Weight Loss After Bariatric Surgery Reverses Insulin-Induced Increases in Brain Glucose Metabolism of the Morbidly Obese

  1. Pirjo Nuutila1,4
  1. 1Turku PET Centre, University of Turku and Turku University Hospital, Turku, Finland
  2. 2Department of Radiology, University of Turku and Turku University Hospital, Turku, Finland
  3. 3Department of Surgery, University of Turku and Turku University Hospital, Turku, Finland
  4. 4Department of Medicine, Turku University Hospital, Turku, Finland
  5. 5Department of Biomedical Engineering and Computational Science, School of Science, Aalto University, Espoo, Finland
  6. 6Brain Research Unit, O.V. Lounasmaa Laboratory, School of Science, Aalto University, Espoo, Finland
  1. Corresponding author: Pirjo Nuutila, pirjo.nuutila{at}utu.fi.

Abstract

Obesity and insulin resistance are associated with altered brain glucose metabolism. Here, we studied brain glucose metabolism in 22 morbidly obese patients before and 6 months after bariatric surgery. Seven healthy subjects served as control subjects. Brain glucose metabolism was measured twice per imaging session: with and without insulin stimulation (hyperinsulinemic-euglycemic clamp) using [18F]fluorodeoxyglucose scanning. We found that during fasting, brain glucose metabolism was not different between groups. However, the hyperinsulinemic clamp increased brain glucose metabolism in a widespread manner in the obese but not control subjects, and brain glucose metabolism was significantly higher during clamp in obese than in control subjects. After follow-up, 6 months postoperatively, the increase in glucose metabolism was no longer observed, and this attenuation was coupled with improved peripheral insulin sensitivity after weight loss. We conclude that obesity is associated with increased insulin-stimulated glucose metabolism in the brain and that this abnormality can be reversed by bariatric surgery.

Insulin increases glucose uptake, storage, and usage in tissues, such as skeletal muscle and liver (1). Evidence from animal studies shows that insulin enters brain tissue via a saturable receptor-mediated transport (2), and receptors for insulin are widely expressed in the central nervous system (CNS) (1). Thus, insulin has the potential to act in the CNS, but in humans insulin-mediated effects on glucose metabolism in the CNS are thought to be minimal (3,4). Correspondingly, hyperinsulinemia does not increase brain glucose metabolism in healthy subjects (3,57).

We previously found that brain glucose metabolism increases during hyperinsulinemia in subjects with obesity-related impaired glucose tolerance (IGT) and peripheral insulin resistance compared with healthy lean subjects (6). Weight gain causes insulin resistance in peripheral tissues (8), and brain insulin resistance may develop parallel to peripheral insulin resistance and plasma hyperinsulinemia (9). However, altered insulin responses seem to differ from those observed in other tissues (6,7). Weight loss decreases peripheral insulin resistance (8,10), but the effect of weight loss on brain glucose metabolism is unknown.

In the current study, we sought to determine whether rapid weight loss after bariatric surgery and a subsequent decrease in peripheral insulin resistance would result in decreased brain glucose metabolism during insulin stimulation. We used direct measurements of brain glucose uptake with [18F]fluorodeoxyglucose (FDG)–positron emission tomography (PET) during euglycemic-hyperinsulinemic clamp and during fasting condition. Subjects were morbidly obese patients who underwent bariatric surgery and healthy control subjects. Preoperatively, we aimed to replicate the previous findings of increased brain glucose metabolism in morbidly obese patients in contrast to healthy control subjects.

RESEARCH DESIGN AND METHODS

For the current study, 22 obese patients were recruited as a part of a larger randomized, controlled clinical study evaluating the effect of sleeve gastrectomy and Roux-en-Y gastric bypass (Sleevepass study) (11). Four had type 2 diabetes, 4 had IGT or impaired fasting glucose (IFG), and 14 had normal results in oral glucose tolerance test (OGTT). The four diabetic subjects used metformin (from 1 to 3 g daily). Subjects with insulin treatment or fasting glucose >7 mmol/L were excluded. Other inclusion criteria have previously been described (11). Ten healthy subjects were recruited via an advertisement in local newspapers. Inclusion criteria were normal glucose tolerance test values, fasting plasma glucose <6.1 mmol/L, and chronic disease (Table 1 and 2). Demographic information for the groups is provided in Tables 12.

TABLE 1

Characteristics of the subjects in preoperative comparison of patients and control subjects

TABLE 2

Comparison of patients in preoperative and postoperative states

In this prospective study, obese patients were studied at two time points: before and 6 months after bariatric surgery. Healthy subjects were studied once. The baseline studies were carried out before the patient started a 4-week very-low-energy diet before the surgery. The PET studies were performed both in a fasting state and during euglycemic-hyperinsulinemic clamp (12), on separate days, <2 weeks apart. Strenuous physical activity was prohibited from the preceding evening. Antidiabetes treatment was withheld 24 h before the metabolic studies. Prior to inclusion, each participant gave written consent. The study protocol was approved by the ethics committee of the Hospital District of the Southwestern Finland and conducted in accordance with the Declaration of Helsinki.

PET study protocol and data acquisition.

The studies were performed after a 12-h fast using the GE Advance PET camera (General Electric Medical Systems, Milwaukee, WI). The euglycemic-hyperinsulinemic clamp technique was used as previously described (6,12). [18F]FDG-PET (187 ± 9 MBq) was injected intravenously over 15 s, and radioactivity in brain was followed thereafter for 40 min (4 × 30, 3 × 60, and 7 × 300 s frames).

[18F]FDG was synthesized with a computer-controlled apparatus (18). All data were corrected for dead time, decay, and measured photon attenuation and reconstructed using a Hann filter with a cutoff frequency of 0.5 and an mannosylrentinyl phosphate reconstruction method (13). Arterialized blood samples were drawn during the scan and analyzed for radioactivity concentration in plasma using an automatic γ counter (6,12). In the clamp study, plasma glucose and serum insulin concentrations were taken at baseline and every 5, 30, and 60 min, respectively.

In the preoperative setting, scans of morbidly obese subjects were compared with the scans of the control group. In the preoperative versus postoperative comparison, only subjects who participated in all four scans were included in the analysis (n = 17). We lost two control scans as a result of unsuccessful preprocessing, and one control subject was excluded because of an elevated BMI. In the postoperative setting, four subjects did not undergo bariatric surgery and one subject missed one scan.

Quantification of brain glucose metabolism.

The influx constant (Ki) was calculated for each voxel separately using the linear Gjedde-Patlak plot with arterial plasma input function, with a linear phase start time of 20 min. Glucose uptake estimate of the cerebral metabolic rate (CMR) (CMRglu [μmol × 100 g − 1 × min−1]) was then calculated at voxel level as follows: CMRglu = Ki × Cp/LC, where Cp is the plasma glucose concentration and LC is the lumped constant (which was set at 0.80).

Summed PET images were normalized spatially to a ligand-specific template in to the Montreal Neurological Institute (MNI) space (MNI International Consortium for Brain Mapping) using SPM8 (www.fil.ion.ucl.ac.uk/spm/) running on Matlab for Windows (version 7.7.0; Math Works, Natick, MA). Normalization parameters were subsequently applied to corresponding parametric glucose metabolism images. Parametric images were smoothed at 10 mm full width at half-maximum.

Statistical analysis.

Preoperative data were analyzed with a 2 (fast, clamp) × 2 (patient, control) ANOVA and the follow-up data for patients with a 2 (fast, clamp) × 2 (preoperative, postoperative) ANOVA. The statistical threshold in SPM analysis was set at voxel level false discovery rate (FDR) corrected P < 0.05, uncorrected at cluster level. Lowering of the statistical threshold did not yield statistically significant results. Further statistical analyses were done using SPSS, version 18.0, for Windows (SPSS, Chicago, IL). A more detailed complementary analysis included quantification of brain glucose metabolism by defining the LC in a region-specific manner instead of using a fixed lumped constant as in the analysis mentioned above and region of interest–based analysis (see Supplementary Data).

RESULTS

Effects of insulin on brain glucose metabolism in control subjects and patients prior to the surgery.

In the morbidly obese patients, we found a global increase in brain [18F]FDG-PET uptake during clamping compared with fasting (Fig. 1B). This difference was not statistically significant in the control group. The regional maximum of the increased glucose metabolism was the region of the right caudate nucleus (Fig. 1A). There were no statistically significant differences between the groups in the fasting condition. Thus, the euglycemic-hyperinsulinemic clamp increased brain glucose metabolism only in the morbidly obese subjects.

FIG. 1.

Insulin-induced changes of brain glucose metabolism. A: Preoperative (Preop) clamp scan comparison of morbidly obese patients vs. control group. The bar represents T values. MNI coordinates −9, 2, and 10 are chosen so that regional maximum in right caudate nucleus can be seen. The data were thresholded at P < 0.05 FDR corrected, T > 3.02. B: Preoperative CMRglu values of control subjects (n = 7) and morbidly obese patients (n = 22). Error bars represent the SEM. *Significantly increased CMRglu values during clamp scans in comparison with corresponding fasting scans. C: Preoperative comparison of CMRglu of morbidly obese patients for clamp vs. fast. P value 0.05 voxel level uncorrected; P < 0.05 cluster level FDR corrected. The bar represents T values. MNI coordinates: −16, 8, and 28. The data were thresholded at P < 0.05 FDR corrected, T > 3.52. D: Postoperative (Postop) comparison of CMRglu of morbidly obese patients for clamp vs. fast. P value 0.05 voxel level uncorrected; P < 0.05 cluster level FDR corrected. The bar represents T values. MNI coordinates −16, 8, and 28. E: Pre- and postoperative CMRglu values of morbidly obese patients (n = 17). Error bars represent the standard SEM. *Significantly increased CMRglu values during clamp scans in comparison with corresponding fasting scans. CER-A, anterior cerebellum; CER-P, posterior cerebellum; FRO, frontal lobe; LIMB, limbic lobe; MID, midbrain; OCC, occipital lobe; PAR, parietal lobe; TEMP, temporal lobe.

Effects of bariatric surgery on brain glucose metabolism.

Elevation of glucose metabolism during clamp was markedly diminished after bariatric surgery (Fig. 1D). The postoperative regional CMRglu values in the morbidly obese patients were similar to the values of control subjects in a preoperative setting (Fig. 1E). Bariatric surgery resulted in rapid weight loss and improvement in glycemic control and whole-body insulin sensitivity (M values in Table 2). One of four in both diabetic and IGT/IFG patients was in diabetic remission postoperatively, and for the rest the diabetes status, defined by OGTT and American Diabetes Association criteria, was unchanged.

DISCUSSION

Our results indicate that insulin-induced stimulation during euglycemic-hyperinsulinemic clamp scans of the morbidly obese is reversible through weight loss and consequent reduced peripheral insulin resistance 6 months after bariatric surgery (Fig. 1C and D). Preoperatively, we detected elevations in glucose metabolism in morbidly obese subjects compared with normal-weight subjects in striatum and cerebellum during euglycemic-hyperinsulinemic clamp but not in the fasting condition (see Supplementary Data), which partly replicates our previous findings (6). These observations are not explained merely by increases in plasma levels of insulin during PET scans (10) (Tables 1 and 2).

Insulin effects on brain glucose metabolism.

In individuals with normal weight, insulin has little effect on brain glucose metabolism (3,57). Nevertheless, normal-weight individuals may also have minor insulin-induced elevations in brain glucose metabolism that might not be detected as statistically significant in limited sample sizes. Obesity, however, seems to be associated with increased sensitivity to insulin in brain tissue (6).

How do we explain the increased metabolic sensitivity of brain to insulin? The ratio of insulin in cerebrospinal fluid to that in plasma is decreased in obesity (14), and obese individuals also show diminished catabolic responses after intranasal insulin administration (15). These findings could be a sign of brain insulin resistance (9,15,16), which could be caused by impeded insulin transport through the blood-brain barrier (BBB) or a weakened neuronal responses to insulin. On the other hand, in the fasting state glucose metabolism is not decreased in overweight subjects in comparison with a control group (6). Chronically elevated plasma levels of insulin could also increase the delivery of insulin across the BBB. If the cells in the CNS adapt to this change by increasing the rate of insulin clearance or internalization, then decreased cerebrospinal fluid–to–plasma insulin ratio of the obese individuals may reflect greater insulin usage. This would also explain the diminished catabolic responses to intranasal insulin, as the given amounts of insulin would presumably be too small for cells accustomed to increased insulin levels.

Taken together, by increasing peripheral insulin resistance the brain tissue's glucose metabolism is increasingly stimulated independent of changes in plasma glucose levels. Molecular mechanisms for the changes are currently unclear. Clinically, intensifying the pharmacological therapy for glycemic control keeps the plasma glucose levels at bay and prevents glucose neurotoxicity in peripheral organs (4). The opposite might be true for the brain tissue (17).

Limitations in the study.

The radioactive decay of the [18F]FDG-PET tracer can occur at any level between BBB and intracellular space of neurons and glial cells. The clamp as an experimental tool also has some limitations (18). The subjects of this study are heterogeneous in terms of glucose metabolism and differ in terms of endogenous peripheral insulin resistance. Excluding diabetic/IGT subjects from SPM analysis or grouping subjects according to diabetes status did not affect the results, and the effects of medication and diabetes status are minimal. Hyperglycemia during fasting condition might diminish the [18F]FDG uptake to brain tissue. However, obese subjects were not hyperglycemic and the scans and clamping conditions were similar in both groups (Tables 1 and 2). Postoperative comparison necessitated participation in four scans, and because of this the study had a relatively large drop-out rate.

Conclusions.

The results of our study support the evidence showing that insulin stimulates brain glucose metabolism of the morbidly obese. We report, for the first time, that weight loss after bariatric surgery reverses increased insulin stimulation. This supports that bariatric surgery not only results in weight loss but also improves overall metabolic health (10)—this might include the brain.

ACKNOWLEDGMENTS

The study was conducted within the Finnish Center of Excellence in Molecular Imaging in Cardiovascular and Metabolic Research. It was supported by Academy of Finland (grants 256147 and 251125 to L.N.), University of Turku, Turku University Hospital, Åbo Akademi University, Finnish Diabetes Foundation, an Aivo-Aalto grant of Aalto University, and Sigrid Juselius fund.

M.S. has participated in a surgical congress with Pfizer's support and held lectures at scientific meetings organized by Roche and Merck Sharp & Dohme. M.H. has participated in a surgical congress with the support of Covidien, Stryker, and Leiras. No other potential conflicts of interest relevant to this article were reported.

J.J.T., H.K.K., and J.H. analyzed data and wrote the manuscript. J.C.H. conducted the clinical PET studies and wrote the manuscript. M.B. analyzed the compartmental data analyses and wrote the manuscript. M.H. and J.O. performed patient recruitment and screening and wrote the manuscript. M.S. and P.S. performed euglycemic clamps and wrote the manuscript. N.S. synthesized radioligands and wrote the manuscript. L.N. analyzed data and wrote the manuscript. P.N. planned the experiments and wrote the manuscript. J.J.T. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors thank the staff of the Turku PET Centre for performing the PET imaging. The authors also thank Robert M. Badeau, with the Turku University Graduate School Language Centre, for contributing to the grammatical refinement of the manuscript. The authors also thank the researchers and research nurse Mia Koutu, of Turku PET Centre, for data collection.

Footnotes

  • Received October 22, 2012.
  • Accepted March 11, 2013.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

REFERENCES

| Table of Contents

This Article

  1. Diabetes vol. 62 no. 8 2747-2751
  1. Supplementary Data
  2. All Versions of this Article:
    1. db12-1460v1
    2. 62/8/2747 most recent