross-ApEn significantly exceeded 1 for each of the following circumstances, thus indicating nonrandom synchrony for the hormone/metabolite pairs: insulin/lactate (at all times); insulin/glucose (total 24-h period); and lactate/glucose (total 24-h period).

The mean M value of the euglycemic hyperinsulinemic clamp studies was 8.4 ± 2.7 mg/kg/min, resulting in an M/I ratio of 14.6 ± 7.5. An MCR of 936 ± 323 mL/m²·min was observed. The results of the euglycemic hyperinsulinemic clamp studies did not correlate with age, BMI, or the waist-to-hip ratio.

Factors possibly affecting the degree of coupling between insulin and lactate concentration changes were evaluated by Pearson univariate correlation analysis. A significant correlation was noted between the waist-to-hip ratio and the insulin-lactate cross-correlation coefficients (r = 0.66, P < 0.01) and the insulin-lactate Cross-ApEn values during nighttime (r = 0.69, P < 0.05) (Fig. 4). Moreover, the insulin-lactate cross-correlation coefficient values (r = -0.69, P < 0.05) and the insulin-lactate Cross-ApEn values during nighttime (r = -0.63, P < 0.05) were inversely correlated with the glucose uptake rate (M value) in the euglycemic clamp study. The waist-to-hip ratio was not correlated with the M value in this population. Both abdominal obesity and insulin sensitivity seemed to be independent predictors of insulin-lactate copulsatility. Other anthropometric parameters and age were not correlated with the degree of coupling between insulin and lactate concentrations.

View larger version (10K): [in this window] [in a new window]   Figure 4. Relationship between insulin-lactate coupling (z-score at lag = 0 min) and waist-to-hip ratio.

	Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References

To our knowledge, the present observation of episodic pulses of blood lactate concentrations occurring throughout the day and night is the first description of spontaneous fluctuations of this metabolite in vivo. The observed lactate concentration pulses not only exceed the technical error associated with the measurements by far (>1000%) but also span virtually the full normal resting range of random single measurements of blood lactate. The reliability and specificity of the observed blood lactate fluctuations are supported by the complete absence of concentration pulses in the sodium and pH profiles, which were determined in the same samples as internal controls.

Most of the lactate circulating in the resting state is derived from anaerobic glycolysis in erythrocytes, adipocytes, and myocytes (27, 28, 29). Lactate, in turn, serves as a substrate for either gluconeogenesis or (via decarboxylation to acetyl-CoA) for energy production or lipogenesis. Whereas liver and kidney are the exclusive sites of gluconeogenesis, the heart and other organs metabolize lactate as a major source of ATP synthesis. Hence, the periodic fluctuations of blood lactate observed here may reflect synchronized variations of glycolytic lactate production, lactate use, or both.

The observed periodicity of blood lactate concentrations raises the question as to the mechanism which coordinates the synchronous release and/or uptake of this metabolite. In principle, humoral or neuronal mechanisms are potential candidates.

Insulin, the key hormone of energy metabolism, exerts multiple physiological functions that make it a candidate for synchronization of blood lactate concentrations. Insulin is released from the pancreas in episodic pulses. It modulates the activity of phosphofructokinase, a key enzyme of glycolysis, and that of several other enzymes involved in gluconeogenesis and glycogen synthesis. Using cross-correlation analysis, we observed close temporal coupling between insulin and lactate concentrations. Our results suggest that changes in insulin concentrations precede changes in lactate concentrations by 0–15 min. Moreover, Cross-Apen analysis disclosed significantly nonrandom synchrony of the patterned release of insulin, lactate, and glucose. This measure is complementary to conventional cross-correlation assessment, because cross-ApEn is independent of lag. This synchrony was statistically significant for insulin/lactate, insulin/glucose, and lactate/glucose during the daytime and integrated 24-h time periods. In contrast, during the nighttime interval of observation, only insulin/lactate showed substantial cross-synchrony. Consequently, the tightest pattern coupling exists between insulin/lactate, rather than insulin/glucose or lactate/glucose.

These findings support (but do not prove) a driving role of insulin in the genesis of blood lactate concentration pulses. Because intermittent intake of nutrients induced large increases of glucose, insulin, and lactate concentrations, it may be argued that both insulin and lactate concentration pulses were the result of cellular glucose uptake stimulating insulin release and increasing substrate availability for lactate production. Though such a mechanism would explain insulin-lactate copulsatility in the postprandial periods, it cannot be responsible for the persistent temporal association between insulin and lactate in the fasting state, particularly at nighttime.

Further studies will be required to test whether the observed blood lactate fluctuations are caused by low-frequency insulin pulses. Alternative possibilities also can be considered, including a common regulation of insulin and lactate pulses by periodic changes of autonomic nervous system inputs to the pancreas and the tissues involved in lactate metabolism, simultaneous changes in the perfusion of these tissues, and an unidentified third agent triggering both the insulin and lactate pulses. A common neuronal mechanism seems less likely, because both high- and low-frequency insulin oscillations persist in the isolated pancreas ex vivo (30) and even in isolated Langerhans’ islets (31), suggesting that insulin pulsatility is an intrinsic feature of the ß-cell.

Alternatively, because insulin modulates vascular tone (32, 33), insulin pulses might also cause periodic changes in tissue metabolism by periodically altering tissue perfusion. Finally, we cannot exclude potential roles of other hormones (such as glucagon, gastrointestinal peptides, or catecholamines) in synchronizing plasma lactate concentrations (34, 35).

We observed a relationship between the strength of insulin-lactate coupling and insulin sensitivity, as well as the degree of abdominal obesity. Lactate pulses were apparently more closely coupled to insulin pulses in relatively obese subjects and in those with low-normal insulin sensitivity. Adipose tissue is an important source of circulating lactate (36, 37, 38), and insulin synchronizes oscillatory glycolysis in rat fat cells (39). In obese individuals, a higher proportion of lactate may be released from the insulin-sensitive visceral or other adipose tissue, compared with lean subjects, in whom a greater proportion of circulating lactate may be derived from the insulin-insensitive erythrocytes (39, 40). Therefore, the observed relationship would be compatible with insulin synchronizing periodic lactate release.

In summary, we have demonstrated that blood lactate concentrations oscillate in humans. Such oscillations of lactate concentrations are closely coupled (at a 0–15 min lag) to low-frequency insulin pulsations. We hypothesize that insulin synchronizes intracellular metabolic processes in a pulsatile fashion. The full implications of this dynamic linkage in health and disease remain to be explored.

	Acknowledgments

 
The authors are indebted to the contribution and patience of the participating volunteers. Ilona Gomminginger (AOK Heidelberg) kindly supported the recruitment of volunteers. Expert laboratory assistance was provided by Renate Nitze (from the Department of Medicine) and Heiko Kienzler, Elke Kohl, Jutta Mühlemeier, and Karin Schnorr (from the Department of Sports Medicine at the University of Heidelberg). We appreciate the helpful comments of Professor H. J. Bremer and Professor P. Bärtsch. We are also grateful to P. Bahrmann and D. Fliser for their helpful instructions with the euglycemic clamp technique.

	Footnotes

 
¹ This work was supported by the Else Kröner-Fresenius Stiftung (Bad Homburg, Germany), Novo Nordisk A/S (Copenhagen, Denmark), Dako, Diagnostika, (Hamburg, Germany), the NSF Science and Technology Center for Biological Timing, and NIH Grant RCDA-1-K-04-HD-00634 (IPU).

² Recipient of a scholarship of the Deutsche Forschungsgemeinschaft.

Received August 17, 1998.

Revised September 11, 1998.

Accepted September 21, 1998.

	References

Top Abstract Introduction Subjects and Methods Results Discussion References

 

1. Goodner CJ, Walike BC, Koerker DJ, et al. 1977 Insulin, glucagon and glucose exhibit synchronous, sustained oscillations in fasting monkeys. Science. 195:177–179.[Abstract/Free Full Text]
2. Chou HF, McGivern R, Berman N, Ipp E. 1991 Oscillations of circulating plasma insulin concentrations in the rat. Life Sci. 48:1463–1469.[CrossRef][Medline]
3. Koerker DJ, Goodner CJ, Hansen BW, Brown AC, Rubenstein AH. 1978 Synchronous, sustained oscillation of C-peptide and insulin in the plasma of fasting monkeys. Endocrinology. 102:1649–1652.[Abstract]
4. Lang DA, Matthews DR, Peto J, Turner RC. 1979 Cyclic oscillation of basal plasma glucose and insulin concentrations in human beings. N Engl J Med. 301:1023–1027.[Abstract]
5. Polonsky KS, Given BD, Van Cauter E. 1988 Twenty-four-hour profiles and pulsatile patterns of insulin secretion in normal and obese subjects. J Clin Invest. 81:442–448.
6. Simon C, Brandenberger G, Follenius M. 1987 Ultradian oscillations of plasma glucose, insulin, and C-peptide in man during continuous enteral nutrition. J Clin Endocrinol Metab. 64:669–674.[Abstract]
7. Simon C, Brandenberger G, Saini J, Ehrhart J, Follenius M. 1994 Slow oscillations of plasma glucose and insulin secretion rate are amplified during sleep in humans under continuous enteral nutrition. Sleep. 17:333–338.[Medline]
8. Sturis J, Van Cauter E, Blackman JD, Polonsky KS. 1991 Entrainment of pulsatile insulin secretion by oscillatory glucose infusion. J Clin Invest. 87:439–445.
9. Cunningham BA, Deeney JT, Bliss JC, Corkey BE, Tornheim K. 1996 Glucose-induced oscillatory insulin secretion in perifused rat pancreatic islets and clonal beta-cells (HIT). Am J Physiol. 271:E702–E710.
10. Molnar GD, Taylor WFI. 1998 Plasma immunoreactive insulin patterns in insulin-treated diabetics. Mayo Clin Proc. 47:709–719.
11. Tornheim K. 1997 Are metabolic oscillations responsible for normal oscillatory insulin secretion? Diabetes. 46:1375–1380.[Abstract]
12. Chance B, Estabrook RW, Ghosh A. 1964 Damped sinusoidal oscillations of cytoplasmic reduced pyridine nucleotides in yeast cells. Proc Natl Acad Sci USA. 51:1244–1251.[Free Full Text]
13. Betz A, Chance B. 1965 Phase relationship of glycolytic intermediates in yeast cells with oscillatory metabolic control. Arch Biochem Biophys. 109:585–594.[CrossRef][Medline]
14. Boiteux A, Goldbeter A, Hess B. 1975 Control of oscillating glycolysis of yeast. Proc Natl Acad Sci USA. 72:3829–3833.[Abstract/Free Full Text]
15. Goldbeter A, Nicolis G. 1976 An allosteric enzyme model with positive feedback applied to glycolytic oscillations. In: Rosen R, ed. Progress in theoretical biology. New York: Academic Press; 65–160.
16. Chou HF, Berman N, Ipp E. 1992 Oscillations of lactate released from islets of Langerhans: evidence for oscillatory glycolysis in beta-cells. Am J Physiol. 262:E800–E805.
17. Prenki M. 1996 New insights into pancreatic beta-cell metabolic signaling in insulin secretion. Eur J Endocrinol. 134:272–286.[Abstract]
18. Sobey WJ, Beer SF, Carrington CA, et al. 1989 Sensitive and specific two-site immunoradiometric assays for human insulin, proinsulin, 65–66 split and 32–33 split proinsulins. Biochem J. 260:535–541.[Medline]
19. DeFronzo RA, Tobin JD, Andres R. 1979 Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 237:E214–E223.
20. Veldhuis JD, Johnson ML. 1986 Cluster analysis: a simple, versatile and robust algorithm for endocrine pulse detection. Am J Physiol. 250:E486–493.
21. Veldhuis JD, Johnson ML, Seneta E. 1991 Analysis of the co-pulsatility of anterior pituitary hormones. J Clin Endocrinol Metab. 73:589–576.
22. Veldhuis JD, Johnson ML, Faunt LM, Seneta E. 1994 Assessing the temporal coupling between two, or among three or more, neuroendocrine pulse trains: cross-correlation analysis, simulation methods, and conditional probability testing. Methods Neurosci. 20:336–376.
23. Pincus SM. 1998 Approximate entropy as a measure of system complexity. Proc Natl Acad Sci USA. 88:2297–2301.[Abstract/Free Full Text]
24. Pincus SM, Keefe DL. 1992 Quantification of hormone pulsatility via an approximate entropy algorithm. Am J Physiol. 262:E741–E754.
25. Hartman ML, Pincus SM, Johnson ML, et al. 1994 Enhanced basal and disorderly growth hormone (GH) secretion distinguish acromegalic from normal pulsatile GH release. J Clin Invest. 94:1277–1288.
26. Pincus SM, Mulligan T, Iranmanesh A, Gheorghiu S, Godschalk M, Veldhuis JD. 1996 Older males secrete luteinizing hormone and testosterone more irregularly, and jointly more asynchronously, than younger males: dual novel facets. Proc Natl Acad Sci USA. 93:14100–14105.[Abstract/Free Full Text]
27. Dombrowski Jr GJ, Swiatek KR. 1991 Lactate genesis by rat liver and muscle during development. Pediatr Res. 30:331–336.[Medline]
28. Hagstrom E, Arner P, Ungerstedt U, Bolinder J. 1990 Subcutaneous adipose tissue: a source of lactate production after glucose ingestion in humans. Am J Physiol. 258:E888–893.
29. Kreisberg RA. 1972 Glucose-lactate inter-relations in man. N Engl J Med. 287:132–137.
30. O’Meara NM, Sturis J, Blackman JD, et al. 1993 Oscillatory insulin secretion after pancreas transplant. Diabetes. 42:855–861.[Abstract]
31. Chou HF, Berman N, Ipp E. 1994 Evidence for pancreatic pacemaker for insulin oscillations in low-frequency range. Am J Physiol. 266:R1786–R1791.
32. Tack CJ, Schefman AE, Willems JE, Thien T, Lutterman JA, Smits P. 1996 Direct vasodilator effects of physiological hyperinsulin-aemia in human skeletal muscle. Eur J Clin Invest. 26:772–778.[CrossRef][Medline]
33. Baron AD, Brechtel-Hook G, Johnson A, Hardin D. 1993 Skeletal muscle blood flow. A possible link between insulin resistance and blood pressure. Hypertension. 21:129–135.[Abstract/Free Full Text]
34. Faintrenie G, Gloen A. 1996 Lactate production by white adipocytes in relation to insulin sensitivity. Am J Physiol. 270:C1061–C1066.
35. Yarimizu K, Kawano N, Ono J, Takaki R. 1992 Periodicity of insulin secretion comprises multiple cycles with different duration in perfused rat islets. Diabetes Res Clin Pract. 17:27–32.[CrossRef][Medline]
36. Newby FD, Wilson LK, Thacker SV, DiGirolamo M. 1990 Adipocyte lactate production remains elevated during refeeding after fasting. Am J Physiol. 259:E865–E871.
37. Jansson PA, Krogstad AL, Lönnroth P. 1996 Microdialysis measurements in skin: evidence for significant lactate release in healthy humans. Am J Physiol. 271:E138–E142.
38. Thacker SV, Nickel M, DiGirolamo M. 1987 Effects of food restriction on lactate production from glucose by rat adipocytes. Am J Physiol. 253:E336–E342.
39. Gambhir KK, Agarwal VR. 1991 Red blood cell insulin receptors in health and disease. Biochem Med Metab Biol. 45:133–153.[CrossRef][Medline]
40. Hjollund E. 1991 Insulin receptor binding and action in human adipocytes. A critical approach to methods, correlations with receptor binding to other cell types, and relations between insulin binding and action. Dan Med Bull. 38:252–70.[Medline]

This article has been cited by other articles:

				T Meissner, T Otonkoski, R Feneberg, B Beinbrech, S Apostolidou, I Sipilä, F Schaefer, and E Mayatepek Exercise induced hypoglycaemic hyperinsulinism Arch. Dis. Child., March 1, 2001; 84(3): 254 - 257. [Abstract] [Full Text]

This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Feneberg, R. Articles by Schaefer, F. Search for Related Content PubMed PubMed Citation Articles by Feneberg, R. Articles by Schaefer, F.