| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Pediatrics (T.K., O.S.), Medicine (P.N., T.R.), and Radiology (E.S.) and Turku PET Centre (P.N., K.A.V., P.I., M.B., A.R., K.N., T.V., H.M., J.K.), University of Turku, 20521 Turku, Finland; and Department of Pediatrics (T.K.), Central Hospital of Seinäjoki, 60220 Seinäjoki, Finland
Address all correspondence and requests for reprints to: Teemu Kalliokoski, Department of Pediatrics, Central Hospital of Seinäjoki, Hanneksenrinne 7, 60220 Seinäjoki, Finland. E-mail: teemu.kalliokoski{at}utu.fi.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Design: To evaluate possibilities to visualize insulitis in man in vivo with positron emission tomography, we studied 12 male patients (age 26 ± 7 yr) with newly diagnosed type 1 diabetes (duration range 0–7 months) and nine age- and sex-matched healthy controls after an overnight fast using 2-[18F]fluoro-2-deoxy-D-glucose and [11C]methionine. For definition of the regions of interest, pancreas was localized with magnetic resonance imaging or computed tomography-positron emission tomography.
Results: Glucose uptake to the pancreas was markedly higher in the patients with type 1 diabetes than in the healthy controls (22.9 ± 6.4 vs. 17.8 ± 6.0 µmol/kg·min, P = 0.039). Glucose uptake to the pancreas of the patients was inversely associated with the duration of diabetes (r = –0.58; P = 0.024), so that in patients with newly diagnosed type 1 diabetes, glucose uptake was higher than in the healthy controls or patients with long duration of diabetes. Methionine uptake to the pancreas of the patients was similar as in the controls (3.7 ± 1.9 vs. 4.6 ± 2.4 µmol/kg·min, P = 0.21).
Conclusions: In patients with type 1 diabetes, glucose uptake to the pancreas is enhanced at or soon after the time of diagnosis.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Imaging of inflammation by nuclear medicine has earlier been based on nonspecific visualization due to increased vascular permeability in the acute phase (3, 4, 5). The first methods for in vivo imaging of inflammation in the pancreatic islets were developed by Signore and co-workers (6, 7), who visualized insulitis in vivo in the NOD mouse using scintigraphy and 123I-labeled IL-2. In some clinical applications, acute inflammatory lesions have been accurately localized using labeled white blood cells and scintigraphy. However, when the inflammatory lesion ages, neutrophil turnover at the site decreases as they are increasingly replaced by mononuclear cells, diminishing accuracy of the leukocyte scan. Recently, new approaches toward β-cell and inflammatory cell imaging have been introduced, but they have offered no solution to the in vivo imaging of insulitis (8, 9, 10).
Positron emission tomography (PET) makes quantitative and accurate measurement of regional metabolism possible. It also overcomes most of the physical limitations of other imaging or invasive techniques by enabling true quantification of physiological processes (11). 2-[18F]Fluoro-2-deoxy-D-glucose ([18F]FDG) has turned out to be a useful tumor-detecting agent due to the enhanced glucose utilization by the neoplastic and inflammatory cells in tumors (12). It is efficiently taken up by activated inflammatory cells and can be used to quantify their glucose uptake (13). Studies from the 1980s show that glucose uptake by proliferating thymocytes during an acute experimental inflammation can be as high as 56-fold compared with the uptake by resting thymocytes (14, 15). Increased glucose metabolism is reflected in the higher [18F]FDG uptake, suggesting that also subacute and chronic inflammations might be detectable with [18F]FDG PET.
We have earlier shown with [18F]FDG in prediabetic NOD mice that uptake of radioactivity to the islets affected with insulitis is up to 2.3 times higher than uptake to islets in the same pancreas that remained unaffected by insulitis or to the islets in the healthy BALB/c mice (16).
Methionine is an essential amino acid needed for protein synthesis, transmethylation reactions and polyamine synthesis and as a precursor for transsulfuration pathways. Amino acids like [11C]methionine (T
= 20.4 min) have been widely used as tracers for detection of malignancies with PET. The uptake of [18F]FDG to normal human pancreas is modest, but methionine is taken up by the pancreas much more intensely than by the surrounding tissues (17). In the rat, [11C]methionine accumulates in the pancreas, liver, kidney, and spleen, in decreasing order (18). [18F]FDG is a labeled glucose analog in which one hydroxyl group is replaced by fluorine-18 (T = 109.8 min). [18F]FDG is transported into the cell by the same carriers as glucose and phosphorylated to [18F]FDG-6-phosphate ([18F]FDG-6-P) by hexokinases. After phosphorylation, [18F]FDG-6-P can be incorporated into glycogen, but it cannot enter glycolysis and is irreversibly trapped in the cell.
We have now investigated adult male patients with newly diagnosed type 1 diabetes and their healthy age- and sex-matched controls using [11C]methionine and [18F]FDG PET to evaluate whether these methods enable visualization of the inflammatory process in the islets and their surrounding tissue in the pancreas.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A total of 16 adult male patients (age 27 ± 8 yr, mean ± SD) with newly diagnosed multiple autoantibody-positive type 1 diabetes participated in the study. The patients had classical symptoms of type 1 diabetes and fulfilled the World Health Organization diagnostic criteria. The control group included nine age-matched healthy men without any medication. None of the subjects had clinical symptoms of pancreatitis. The nature, purpose, and potential risks of the study were explained to all subjects as part of the written informed consent procedure. The Ethics Committee of the Turku University Hospital approved the study. PET data were not obtained in one patient due to technical problems during the study. One patient was in clinical remission requiring daily only 2 IU of short-acting insulin before breakfast. Therefore, he was excluded from the final analysis. Two additional patients were excluded from the study because normoglycemia was not achieved during the PET study. Twelve type 1 diabetic patients and nine control subjects were included in the final analysis (Table 1
).
|
All patients underwent [18F]FDG PET studies no later than 7 months after the diagnosis of type 1 diabetes. The [11C]methionine study was performed with nine diabetic patients and seven healthy controls. All subjects were studied after an overnight fast (12 h). Patients with diabetes were rendered normoglycemic before the start of the study. The regular dose of intermediate or long-acting insulin was reduced by half, and short-acting insulin was withdrawn on the morning of the study. For anatomical localization of the pancreas, either magnetic resonance imaging (MRI) scan (n = 18) or computed tomography (CT) (n = 7) was used.
PET study protocol
GE Advance Scanner (General Electric Medical Systems, Milwaukee, WI) and Discovery VCT scanner (General Electric Medical Systems) were used for imaging. Final in-plane resolutions of the reconstructed PET images were 7.4 and 4.8 mm (19), respectively. Photon attenuation was measured with a 5-min transmission scan, using a removable ring source containing 68Ge, or with CT. Either CT or MRI was used for anatomical reference. In the beginning of the PET study, a catheter was inserted in the antecubital vein for infusion of glucose, insulin, and bolus injections of [11C]methionine and [18F]FDG. Another catheter was inserted in a vein of the contralateral hand, which was warmed for sampling of arterialized venous blood. The study subject was placed in a supine position in the gantry of the PET camera. At 0 min, 280–360 MBq of [11C]methionine was injected into the antecubital vein over 2 min, and a dynamic scan (40 min; four 30-sec, three 60-sec, five 180-sec, and four 300-sec frames) of the abdominal region was obtained. Plasma radioactivity samples were collected during each time frame. [18F]FDG scanning was started more than 2 h (six times 11C half-life) after the methionine injection to allow decay of the [11C] radioactivity. Then, 250 MBq [18F]FDG was injected iv over 2 min, and a dynamic PET scan (60 min; four 30-sec, three 60-sec, five 180-sec, four 300-sec, and two 600-sec frames) was initiated at the time of the injection. Plasma radioactivity samples were collected during each time frame. During PET studies, normoglycemia was maintained using a variable-rate infusion of 20% glucose. If blood glucose concentration exceeded 7 mmol/liter during the study, a euglycemic insulin clamp was started with low insulin dose (0.2 mU/kg·min) (20).
Data analysis
Tissue time-activity curves were derived from regions of interest (ROI) placed in the body and head of the pancreas as well as in the erector spinae muscle, according to the anatomical reference of MRI or CT (Figs. 1 and 2). In studies using MRI to locate the pancreas in PET images, ROI were placed in the transaxial MRI picture with the same zoom as the PET images (Fig. 1). ROI were copied from MRI sections to corresponding planes in the [11C]methionine image to confirm their correct positioning, and then ROI were copied to the [18F]FDG planes using the same zoom and thickness. Radioactivity of [11C]methionine and [18F]FDG, measured in arterialized venous blood samples over time, was used as input function. The three-compartment model of [11C]methionine and [18F]FDG kinetics (21) and graphical analysis (22) were used to quantify the fractional uptake rates. The plasma curve was used for graphic analysis of [18F]FDG and [11C]methionine influx as described (23). The slopes of the plot (Ki) in graphical analysis are equal to the utilization rate constants of [11C]methionine and [18F]FDG. The rate of glucose uptake is calculated by multiplying Ki by the plasma glucose concentration and then dividing this by the lumped constant, which is a correction factor for the tissue. A lumped constant value of 1.0 for pancreas and 1.2 for skeletal muscle was used, as described (24). The rate of methionine uptake was calculated by multiplying Ki by plasma methionine concentration.
|
|
[11C]Methionine was prepared as described (25). The radiochemical purity exceeded 92%, and the specific radioactivity was 37 MBq/nmol (26). [18F]FDG was synthesized with a computer-controlled apparatus, according to a modified method of Hamacher et al. (27). The radiochemical purity exceeded 98%, and the specific radioactivity was 76 MBq/nmol.
MRI
MRI was done with GE Signa Horizon LX Echo Speed 1.5T (General Electric Medical Systems) with a body coil. Transverse T1-weighted field echo images with time of repetition of 124 msec and time of echo of 5 msec were obtained with the same pixel size as the PET images. Subjects fasted for at least 6 h before the MRI to minimize bowel movements.
Biochemical and radiochemical analysis
Before the PET study, blood samples were drawn after a 10- to 12-h overnight fast for measurement of islet cell autoantibodies (ICA), glucose, and insulin in serum or plasma. ICA were measured using a standard indirect immunofluorescence assay on a section of frozen human pancreas from a blood group O donor (28, 29). ICA-positive sera were diluted sequentially to find endpoint titers, and the results were expressed in Juvenile Diabetes Foundation units (JDFU). The assay was 100% sensitive and 98% specific in the International Workshop of the Standardization of the ICA (30). The detection limit of the assay is 2.5 JDFU. Autoantibodies to insulin, glutamic acid decarboxylase, and IA-2 protein were measured as described (31). Plasma samples for immunoreactive free insulin were immediately precipitated with polyethylene glycol (32). Fasting insulin and plasma C-peptide were assessed by RIA using commercial kits (Pharmacia Insulin RA kit from Pharmacia Diagnostica, Uppsala, Sweden, and Euria-C-peptide from Eurodiagnostica, Malmö, Sweden). Glycosylated hemoglobin (HbA1c) was measured with HPLC (Variant Analyzer; Bio-Rad, Hercules, CA) with reference values of 4.0–6.0%. Both fasting C-peptide and HbA1c were measured at the diagnosis of type 1 diabetes. Glucose concentration in arterialized venous blood was determined every 10 min using glucose oxidase method (Analox GM7 Analyser; Analox Instruments, Hammersmith, UK). Heart rate and blood pressure were monitored every 15 min during the PET studies.
Plasma radioactivity was measured with an automatic 
-counter (Wizard 1480 3"; PerkinElmer-Wallac, Turku, Finland) from blood samples drawn at every time frame. All data were corrected for dead time, decay, and measured photon attenuation. Radioactive [11C]methionine metabolites of plasma were analyzed at 0, 20, and 40 min after the injection. The low molecular weight fraction of [11C]methionine metabolites in plasma were separated by fast-gel filtration using Sephadex PD-10 column (Pharmacia Fine Chemicals, Uppsala, Sweden).
Statistical analysis
Statistical analysis was performed using SAS System for Windows release 8.2 (SAS, Cary, NC). Shapiro-Wilks test was performed to test normal distribution, and one-way ANOVA was used as a parametric method. All data are presented as mean ± SD. Pearson correlation test was used for analysis of correlation. Significance was set at <0.05.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). Plasma glucose concentration was within the normoglycemic range during the study sessions in both groups but slightly higher in the patients than in the healthy control subjects during the PET studies (5.9 ± 0.5 vs. 5.3 ± 0.5 mmol/liter, P < 0.01). Plasma methionine concentration was similar in the patients and in the healthy subjects (18.8 ± 5.4 vs. 22.1 ± 5.0 µmol/liter, P = 0.12). Insulin concentration was higher in the patients with diabetes than in the controls (11.1 ± 6.1 vs. 5.5 ± 4.1 mU/liter, P = 0.02).
Glucose uptake to the pancreas was higher in the patients with newly diagnosed diabetes than in the healthy subjects (22.9 ± 6.4 vs. 17.8 ± 6.0 µmol/kg·min, P = 0.039). Glucose uptake associated inversely with the duration of diabetes (r = –0.58; P = 0.024) (Fig. 3A) and was markedly higher in the patients who were studied shortly after the diagnosis of type 1 diabetes than in the patients examined a few months later. It is of interest that the highest glucose uptake to the pancreas (38.9 µmol/kg·min) was found in the patient who had type 1 diabetes diagnosed at a very early stage in normal routine checkup at the age of 40 yr. He had abnormal oral glucose tolerance test, and therefore, insulin medication was started. He was positive for three of four type 1 diabetes-specific autoantibodies studied (Table 1). His PET study was done 7 months after the initiation of daily sc insulin. At the time of scanning, he needed only a daily dose of 2 IU of short-acting insulin at breakfast and was considered to be in remission. Therefore, his pancreatic glucose uptake has not been included in the calculation of the mean values.
|
). Pancreatic glucose uptake was not associated with plasma insulin concentrations, daily sc insulin dose, or autoantibody titers. Pancreatic glucose uptake values and fasting C-peptide concentrations showed rather good correlation (r = 0.62; P < 0.05). However, pancreatic glucose uptake and HbA1c values were not associated (r = –0.35; P = 0.26). In contrast to glucose, methionine uptake to the pancreas and skeletal muscle was similar in the patients and in the controls (3.7 ± 1.9 vs. 4.6 ± 2.4 µmol/kg·min and 0.38 ± 0.32 vs. 0.32 ± 0.09 µmol/kg·min, respectively; P for both was nonsignificant). Methionine uptake was not associated with pancreatic glucose uptake or duration of diabetes.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We have previously shown that the mean [18F]FDG uptake by the islets of Langerhans is elevated in prediabetic and diabetic NOD mice with insulitis. The extent of insulitis varied markedly in the NOD mice of all ages, so that in pancreas sections, only some of the islets were infiltrated. All NOD mice with insulitis had also regions in the pancreas with entirely intact islets, often in groups of several islets in the same intact tissue section. NOD islets with no inflammation showed similar [18F]FDG uptake ratios as the healthy BALB/c islets and the surrounding exocrine pancreas (16).
Our current findings suggest that glucose uptake to the human pancreas is increased at the time of diagnosis of type 1 diabetes (Fig. 3A). After the diagnosis and onset of insulin therapy, the number of the insulin-producing β-cells as well as the activity of the islet-infiltrating lymphocytes probably gradually decreases, leading to a return of the uptake of glucose in the pancreas to the baseline values within 3–6 months of the diagnosis of type 1 diabetes.
Previous studies suggest that amino acids are taken up by inflammatory cells less actively than [18F]FDG, probably because protein metabolism in inflammatory cells is less active than glucose metabolism (33). This may to some extent explain why in this study methionine uptake to the pancreas was slightly lower in the patients with type 1 diabetes than in the controls. It is also possible that the patients with diabetes may have been in a mild catabolic state due to fasting, which could also explain the decrease in the methionine uptake to the pancreas.
Altogether, our results obtained with the glucose and methionine tracers are in accord with the findings of a local inflammation, characterized typically by higher uptake of glucose than of methionine or other amino acids. Although insulitis and other inflammatory processes like pancreatitis cannot be absolutely differentiated from each other based on the [18F]FDG and [11C]methionine uptake only (34), the fact that our patients had no history or clinical evidence of pancreatitis suggests that the patients had insulitis, possibly in combination with other type 1 diabetes-associated metabolic and circulatory findings. Interestingly, the time period during which glucose uptake showed the strongest increase overlapped quite well with the time of the expected peak activity of insulitis.
It is of note that in this study, glucose uptake correlated rather well with the markers of severity of the metabolic derangement at the time of diagnosis of type 1 diabetes. Pancreatic glucose uptake showed positive association with serum fasted C-peptide values, suggesting that additional unknown factors may have modified the events that precede development of overt type 1 diabetes. We thus propose that the overall rate and extent of accumulation of inflammatory cells and velocity of β-cell destruction may markedly differ among individuals. The severity of the metabolic derangement at diagnosis may reflect the immensity of the preceding inflammatory destruction. Furthermore, there may be some delay in the manifestation of the metabolic imbalance with respect to the causal inflammatory reaction. Finally, several potential confounding factors beyond destructive autoimmunity may regulate glucose control at this stage, so that glucose itself may be a poor marker of the overall functional status of the remaining β-cells.
In animal experiments, [18F]FDG uptake to inflammatory lesions is markedly reduced during hyperglycemia, whereas uptake to tumors remains unaltered (35). In our study, the diabetic subjects had slightly higher mean plasma glucose concentration than the healthy controls. Thus, the difference in glucose uptake between the diabetic and control subjects might have been slightly more pronounced if both groups had had strictly identical glucose concentration during the PET study. However, multiple regression analysis showed that in this study, plasma insulin had no influence on glucose uptake to the pancreas. In simple regression analysis, the time since diagnosis of type 1 diabetes correlated nicely with glucose uptake to the pancreas.
In conclusion, glucose uptake to the pancreas of patients with newly diagnosed type 1 diabetes is higher than that of healthy age- and sex-matched control subjects. Despite the fact that the number of subjects studied was small, the data suggest that [18F]FDG-based PET technology can possibly be used for rough estimation of the extent of insulitis in man. To confirm the validity of this imaging approach, repeated studies of subjects with increased risk of developing type 1 diabetes are needed to evaluate the method used. Use of the recently introduced CT-PET, which further improves anatomical localization of the pancreas in the images, in combination with new, more specific signaling molecules, may lead to discovery of new imaging technologies that may markedly improve early identification of subjects with increased risk of progression to overt type 1 diabetes.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Summary: T.K., P.N., P.I., M.B., E.S., A.R., T.V., H.M., T.R., and O.S. have nothing to declare. K.N. has received consulting fees from IAEA of less than $10,000. K.A.V. and J.K. have received lecture fees of less than $10,000.
First Published Online February 19, 2008
Abbreviations: CT, Computed tomography; FDG, fluoro-2-deoxy-D-glucose; HbA1c, glycosylated hemoglobin; ICA, islet cell autoantibodies; JDFU, Juvenile Diabetes Foundation units; MRI, magnetic resonance imaging; NOD, nonobese diabetic; PET, positron emission tomography; ROI, regions of interest.
Received November 17, 2006.
Accepted February 11, 2008.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
). The studies were performed using a PET scan and [18F]FDG. One of the patients was considered to be in remission (x) (for details, see Subjects and Methods). Glucose uptake of control subjects has been shown in the same scale as a group (
). Glucose uptake (millimoles per kilogram per minute) to the pancreas shows good correlation with the duration of diabetes (r = –0.58; P value = 0.024). Glucose uptake to the muscle differs clearly from the uptake to the pancreas.