| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
-cell mass in normal and -cell-reduced lean and obese Göttingen minipigs in vivo
1Department of Pharmacology Research I and 2Department of Pharmacology Research III, Novo Nordisk A/S, Maaloev; 3Department of Assay and Cell Technology, Novo Nordisk A/S, Bagsvaerd; 4Discovery Management, Novo Nordisk A/S, Bagsvaerd; 5Medical Department C, Aarhus University Hospital, Aarhus; and 6Department of Pharmacology Research IV, Novo Nordisk A/S, Maaloev, Denmark
Submitted 6 June 2005 ; accepted in final form 1 November 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-cell mass in humans are those based on measurement of insulin levels or appearance rates in the circulation. In several animal models, these estimators have been validated against -cell mass in lean animals. However, as many diabetic humans are obese, a correlation between in vivo tests and -cell mass must be evaluated over a range of body weights to include different levels of insulin sensitivity. For this purpose, obese (n = 10) and lean (n = 25) Göttingen minipigs were studied. -Cell mass had been reduced (n = 16 lean, n = 5 obese) with a combination of nicotinamide (67 mg/kg) and streptozotocin (125 mg/kg), acute insulin response (AIR) to intravenous glucose and/or arginine was tested, pulsatile insulin secretion was evaluated by deconvolution (n = 30), and -cell mass was determined histologically. AIR to 0.3 (r2 = 0.4502, P < 0.0001) or 0.6 g/kg glucose (r2 = 0.6806, P < 0.0001), 67 mg/kg arginine (r2 = 0.5730, P < 0.001), and maximum insulin concentration (r2 = 0.7726, P < 0.0001) were all correlated to -cell mass when evaluated across study groups, and regression lines were not different between lean and obese groups except for AIR to 0.3 g/kg glucose. Baseline pulse mass was not significantly correlated to -cell mass across the study groups (r2 = 0.1036, NS), whereas entrained pulse mass did show a correlation across groups (r2 = 0.4049, P < 0.001). This study supports the use of in vivo tests of insulin responses to evaluate -cell mass over a range of body weights in the minipig. Extensive stimulation of insulin secretion by a combination of glucose and arginine seems to give the best correlation to -cell mass. animal model; streptozotocin; arginine; pulsatile insulin secretion
-CELL MASS is an intrinsic aspect of diabetes in humans (9, 11, 23, 34, 43, 52, 53). Measurement of glucose levels is used both in the diagnosis of the disease and to evaluate effects of treatment of diabetes (1, 5), but because glucose is tightly controlled by several mechanisms, increased levels of glucose are not seen before a considerable proportion of -cell function and mass is lost (4, 9). Similarly, reductions of -cell mass by 30–50% result in only mild changes in glycemia and insulin levels in humans (21), rats (35, 52), and pigs (33), and severe changes are seen only after more dramatic reductions in -cell mass in humans (13, 23, 42), rats (8), pigs (22, 29, 33), and baboons (36). In both humans (45) and pigs (38), there is a strong relation between islet mass used in transplantation studies and metabolic control obtained as evaluated by glucose levels and insulin secretion, a finding that underlines the importance of -cell mass in maintenance of normal glucose tolerance and insulin secretion.
The possibility of studying the dynamics of -cell mass in humans could supply important information about the development of diabetes as well as effects of pharmacological treatment of the disease as a supplement to measurement of the highly regulated glucose levels.
Nuclear imaging has been validated for this purpose in mice (39), whereas at present it is not possible to use this technology in humans. Therefore, the best available estimates are measurements of insulin levels or appearance rates in the circulation as an indication of -cell mass and/or function in vivo. Several methods are available for such evaluations in humans, all relying on measurements of insulin levels or appearance rates in the circulation, including stimulation with glucose (6, 12) and arginine (50). The use of measurements to estimate -cell mass based on extensive stimulation of insulin secretion using glucose and/or arginine has been validated in primates (36) and pigs (29) and based on detailed evaluation of pulsatile insulin secretion in pigs (22, 26). Whether either of the two methods shows a superior correlation to -cell mass remains to be determined. Furthermore, these validations have been performed only in lean animals, and since many diabetic humans are obese, it is of importance to evaluate whether this correlation is also found across a wider range of body weights in the same strain of animals. We have previously evaluated measurements of insulin responses to stimulations of various types in obese animals in detail (27) but have not related these to -cell mass compared with what is seen in lean animals.
Therefore, the aim of the present study was to evaluate correlations between -cell mass and different methods for evaluation of insulin levels or appearance rates in the circulation, including detailed analysis of pulsatile insulin secretion, over a range of body weights. The Göttingen minipig is a relevant model in the field of diabetes due to its well-described biology with respect to glucose metabolism in normal and diabetic animals (22, 30, 31) and the rapid dynamics of endogenous insulin facilitating evaluation of pulsatile insulin secretion (25). Because there is a close relationship between insulin sensitivity and insulin levels and responses in humans (2, 19, 20, 32, 47), it is important to include different levels of insulin sensitivity when measurements of insulin responses in vivo and -cell mass are compared. We (28) have previously shown a negative correlation between body weight and insulin sensitivity in Göttingen minipigs, so both lean and obese animals were included in the study to cover a wide range of body weights and insulin sensitivity. Similarly, to include a wide range of -cell masses, both lean and obese animals that had had their -cell mass reduced by a combination of nicotinamide (NIA) and streptozotocin (STZ) (31) were included in the study.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The present study was a retrospective analysis summarizing data from two different experiments performed in our laboratory. The major objective was to combine information obtained from these two studies of different aspects of insulin appearance in the circulation and -cell mass. Because data were analyzed retrospectively, group sizes have been dependent on the available data and, therefore, differ across the included study groups. Inclusion criteria for the study was that each animal should have completed at least two in vivo tests, one of these focusing on pulsatile insulin secretion and one focusing on insulin response to glucose and/or arginine. The only deviation from this was five normal animals, which were included for histological assessment of -cell mass only (see below).
Animals
Source and housing. All animals included in the study were adult male Göttingen minipigs 11–14 mo of age. Animals had been obtained from the barrier unit at Ellegaard Göttingen Minipigs ApS (Dalmose, Denmark). All animals had been housed in single pens under controlled conditions (temperature was kept between 18 and 22°C, relative air humidity was 30–70% with 4 air changes/h) with a 12:12-h light-dark cycle and allowed free access to water.
All pigs had been studied 
2 wk after surgical implantation of central venous catheters (see below) and after careful training in all experimental procedures before start of experiments.
Principles of laboratory animal care were followed and the type of study was approved by the Animal Experiments Inspectorate, Ministry of Justice, Denmark.
In total, 35 animals were included in the study.
Diet.
Three different feeding regimens had been used for the animals included in the study. To obtain a wide range of body weights, 25 animals that had been kept on the standard feeding regimen to maintain lean phenotype [140 g of SDS minipig diet (SDS, Essex, England) and 240 g of a commercial swine fodder ("Svinefoder 22" or "Antonio", Østsjællands Andel, Karise, Denmark) fed twice daily] were included. Furthermore, 10 obese animals were included, five of which had been fed SDS ad libitum from weaning until 12 mo of age. From 12 mo of age, these animals had been fed 
300 g of a high-fat diet (20% fat) (Danish pastry, Ganløse, Denmark) plus 240 g of commercial swine fodder twice daily for 7 mo, as previously described (27), to induce obesity. To combine reduced -cell mass and obesity, another five of the obese animals included in the study had had their -cell mass reduced at age 12 mo (see below) followed by high-fat feeding (240 g commercial swine fodder and 300 g high-fat diet twice daily) for 8 mo.
Data obtained.
Data on -cell mass and body weight were collected from all 35 included animals.
For five of the included lean animals, these were the only data obtained. The reason for including these animals was to ensure a robust measurement of -cell mass in lean animals.
Data from three types of in vivo tests were included in the study: insulin secretion in response to glucose and/or arginine, basal pulsatile insulin secretion, and entrained pulsatile insulin secretion (see below for methodology applied). Because of the inclusion criteria, it was not possible to include data from all in vivo tests from all animals, so n is different for each test (see exact number of animals for each test in the descriptions below).
For some of the animals included in this study, data from in vivo tests have been published previously. Detailed information on this is found below.
Surgical Implantation of Central Venous Catheters
All animals from which data from in vivo tests were included had had two central venous catheters (Certo 455; B. Braun Melsungen, Melsungen, Germany) inserted surgically under general anesthesia as described previously (31). Postsurgical analgesia had been maintained by injection of 0.03 mg/kg buprenorfine (Anorfin; GEA, Frederiksberg, Denmark) and carprofen 4 mg/kg (Rimadyl vet. 50 mg/ml; Pfizer, Ballerup, Denmark) intramuscularly before the end of anesthesia and for 3 days postsurgery by injection of 4 mg/kg carporfen once daily intramuscularly. At the start of the study period, all animals had recovered fully from the surgical procedure as evaluated by normal behavior and eating patterns.
Reduction of -Cell Mass
To obtain a wide range of -cell mass, we included 21 animals which had had their -cell mass reduced by intravenous administration of NIA (67 mg/kg, Sigma N-3376; Sigma-Aldrich, Steinheim, Germany) followed by STZ (125 mg/kg, Sigma S-0130) after an 18-h overnight fast, as previously described (31). With respect to reduction of -cell mass, all animals had been offered SDS fodder 2 h after treatment and had been observed frequently during the first 48 h after administration of NIA and STZ, including regular monitoring of blood glucose to detect and treat episodes of hypoglycaemia due to sudden hyperinsulinemia caused by necrosis of -cells.
Five of the 21 -cell-reduced animals had also been fed a high-fat diet for 7–8 mo to induce obesity (obese-STZ animals).
Dynamics of Baseline Insulin Secretion
Baseline pulsatile insulin secretion had been studied in fasted (18 h) conscious animals (n = 4 lean, 12 lean-STZ, 5 obese, and 5 obese-STZ) as previously described (25). In short, blood samples (0.8 ml) had been obtained from a central venous catheter every minute for 40 min. Before each blood sample was collected, 1.5 ml of blood, corresponding to the catheter dead space, were withdrawn and returned aseptically after each sample. Catheters were flushed with 0.8 ml of sterile saline (0.9% SAD, Copenhagen, Denmark) after each blood sample.
Detailed analysis of baseline pulsatile insulin secretion in these animals (except from 2 lean-STZ) has been published elsewhere (26, 27).
Dynamics of Entrained Insulin Secretion
Entrainment of pulsatile insulin secretion had been studied in fasted (18 h) conscious animals (n = 4 lean, 10 lean-STZ, 5 obese, and 5 obese-STZ) as previously described (25). In short, blood samples (0.8 ml) had been obtained from a central venous catheter every minute for 40 min as described above. Every 10th min, starting at 0 min, a bolus of glucose (4 mg·kg–1·min–1 glucose, 200 mg/ml, SAD) had been infused over 1 min via the other central venous catheter.
Detailed analysis of entrained pulsatile insulin secretion in these animals has been published elsewhere (26, 27).
Glucose- and Arginine-Stimulated Insulin Responses
All evaluation of glucose- and arginine-stimulated insulin responses had been performed in fasted (18 h) conscious animals as previously described (29). All compounds had been given intravenously through a central venous catheter, and blood samples had been obtained from the other central venous catheter.
Glucose-stimulated insulin response had been evaluated by administration of glucose (0.3 g/kg, 500 g/l, SAD, given as a bolus over 30 s; n = 4 lean, 15 lean-STZ, 5 obese, and 5 obese-STZ). Furthermore, in some animals (n = 2 lean, 9 lean-STZ, 5 obese, and 4 obese-STZ), the test had been extended with evaluation of insulin secretion in response to 0.6 g/kg glucose (given as a bolus over 30 s 60 min after the 0.3 glucose bolus) followed by an infusion of glucose (2 g·kg–1·h–1, 200 g/l, SAD) for 40 min to maintain hyperglycemia at 20–30 mM. In these animals, a bolus of arginine (67 mg/kg given over 30 s) had been given 30 min after the 0.6 g/kg glucose bolus.
Blood samples had been obtained at –15, –10, –5, 1, 3, 5, 7, and 10 min relative to each of the bolus injections; 24 blood samples were obtained in total.
Details of insulin responses in relation to -cell mass have been published for some of these animals [n = 4 lean, 7 lean-STZ, 5 obese, and 5 obese-STZ (27, 29)].
Handling and Analysis of Blood Samples
Blood samples had immediately been transferred to vials containing EDTA (1.6 mg/ml final concentration) and aprotinin 500 KIU/ml full blood (Trasylol, 10,000 KIU/ml; Bayer, Lyngby, Denmark) and kept on ice until centrifugation. Samples had been centrifuged (4°C, 10 min, 3,500 rpm) and plasma separated and stored at –20°C until analysis. Plasma glucose had been analyzed using the immobilized glucose oxidase method, 10 µl of plasma in 0.5 ml of buffer (EBIO plus autoanalyzer and solution; Eppendorf, Hamburg, Germany). Plasma insulin had been analyzed in a two-site immunometric assay with monoclonal antibodies as catching and detecting antibodies [catching antibody HUI-018 raised against the A-chain of human insulin; detecting antibody OXI-005 raised against the B-chain of bovine insulin (3)] and using purified porcine insulin for calibration of the assay as previously described (31). The minimal detectable concentration was 3.2 pmol/l, the upper limit was 1,200 pmol/l (no sample dilution), and the inter- and intra-assay coefficients of variation at three concentration levels were 15.3 and 3.2% (at 342 pM), 9.9 and 7.6% (at 235 pM), and 14.6 and 4.4% (at 87 pM). Recovery at high, medium, and low concentration levels was 97.1, 97.9, and 101%, respectively.
Histological Examination of Pancreas
Histological examination had been performed in all animals. After euthanasia with pentobarbitone (20 ml per animal, 200 mg/ml; Pharmacy of the Royal Veterinary and Agricultural University, Copenhagen, Denmark), the pancreas had been isolated in toto and fixed in paraformaldehyde (Bie & Berntsen) for 24 h. In all animals, the pancreas had been obtained at least 4 days after the last in vivo test to ensure recovery of insulin stores. To compensate for a possible heterogeneous distribution of the islets in the pancreas, as described for dog pancreas (24), an unbiased fractionator sampling scheme had been used. In short, pancreata were embedded in 3% agar solution (Meco-Benzon cat. no. 303289, Copenhagen, Denmark) and physically sectioned. The pancreas was cut into 3-mm slices, and every fifth tissue slice, starting at slice 1, 2, 3, 4, or 5, determined from a table of random numbers, was retained for sectioning, so that, in total, one-fifth of the whole pancreas was included in 9–10 slices. All of the tissue included in these 9–10 slices was then cut into 80–90 cubes of roughly equal size, still representing 20% of the total pancreas. These cubes were arranged on a line according to size, as practiced in the smooth fractionator method, with the largest cubes in the middle and the smallest cubes on the ends (7, 15), a recent modification of the classical fractionator scheme (16). Every eighth cube, starting at cube 1, 2, 3 ... 8, determined from a table of random numbers, was retained and placed randomly oriented in cassettes for dehydration and paraffin infiltration in a tissue processor (Leica TP 1050, Copenhagen, Denmark). On the basis of the described sampling schedule, the resulting paraffin block from each pancreas contained 10–12 cubes of tissue representing different anatomic areas selected randomly. This tissue sample equals 2.5% of the total pancreas mass from each animal. From each paraffin block, 2–3 sections (representing all 10–12 cubes of tissue) 3 µm thick, 250 µm apart in depth, were cut on a Leica RM 2165 microtome.
The deparaffinized sections were stained for insulin to visualize -cells. Furthermore, sections were counterstained with Mayer’s hematoxylin. -Cell mass was evaluated stereologically in an Olympus BX-50 microscope (Olympus, Copenhagen, Denmark) with video camera and monitor at a total on-screen magnification of x1,025, resulting in a rectangular screen of 310 x 230 µm. The sections were analyzed by point-counting of frames after systematic, uniformly random sampling (SURS) using a PC-controlled motorized stage and the CAST-GRID software (Olympus).
The sections were examined with the observer blinded to the origin of the sections. Initially, the tissue sections were circumscribed using a x1.25 objective, and the counting of endocrine and exocrine structures took place within this area. The volume fraction of tissues was estimated by point-counting stereological techniques on the 310 x 230 µm screens of tissue (14). -Cell volume was estimated using a grid of 4 x 64 points per screen, nonpancreatic tissue (fat and connective tissue) volume was estimated using a grid of four points per screen, and pancreas volume was estimated using a grid of one point per screen. Step length between screens was maximally 900 x 600 µm and controlled by the CAST-GRID software. Thus 800 (SD200) screens were analyzed per pancreas, resulting in 534 (SD132, n = 35) screens scored for pancreas volume and 437 (SD341) hits for -cell volume. Mean values of estimated volume fractions were calculated from area weighted mean values, so that the volume fraction of -cells was calculated as the total number of hits for -cell volume over the two to three sections analyzed divided by the total number of hits for pancreas volume over the two to three sections and the grid ratio (256). Total -cell mass was calculated from the volume fraction of -cells and the weight of the pancreas, corrected for the presence of nonpancreatic tissue, and assuming that -cells and the rest of the pancreas have the same density. -Cell mass is expressed as absolute mass (mg) for each individual animal.
In groups of identically treated animals (non-STZ or STZ animals from previous studies in our laboratory), the total coefficient of error (CE) for determination of -cell mass (variation between animals and from method) is 25–40%. In the material included in the present study, the between-sections 250 µm apart CE was 16%. The involved SURS method alone contributes with an estimated CE of 3% and the point-counting "noise" with a CE of 3% with the number of fractions (10–12 fractions) and "hits" used for the present studies (7, 15). An increase in the number of sections or points counted would only marginally have reduced the overall CE by 2–3%. Therefore, the methodology used in the present study fairly represents the true difference in -cell mass between animals.
The technique of using point counting has been described in detail by other laboratories (7, 15), and we have previously compared the use of point counting with image analysis of whole sections of the pancreas to estimate -cell mass in our laboratory (Gotfredsen CF, unpublished results). The correlation between the two methods in our laboratory was r2 = 0.8545, P < 0.0001, n = 18.
The average number of days between evaluation of pulsatile insulin secretion and histology was 40 (SD 28), and this period was significantly longer in obese vs. lean-STZ animals [90 (SD 15) vs. 20 (SD 8), P < 0.01], whereas no significant differences were found between any other pairs of groups [lean 34 (SD 9), obese-STZ 50 (SD 15)]. Similarly, the average number of days between glucose and arginine stimulation and histology was 30 (SD 25), and this period was significantly longer in obese vs. lean-STZ animals [77 (SD 14) vs. 14 (SD 9), P < 0.001], whereas no significant differences were found between any other pairs of groups [lean 30 (SD 7), obese-STZ 25 (SD 10)].
Evaluation of Results
Acute insulin response. Acute insulin response (AIR) to glucose and arginine was calculated as the average increase in insulin levels during the first 10 min after dosing: AIR = (mean insulin 0–10 min) – (baseline before dosing).
Maximum insulin.
Maximum insulin concentration during stimulation with glucose and arginine was used as an additional measure of -cell response. Maximum insulin concentration was generally seen in one of the blood samples taken after the arginine bolus.
Detection and quantification of pulsatile insulin secretion by deconvolution.
The plasma insulin concentration time series were analyzed by deconvolution for the purpose of detection and quantification of insulin secretory bursts, as previously described (25). In short, deconvolution of venous insulin concentration data was performed with a multiparameter technique (49), which requires the following assumptions. The venous plasma insulin concentrations measured in samples collected at 1-min intervals were assumed to have resulted from five determinable and correlated parameters: 1) a finite number of discrete insulin secretory bursts occurring at specific times and having 2) individual amplitudes (maximal rate of secretion attained within a burst); 3) a common half-duration (duration of an algebraically Gaussian secretory pulse at half-maximal amplitude), which is superimposed on a 4) basal time-invariant insulin secretory rate; and 5) a biexponential insulin disappearance model in the systemic circulation. The mass of hormone secreted per burst (time integral of the calculated secretory burst) was computed as picomoles of insulin released per liter of systemic distribution volume, and pulse mass during both baseline and entrained conditions was evaluated in relation to -cell mass. All data analysis by deconvolution was performed in a blinded manner by a single person.
Statistics
All calculations and statistical evaluation of results were performed using Excel (2000) and GraphPad Prism v. 3.00 for Windows (GraphPad Software, San Diego, CA). Statistical evaluation using Prism included one-way ANOVA (with Tukey’s multiple comparison test as a post hoc test) or the Kruskal-Wallis test (with Dunn’s test as post hoc test), Student’s t-test, or Mann-Whitney test. Linear regression was done using Prism. P values of 0.05 or less were considered significant. Data are presented as means (SD) in the text.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
-Cell Mass
The range of -cell mass was from 28 to 931 mg (Fig. 1), and there was considerable overlap between groups. The STZ-dosed animals had significantly reduced -cell mass compared with animals not dosed with STZ [168 (SD 89) vs. 557 (SD 157) mg, P < 0.001]. Obese non-STZ animals tended to have increased -cell mass compared with lean non-STZ animals [643 (SD 162) vs. 508 (SD 108) mg, P = 0.06; Fig. 1]. On the basis of the overlap between the groups, correlations between in vivo tests and -cell mass were analyzed across the entire data set included in the study.
|
The range of body weights was from 24 to 65 kg, with lean animals weighing significantly less than obese animals [33 (SD 5) (range 24–41) vs. 54 (SD 6) (range 47–65) kg, P < 0.0001; Fig. 1].
Similarly, the range of fasting plasma glucose values was 2.6 to 9.2 mM, with STZ-dosed animals having significantly higher fasting plasma glucose than animals not dosed with STZ [4.5 (SD 1.5) (range 3.4–9.2) vs. 3.4 (SD 0.5) (range 2.6–4.1) mM, P < 0.01; Fig. 1].
Obese animals tended to have slightly lower fasting plasma glucose than lean animals [3.2 (SD 0.5) vs. 3.7 (SD 0.3)], but this was not significant.
Relation Between Parameters from In Vivo Tests and -Cell Mass
AIR 0.3, AIR 0.6, and AIR arginine, as well as maximum insulin and entrained pulse mass, were all significantly correlated to -cell mass across the study groups (Fig. 2), whereas basal pulse mass was not. Extensive stimulation of -cells gave the best correlations to mass across the groups (AIR 0.6, AIR arginine, and maximum insulin), with maximum insulin concentrations during the test giving the best correlation to -cell mass. To explore possible reasons for poor correlations for AIR 0.3 and pulse mass, regression lines for lean and obese animals were compared. For AIR 0.3, slopes (pM/mg) were significantly different in lean (0.7503 ± 0.1032) vs. obese animals (0.2818 ± 0.1188, P = 0.005). It was, therefore, not possible to evaluate whether the difference seen for intercept on the y-axis for -cell mass = 0 was significantly different between obese and lean animals (2.141 ± 25.88 pM for lean; 57.42 ± 59.77 pM for obese animals). Similarly, slopes for basal pulse mass were different between lean (0.2394 ± 0.0746) and obese (0.0126 ± 0.0247, P = 0.01) animals (y-axis intercept for lean: –6.168 ± 18.99; for obese: 29.80 ± 12.40). Slopes for entrained pulse mass were not significantly different between lean (0.3306 ± 0.0994) and obese (0.1892 ± 0.0400, P = 0.2) animals. Furthermore, the y-axis intercept did not differ significantly between lean (15.35 ± 27.16) and obese (10.66 ± 20.10, P = 0.056), although there was a tendency for obese animals to have lower intercept. For AIR 0.6, AIR arginine, and maximum insulin, no differences were found between slopes or intercepts between lean and obese animals (data not shown).
|
-cell mass than any of the other tests. Similarly, when lean animals were considered separately, baseline or entrained pulse mass did not show stronger correlation with -cell mass than any of the other tests, whereas, when considering obese animals, entrained pulse mass showed correlations in the same range, or slightly stronger than the other tests (data not shown). Relation Between In Vivo Tests
To be able to evaluate the strength of the correlations between in vivo tests and -cell mass, we looked at the correlation between parameters such as AIR 0.6 and maximum insulin (r2 = 0.9530, P < 0.0001) and AIR arginine and AIR 0.6 (r2 = 0.6595, P < 0.0001) across the study groups. These parameters are very closely related, since they all reflect insulin responses during extensive -cell stimulation. The observation that the correlations between closely related parameters from in vivo tests and -cell mass are in the same range is an indication of a very strong biological relationship between these parameters, further supporting the use of parameters based on in vivo insulin responses as a surrogate measure of -cell mass in this model.
In lean animals, both basal and entrained pulse mass were strongly correlated with insulin responses to 0.3 and 0.6 g/kg glucose, arginine, and maximum insulin (r2 ranging from 0.5515 to 0.9471), whereas no strong correlation was seen between these parameters in the obese animals (r2 ranging from 0 to 0.4466).
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
-cell mass across a group of lean and obese animals. Furthermore, the relationships were similar for those parameters that involve extensive stimulation of insulin secretion (AIR 0.6, AIR arginine, and maximum insulin concentration) when lean and obese animals were evaluated separately. However, with parameters based on either no (baseline pulse mass) or considerably reduced stimulation of insulin secretion (entrained pulse mass and AIR 0.3), a tendency for a steeper decline with reduction of -cell mass in lean compared with obese animals was observed.
The dose-response curve for glucose-stimulated insulin secretion in humans is very steep at glucose concentrations between 7.5 and 10 mM, whereas at glucose concentrations between 10 and 15 mM the increase is less marked (41). Thus the use of extensive stimulation of insulin secretion with plasma glucose concentrations of 10 mM during in vivo tests should provide the most useful estimate for -cell mass in the minipig over a range of body weights, and this is in accord with what we have previously reported in lean minipigs (29).
An argument against extensive stimulation of -cells could be the risk of reducing insulin clearance, because this phenomenon has been reported in humans (46) and would result in an overestimate of -cell mass. However, the present results show a strong correlation between -cell mass and maximum insulin concentration, so any overestimate of -cell mass seems to be similar at low and high mass, enabling the relationship between maximum insulin concentration and mass to be described with a linear function in the range studied.
Because insulin levels and responses and insulin resistance are closely related in humans (2, 19, 20, 32, 47), it is of special interest that the strong correlation between parameters from in vivo tests and mass is present across a group of lean and obese animals. These findings in an animal model could indicate that the use of methods involving extensive stimulation of insulin secretion may possibly provide an estimate of -cell mass in population studies in humans including a broad range of body weights, such as, for instance, patients with the metabolic syndrome and/or type 2 diabetes. However, before the use of such an estimate in a human population, it would be necessary to validate the method in more detail in animals. In particular, it would be of relevance to include obese animals with more marked hyperglycemia. We (29) have previously reported that there is a good correlation between -cell mass and insulin responses in severely hyperglycemic lean animals, but this correlation still remains to be demonstrated in obese animals with marked hyperglycemia. Furthermore, it remains to be determined whether the same methodology can be used to evaluate changes in -cell mass after pharmacological intervention in this model.
The weaker correlation with -cell mass seen after 0.3 g/kg glucose could indicate that not all -cells respond to the more moderate stimuli. A weaker correlation might be due to different secretory stimulation in the population of -cells, as has previously been shown in vitro (40, 48). Alternatively, it could be speculated that less extensive stimulation of insulin secretion would allow animals with reduced -cell mass to compensate to a larger extent than when more extensive stimulation is applied, as has been observed in both dogs (51) and baboons with reduced -cell mass (37).
Evaluation of pulsatile insulin secretion has been shown to be a very sensitive method for detection of changes in -cell responses (17, 18, 27, 44) and to correlate with -cell mass in lean minipigs (22, 26). The present results confirm that deconvolution analysis of pulsatile insulin secretion can provide a useful estimate of -cell mass. However, this method does not seem to provide better estimates of -cell mass compared with some of the more simple tests (stimulation with glucose and/or arginine).
Even though the present study was focused on determining surrogate markers of -cell mass, the glycemic levels in obese animals deserve attention in their own right. First, the slightly lower fasting plasma glucose in the obese animals could indicate a powerful compensation for the obesity in the fasted state, and a similar tendency for slightly lower glucose levels has recently been reported in obese humans (10). We have previously shown in these animals that around two years of obesity is mostly linked to subtle changes in the coordination of pulsatile insulin secretion (27), whereas it remains to be determined whether longer-term obesity will have an effect on insulin responses in vivo and their relation to -cell mass in this model.
Second, it is surprising that obese-STZ animals did not show increased fasting hyperglycemia compared with lean-STZ animals, since the combination of obesity and reduced -cell mass could be expected to further deteriorate metabolic control. Whether such abnormalities would develop over a longer period still remains to be determined. On the basis of our results, further studies of the relation between obesity and compensatory mechanisms to maintain glucose tolerance in this animal model would be very interesting.
In conclusion, evaluation of insulin concentrations and appearance rates in vivo can provide a highly predictive estimate of -cell mass across a wide range of body weights in the Göttingen minipig. Evaluation of pulsatile insulin secretion did not show a stronger relationship to -cell mass compared with some of the simpler tests used. The strongest correlation with -cell mass across the groups was seen after extensive stimulation of insulin secretion with both glucose and arginine when maximum insulin concentration was measured during the in vivo test. This finding in an animal model could support the evaluation of similar methods in the future to estimate -cell mass in studies in humans over a range of body weights. Furthermore, the ability to estimate -cell mass on the basis of in vivo experiments would allow longitudinal studies on changes in -cell mass in minipigs to evaluate the dynamics of this process in vivo.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
-Cell function and islet morphology in normal, obese, and obese -cell mass-reduced Göttingen minipigs. Am J Physiol Endocrinol Metab 288: E412–E421, 2005.This article has been cited by other articles:
|
|
 |
|
  R. P. Robertson Estimation of {beta}-Cell Mass by Metabolic Tests: Necessary, but How Sufficient? Diabetes, October 1, 2007; 56(10): 2420 - 2424. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
, lean (A and C, n = 9; B and D, n = 4); 
, lean-STZ (n = 16); 
, obese (n = 5); 
, obese-STZ animals (A and B, n = 5).