- Email: [email protected]
The endocrine pancreas in virgin and pregnant offspring of diabetic pregnant rats
Diabetes Research and Clinical Practice 38 (1997) 9 – 19
The endocrine pancreas in virgin and pregnant offspring of diabetic pregnant rats L. Aerts *, L. Vercruysse, F.A. Van Assche Laboratory of Obstetrics and Gynaecology, U.Z. Gasthuisberg, Herestraat 49, B3001 Leu6en, Belgium Received 17 February 1997; received in revised form 14 May 1997; accepted 9 July 1997
Abstract Diabetes of the mother during pregnancy induces structural and functional adaptations in the fetal endocrine pancreas. We have previously shown in our experimental rat model, that the impact of this abnormal intra-uterine milieu leads, in the adult offspring, to a disturbance of the glucose homeostasis and to the development of gestational diabetes. The aim of the present work is to investigate wether these functional differences can be explained by structural differences at the level of the endocrine pancreas. Therefore the size and the structure of the endocrine pancreas, as well as the contribution of the insulin-, glucagon-, somatostatin- and PP-cells, were investigated morphometrically in the adult youngsters of mildly and of severely diabetic mothers, since both display a disturbed glucose tolerance but with divergent characteristics. Also the adaptation of their endocrine pancreas to pregnancy was measured and compared to that of a control pregnancy. In the offspring of mildly diabetic mothers, the size of the endocrine pancreas and the distribution of the islets of Langerhans are normal. Also the doubling of the endocrine mass during pregnancy is similar to controls. The high proportion of A-cells, especially in relation to a normal B-cell mass and the low amount of PP-cells, might play a role in the impairment of the insulin response in these animals and in the development of gestational diabetes. In the offspring of severely diabetic mothers a clear hypertrophy of the endocrine pancreas is noted, which is mainly due to the presence of numerous small islets and which does not increase further during pregnancy. In these animals, the size of the endocrine pancreas and of the B-cell mass have reached ‘pregnant’ values without pregnancy, which coincides with an exaggerated insulin output and peripheral insulin resistance, as during normal pregnancy. No further increase in islet mass is seen during pregnancy, which is associated with gestational diabetes. © 1997 Elsevier Science Ireland Ltd. Keywords: Endocrine pancreas; Islets of Langerhans; Morphometry; Glucose tolerance; Gestational diabetes; Insulin; Glucagon; Somatostatin; Pancreatic polypeptide
* Corresponding author. 0168-8227/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved. PII S 0 1 6 8 - 8 2 2 7 ( 9 7 ) 0 0 0 8 0 - 6
10
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
1. Introduction Diabetes during pregnancy induces adaptations in the fetal metabolism, which result in persistent consequences in later life. Glucose homeostasis is impaired in the adult youngsters of diabetic mothers and the characteristics of this impairment are divergent according to the degree of the maternal diabetes: deficient insulin to glucose response in the youngsters of mildly diabetic mothers, exaggerated response associated with peripheral insulin resistance in the youngsters of severely diabetic mothers. When pregnant, all animals develop gestational diabetes. The characteristics of these alterations in experimental work have been extensively described by our [1 – 5] and other [6 – 8] laboratories. Epidemiological studies have confirmed similar effects in humans [9 – 13]. The occurrence of gestational diabetes in these offspring is of crucial importance, since it presents yet again an abnormal intra-uterine milieu to the developing fetuses and so transmits the effects to the next generation. This experimental design might therefore present a model for the ontogeny of disturbed glucose tolerance and gestational diabetes. The aim of the present study in the rat is to investigate whether the metabolic alterations in these youngsters from diabetic mothers can be explained by structural characteristics of their endocrine pancreas and of its adaptation during pregnancy. An extensive morphometric analysis has therefore been performed on pancreatic biopsies of control rats (CO) and of adult female youngsters of both mildly (MD) and severely (SD) diabetic mothers, non-pregnant and at term pregnancy. In each group, the mass of endocrine tissue is evaluated, the number and size of the islets of Langerhans and the islet size distribution. Also the contribution of the different endocrine celltypes, insulin-, glucagon-, somatostatin- and pancreatic polypeptide-cells, is determined.
by a single intravenous injection of streptozotocin (Upjohn) into the tail vein. According to the injected dose (30 or 50 mg/kg body weight), the animals developed mild (MD) or severe (SD) diabetes, with plasma glucose levels, at day 20 of gestation, of 9.8 mM9 0.4 (nine) and 28 mM9 1.6 (ten), versus 7.3 mM9 0.2 (ten) in the controls. These animals were allowed to deliver spontaneously and to raise their pups in standard laboratory conditions. Only litters with eight or more pups were included in the study. Part of the female offspring of these MD- and SD-mothers, as well as control animals (CO), were killed at adulthood (80 days) by cervical dislocation. The others were caged overnight with a control male and examined for copulation plug in the morning. Pregnant animals from each group were killed at day 20 of gestation. General features of these animals have been described previously [4]. Blood was taken from the tip of the tail of the rats without anesthesia for determination of plasma glucose and insulin concentrations. The ‘Principals of Laboratory Animal Care’ (NIH publication No. 85-23, revised 1985) were followed.
2.2. Biopsies The complete pancreas of each animal was removed, coiled into a homogeneuos mass without any orientation, and immediately fixed in Bouin’s solution. The biopsies were embedded in paraffin in random position. From each pancreas three series of ten consecutive sections (3 mm) were made, each series being at a distance of at least 1 mm from the previous one. From each animal, three sections, one from each series, were stained immunohistologically for insulin. Similarly, three consecutive sections from each biopsy were stained for glucagon, somatostatin and pancreatic polypeptide.
2.3. Immunohistological staining 2. Material and methods
2.1. Animals Diabetes was induced in adult female Wistar rats at day 1 of gestation (day of copulation plug)
After dewaxing and blocking endogenous peroxidase with 0.5% hydrogen peroxide in absolute methanol for 15 min, irrelevant binding sites were blocked by a 30 min incubation with either normal goat serum (Dako: 1:30, containing 0.5% Tween-80) in the staining for pancreatic polypep-
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
tide or with bovine serum albumin (BSA fraction V 2% w/v) in all other staining sequences. Subsequently, serial sections were incubated at 4°C overnight with one of the following primary antibodies: guinea-pig anti-porcine insulin (MilesYeda: 1:32 000), rabbit anti-porcine glucagon (Milab: 1:40 960), rabbit anti-somatostatin (kindly provided by W. Gepts, Brussels, Belgium: 1:10 000) rabbit anti-bovine pancreatic polypeptide (Cappel-Organon Technica: 1:2000). After a second protein blocking step with normal goat serum or BSA, anti-insulin and anti-pancreatic polypeptide were detected with peroxidase-conjugated goat anti-guinea pig and goat anti-rabbit immunoglobulins respectively (Jackson Immunoresearch Laboratories: 1:100); anti-glucagon and anti-somatostatin were detected with goat anti-rabbit immunoglobulins (kindly provided by F. Van de Sande, Leuven, Belgium: 1:80) and subsequent incubation with peroxidase – antiperoxidase complex (Dako: 1:300). In order to avoid cross-reaction with tissue immunoglobulins all secondary antibody dilutions were mixed with 4% complement-inactivated normal rat serum, 30 min before use. The peroxidase enzyme activity was demonstrated with diaminobenzidine by standard methods and sections were counterstained lightly with Mayer’s hematoxylin, dehydrated and mounted.
11
each islet included in the 255 randomly chosen fields, the total area of endocrine pancreas, the mean area of islet profiles and the volume density of insulin-positive tissue in the islets were calculated and the number of islets was noted. These data were collected per animal, and expressed as mean9 S.D. per group. The size distribution of the islet profiles was also calculated from these data, but for this purpose all islets (9 500) were pooled per experimental group. On the slides stained immunohistologically for glucagon, somatostatin and pancreatic polypeptide, 100 islets were examined from each animal for each hormone, equally distributed at random over three distended sections. Also here the area of each islet was manually delineated and the positivelystained area within this surface was quantitated automatically. From these data, the volume density of positive tissue in the islet was calculated for each hormone. All data on the endocrine pancreas, as measured in this standard area, can be used to compare absolute amounts of endocrine tissue between the different groups, since all measurements were performed on a similar area of pancreatic tissue for each animal and since pancreatic size is not different between the three experimental groups, nor changes during pregnancy [14].
2.5. Statistics 2.4. Morphometry Sections were examined with a ZEISS Axioskoop microscope with an objective lens magnification of ×20, connected to a VIDAS Automatic Image Analyzer System (KONTRON, Germany). Measurements were performed on histological images as projected on the video screen with a final magnification of ×600, each examined field representing 0.08 mm2 of pancreatic tissue. On each of the three insulin-stained sections per animal, 85 consecutive fields (255 in total) were examined, representing a total surface of 20 mm2. On each field the area of endocrine tissue was measured by manually delineating the border of each islet profile, while the area of insulin-positive tissue in the islet was registered automatically. From these data, registered for
All data were expressed as mean9 S.D. Differences between the six groups, between the three non-pregnant groups and between the three pregnant groups were tested by the Kruskall-Wallis test and if appropriate, further evaluated by the non-parametric Mann-Witney U-test for comparison of groups with non-Gaussian distribution. Differences were considered significant if P5 0.05.
3. Results In the non-pregnant rats plasma insulin and glucose levels were similar in controls and in offspring of mildly and of severely diabetic mothers. During pregnancy insulin levels rise and glu-
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
12
Table 1 Plasma glucose and insulin levels in non-pregnant and pregnant offspring of control, mildly diabetic and severely diabetic mothers Controls
MD
SD
Insulin (mU/ml) Non-preg. Pregnant
1492 (7) 57 97 (5)*
1592 (7) 43911 (9)°*
159 2 (10) 489 14 (6)°*
Glucose (mM) Non-preg. Pregnant
6.790.5 (7) 6.99 0.2 (7) 6.89 0.2 (10) 4.490.2 (5)* 6.09 0.2 (9)°* 5.39 0.2 (6)°*
Mean9S.D. (n); ° PB0.05 versus controls, * PB0.05 versus non-pregnant.
cose levels drop in the youngsters of MD and SD mothers, but not to the same extent as in the controls and at day 20 of gestation both parameters are significantly different from normal (Table 1). In the non-pregnant animals the total area of endocrine pancreas (Fig. 1) is higher in the SDoffspring than in the controls (P B0.01). During normal pregnancy the area of endocrine tissue in the pancreas doubles (P=0.001). A similar in-
Fig. 1. Total area of endocrine pancreas measured on the investigated pancreatic biopsy in non-pregnant (N) and pregnant (P) control rats (CO) and offspring of mildly (MD) and of severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean9 S.D. per group; the number of animals is six for each group; °, significant difference (P5 0.05) of MD or SD versus CO; *, significant difference (P5 0.05) of P versus corresponding N.
Fig. 2. Number of islets on the investigated pancreatic biopsy and mean area of the islet profiles in non-pregnant (N) and pregnant (P) control rats (CO) and offspring of mildly (MD) and of severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean9 S.D. per group; the number of animals is six for each group; °, significant difference (P5 0.05) of MD or SD versus CO; *, significant difference (P5 0.05) of P versus corresponding N.
crease occurs in the offspring of the MD-group (PB0.005), but in the SD-group the difference is not significant. However, the total area occupied by endocrine tissue at the end of gestation is similar in the three groups. The amount of endocrine tissue is determined by two factors: the number of islets and the size of the islets (Fig. 2). In the non-pregnant animals the number of islets is not different in the MDyoungsters but is significantly higher (P= 0.02) in the SD-youngsters as compared to the controls. During normal pregnancy, there is a significant increase (P= 0.02) in the number of islets, but this increase is smaller (× 1.2) than the increase in endocrine area (×2). The increase in islet number during pregnancy is similar to the controls in the
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
13
Fig. 3. Islet size distribution histogram of all the islet profiles (upper panel, A) and of the small islet profiles (B10.000 mm2; lower panel, B) measured on the investigated pancreatic biopsies in non-pregnant (N) and pregnant (P) control rats (CO) and offspring of mildly (MD) and of severely (SD) diabetic mothers. Data are sampled per group; number of islets between 450 and 550 per group. Number of islets on the y-axis.
MD-offspring (×1.2, P =0.03) but is absent in the SD-youngsters. In the pregnant offspring the number of islets is not significantly different between the three groups. The mean area of the islet profiles in the nonpregnant MD- and SD-youngsters is not significantly different from control values. During pregnancy the increase in mean area of islet profiles is approximately ×1.5 in control- and MDyoungsters (P = 0.03 and P =0.07 versus non pregnant, respectively), but is very small (× 1.2) in the SD-youngsters (NS). In the pregnant animals the mean area of islet profiles is similar in the three groups. In control non-pregnant animals the size distri-
bution of the islet profiles ranks from 50 mm2 up to 60.000 mm2 (Fig. 3A). From this islet population 87% consists of ‘small islets’ with an islet profile smaller than 10.000 m 2 (Fig. 3B), constituting together about 15% of the total islet mass. More than half of the islets (55%) is even smaller than 2.000 mm2. For the MD- and SD-youngsters the shape of the histogram for islet size distribution is quite similar to that of the controls, but larger islets do occur, although rarely, and the number of small islets is increased, especially in the SD-group. During normal pregnancy, there is an increase in number in all islet classes and the maximal size of the largest islets increases considerably (up to 100.000 mm2), but still 83% of the
14
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
islet profiles are smaller than 10.000 mm2 and 50% smaller than 2.000 mm2. A similar pattern is seen in the MD-youngsters, but in the SD-group the difference with non-pregnant rats is minimal. Comparing the three groups of pregnant animals, we see very large islets in the CO- and MD-groups but not in the SD-youngsters, where on the contrary the number of small islets is higher. The volume density of the insulin-positive tissue in the endocrine pancreas (Fig. 4) remains remarkably constant in the three groups of nonpregnant animals: 955%. It remains at the same level during pregnancy in the controls and the MD-youngsters, but it is significantly higher (P = 0.05) in the pregnant SD-offspring The total area of insulin-positive tissue, as measured on the standard pancreatic area (Fig. 5), is not significantly different in the MD-youngsters but markedly higher (P =0.004) in the SD-youngsters. During normal pregnancy, there is a doubling (× 2.1) of the insulin-positive area (P= 0.001); in the MD-offspring this increase is somewhat lower (× 1.6) but still very significant (P= 0.005), while in the SD-offspring it is much smaller (× 1.4) and not significant. The difference in insulin-positive area between the groups of pregnant animals is therefore minimal, and not significant. The volume density of glucagon-positive tissue in the islets (Fig. 4) is considerably higher in the non-pregnant MD- and SD-youngsters than in the controls (PB0.0005 and P B0.0001, respectively). It has a tendency to decrease during pregnancy, which is only significant in the SD-group. Comparing the three groups of pregnant animals, the volume density of glucagon-cells within the islets appears to be significantly higher (P B0.05) in the MD-youngsters. The absolute area of glucagon-positive tissue (Fig. 5) is increased in both experimental groups (× 1.7 for the MD youngsters, P B0.05 and × 2.2 for the SD, P= 0.001). The glucagon area does increase during pregnancy in the controls (P=0.04) and in the MD-youngsters (P = 0.05), but not in the SD-animals where its value was already very high before gestation. No significant difference is seen between the three groups of
pregnant animals, although the value is high in the MD-group. The ratio between the area of insulin-containing tissue and the area of glucagon-containing tissue (Fig. 6), although lower than normal in both experimental groups, is only significantly different for the MD-youngsters (P= 0.02). The insulin/ glucagon ratio has a tendency to increase during
Fig. 4. Volume density of insulin-, glucagon-, somatostatinand pancreatic polypeptide-positive tissue in the islets of Langerhans in non-pregnant (N) and pregnant (P) control rats (CO) and in offspring of mildly (MD) and severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean 9 S.D. per group; °, significant difference (P5 0.05) of MD or SD versus CO; *, significant difference (P5 0.05) of P versus corresponding N.
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
15
value of the controls, P=0.001). During normal pregnancy, the relative contribution of PP-cells to the islets decreases (P= 0.05), due to the fact that the absolute amount of PP-cells remains quite stable within the increased islet mass. In the offspring of the MD-mothers the low volume density of PP-tissue in the islets does not change significantly during pregnancy and remains low as compared to normal pregnant values (P= 0.05). In the offspring of the SD-mothers, no changes occur. When expressed as absolute values (Fig. 5), the amount of PP-cells is very low in the non-pregnant MD-youngsters.
4. Discussion
4.1. The endocrine pancreas in non pregnant adult rats
Fig. 5. Area occupied by insulin-, glucagon-, somatostatinand pancreatic polypeptide-positive tissue in the standard area of pancreatic tissue in non-pregnant (N) and pregnant (P) control rats (CO) and in offspring of mildly (MD) and severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean9S.D. per group; °, significant difference (P 5 0.05) of MD or SD versus CO; *, significant difference of P versus corresponding N.
pregnancy (only significant in the SD-group, P B 0.005), and is not different between the three pregnant groups. The volume density of somatostatin-positive tissue in the islets (Fig. 4) is remarkably constant in all groups with or without pregnancy. The somatostatin-containing area (Fig. 5) is similar in all non-pregnant animals. It increases during pregnancy in the controls (P =0.05) and in the MD-group (P =0.004), but not in the SDgroup, and the final area of somatostatin tissue is not different between the three groups of pregnant animals. The volume density of pancreatic polypeptide positive tissue in the islets (Fig. 4) is very low in the non-pregnant MD-youngsters (only half the
Although previous studies have shown a morphologically normal endocrine pancreas in adult offspring of diabetic mothers [15], a more extended quantitation of the islets of Langerhans, performed on insulin-stained material in order to detect even the smallest clusters of islet cells and
Fig. 6. Ratio of the area of insulin-positive tissue and the area of glucagon-positive tissue in non-pregnant (N) and pregnant (P) control rats (CO) and in offspring of mildly (MD) and severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean 9S.D. per group; °, significant difference (P5 0.05) of MD or SD versus CO; *, significant difference of P versus corresponding N.
16
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
Fig. 7. Histological images of islets of Langerhans from a control pregnant rat, stained immunohistologically for insulin (A), glucagon (B), somatostatin (C) and pancreatic polypeptide (D). (final magnification of × 85).
including quantification of the different endocrine celltypes (Fig. 7), reveals significant differences with control animals. These alterations must result from the perinatal influences induced by the maternal diabetes. In the fetuses of mildly diabetic mothers, the permanent mild hyperglycemia has stimulated the fetal endocrine pancreas to islet hyperplasia at term gestation [16]. During suckling, a period of poor carbohydrate supply, the involution in the mass of endocrine pancreas [17] is more pronounced in the pups of diabetic mothers than in the controls and the islet mass in these youngsters at weaning is deficient [15]. By adulthood however, their endocrine pancreas appears to have restored from the perinatal influences. Also the proportion of insulin-containing cells in the islets and the total mass of B-cells are normal. The deficient insulin response in these adult youngsters [1,3] can therefore not be attributed to a deficient islet- or B-cell mass. The high proportion of Acells, especially in relation to the normal B-cell mass, might play a role in the disturbed equi-
librium of glucose handling, although basal glucagon levels are normal in these MD- as in the SD-youngsters [4]. Glucagon, as an antagonist of insulin, increases plasma glucose levels by glycogenolysis and gluconeogenesis from aminoacids in the liver. From our previous data, we know that exactly in these MD-youngsters the livers are heavier than normal [2], the circulating levels of amino-acids are lower and the total nitrogen excretion in the urine is increased [2,18]. These combined data suggest a higher turnover of amino-acids, in compensation for a lower mobilization of hepatic glycogen. The resulting low plasma amino-acid levels as such might be involved in the deficient insulin response in these animals, since normalization of the amino-acid levels normalizes also the stimulated insulin output [2]. No changes are found in the amount of somatostatin-containing D-cells, known to modulate the function of both insulin and glucagoncells by paracrine secretion. The size of the somatostatin cell population in the endocrine pancreas is indeed little or not affected by experimen-
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
tal conditions [19–21]. Pancreatic polypeptide is known to inhibit the pancreatic secretion by modulating the vagal tone on the pancreas, via PP-receptors in the dorsal vagal complex of the brainstem [22,23]. Quantitation of PP-cells is complicated by an unequal distribution of these cells in different parts of the pancreas, which can be divided in a PP-rich (head of the pancreas) and a PP-poor (tail of the pancreas) zone [14]. In the present study, which was mainly focused on insulin- and glucagon-cells, the whole pancreas was collected and randomly examined, as was necessary for that purpose. Results on PP-cells therefore give global data, without regard for zonal differences. The low amount of PP cells in these MD-youngsters suggests an involvement of the nervous system in their deficient insulin to glucose response. In the fetuses of severely diabetic mothers, the permanent severe hyperglycemia has induced islet hyperplasia, overstimulation and exhaustion of the B-cells, resulting in hypoinsulinemia and microsomia [16]. During suckling, the mass of endocrine pancreas in these pups drops in a similar way as in the MD-youngsters [15] and equally recuperates afterwards. By the time of adulthood the endocrine pancreas is clearly hypertrophic and this is mainly due to an increase in the number of small and very small islets. The contribution of B-cells to the islet mass remains normal. This abundance of insulin-producing cells allows the adult youngsters of severely diabetic mothers to maintain their basal glucose homeostasis, despite the high insulin requirements due to insulin resistance in their liver and peripheral tissues [5], but it can not explain per se the exaggerated insulin response to glucose stimulation in these animals [3]. Since the insulin/glucagon ratio is not different from control values and since the contribution of the other endocrine cell-types in the pancreas is also normal, their involvement in the impaired glucose tolerance in these animals seems to be neglectable.
4.2. The endocrine pancreas in pregnancy During pregnancy the increased metabolism of the mother requires extension of the endocrine
17
pancreas. Our data confirm previous publications in control rats [14,24], showing a doubling of the mass of endocrine tissue. This hypertrophy of the total islet mass is associated with a mild increase in the number of islets, which is mainly due to an increased number of small and even very small islets, consisting of only a few cells and often situated in the vicinity of pancreatic ducts. It is also evident that islets grow much larger during pregnancy and that large islets are more numerous than before gestation. Our data thus suggest that the extensive growth of the endocrine pancreas during pregnancy is achieved by the same processes as those involved in normal islet maintenance and in major regeneration processes [25,26]: islet neogenesis and B-cell replication. This hypertrophy of the endocrine pancreas parallels the hyperplasia of B-, A- and D-cells, while the contribution of the PP-cells remains rather constant. In the MD-youngsters the hypertrophy of the endocrine pancreas as a result of pregnancy occurs parallel to that of the controls, including a similar hyperplasia of B-, A- and D-cells. The low contribution of PP-cells to the islets confirms the involvement of the vagal nerve system in the deficient insulin secretion of these animals. However, the structural adaptations of the endocrine pancreas to pregnancy appear to be adequate in these MD-youngsters and the origin of their deficient basal insulin secretion must be found rather in a failure to adapt the B-cell sensitivity and responsiveness to the demands of pregnancy. In the non-pregnant SD-youngsters, all parameters have reached ‘pregnant’ values already before pregnancy, including the total islet and B-cell mass. These values seem to have reached a limiting point, since they do not increase further under the influence of pregnancy, as they do in normal animals. It appears that in the SD-youngsters in basal conditions all means are already mobilized up to their maximal capacity in the B-cells as in the peripheral tissues, where an insulin resistance comparable to that of normal pregnancy is present already before gestation [5,27]. Neither of these two aspects of glucose handling, insulin production by the B-cell and insulin handling by the peripheral tissues, adapts further to the demands of pregnancy, resulting in
18
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
low circulating insulin levels and mild but significant hyperglycemia. It can be concluded that in the adult offspring of mildly diabetic mothers neither the impaired glucose tolerance, nor the gestational diabetes, can be attributed to a lack of B-cells. Differences in the islet composition however might play a role. In the adult offspring of severely diabetic mothers islet hypertrophy and B-cell hyperplasia reach pregnant values before gestation and are associated with increased insulin response and peripheral insulin resistance. They do not adapt further during pregnancy, resulting in gestational diabetes.
[10]
[11]
[12]
[13]
References [14] [1] L. Aerts, F.A. Van Assche, Is gestational diabetes an acquired condition?, J. Dev. Physiol. 1 (1979) 219–225. [2] L. Aerts, K. Holemans, F.A Van Assche, Impaired insulin response and action in offspring of severely diabetes rats, in: E. Shafrir (Ed.), Frontiers in Diabetes Research. Lessons from Animal Diabetes III, SmithGordon and Company Limited (London), 1990, pp. 561 – 566. [3] L. Aerts, F. Sodoyez-Goffaux, J.C. Sodoyez, W.J. Malaisse, F.A. Van Assche, The diabetic intrauterine milieu has a long-lasting effect on insulin secretion by B-cells and on insulin uptake by target tissues, Am. J. Obstet. Gynecol. 159 (1988) 1287–1292. [4] L. Aerts, K. Holemans, F.A. Van Assche, Maternal diabetes during pregnancy: consequences for the offspring, Diabetes Metab. Rev. 6 (1990) 147–167. [5] K. Holemans, L. Aerts, F.A. Van Assche, Evidence for an insulin resistance in the adult offspring of pregnant streptozotocin-diabetic rats, Diabetologia 34 (1991) 81– 85. [6] M.T. Bihoreau, A. Ktorza, M.F. Kinebanyan, L. Picon, Impaired glucose homeostasis in adult rats from hyperglycemic mothers, Diabetes 35 (1986) 979–984. [7] A. Ktorza, D. Gauguier, M.T. Bihoreau, M.F. Berthault, L. Picon, Adult offspring from mildly hyperglycemic rats show impairment of glucose regulation and insulin secretion which is transmissible to the next generation, in: E. Shafrir (Ed.), Frontiers in Diabetes Research. Lessons from Animal Diabetes III, SmithGordon and Company Limited (London), 1990, pp. 555 – 560. [8] D. Gauguier, M.T. Bihoreau, A. Ktorza, M.F. Berthault, L. Picon, Inheritance of diabetes mellitus as consequence of gestational hyperglycemia in rats, Diabetes 39 (1990) 734 –739. [9] J.W. Farquhar, Prognosis for babies born to diabetic
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
mothers in Edinburgh, Arch. Dis. Child 44 (1969) 36 – 47. W. Knowler, D.J. Pettitt, C.I. Kunzelman, J. Everhart, Genetic and environmental determinants of non-insulin diabetes mellitus, Diabetes Res. Clin. Pract. S1 (1985) S309. A.O. Martin, J.L. Simpson, C. Ober, N. Freinkel, Frequency of diabetes mellitus in mothers of probands with gestational diabetes: possible maternal influence on the predisposition to gestational diabetes, Am. J. Obstet. Gynecol. 151 (1985) 471 – 475. D.J. Pettitt, K.A. Aleck, H.R. Baird, M.J. Carraher, P.H. Bennett, W.C. Knowler, Congenital susceptibility to NIDDM role of intrauterine environment, Diabetes 37 (1988) 622 – 628. G. Dorner, E. Steindel, H. Thoelke, Evidence for decreasing prevalence of diabetes mellitus in childhood apparently produced by prevention of hyperinsulinism in the foetus and newborn, Exp. Clin. Endocrinol. 84 (1984) 134 – 142. F.A. Van Assche, W. Gepts, L. Aerts, Immunocytochemical study of the endocrine pancreas in the rat during normal pregnancy and during experimental diabetic pregnancy, Diabetologia 18 (1980) 487 – 491. L. Aerts, F.A. Van Assche, Endocrine pancreas in the offspring of rats with experimentally induced diabetes, J. Endocrinol. 88 (1981) 81 – 88. L. Aerts, F.A. Van Assche, Rat foetal endocrine pancreas in experimental diabetes, J. Endocrinol. 73 (1977) 339 – 346. D.T. Finegood, L. Scaglia, S. Bonner-Weir, Dynamics of B-cell mass in the growing rat pancreas: estimation with a simple mathematical model, Diabetes 44 (1995) 249. L. Aerts, R. Van Bree, V. Feytons, W. Rombauts, F.A. Van Assche, Plasma amino-acids in diabetic pregnant rats and in their fetal and adult offspring, Biol. Neonate 56 (1989) 31 – 39. L.I. Larsson, N. Goltermann, L. De Magitus, J.F. Rehfeld, T.W. Schwartz, Somatostatin cell processes as pathways for paracrine secretion, Science 205 (1979) 1393 – 1395. J.P. Pegorier, P.H. Duee, J. Girard, J. Peret, Development of gluconeogenesis in isolated hepatocytes from fasting or suckling newborn pigs, J. Nutr. 112 (1982) 1038 – 1046. P. Pons, A. Aoki, Differential proliferation of somatostatin and glucagon cells in rat pancreatic islets submitted to various stimuli, Anat. Anz. 177 (1995) 221 – 227. D.M. McTigue, R.C. Rogers, Pancreatic polypeptide stimulates gastric motility through a vagal-dependent mechanism in rats, Neurosci. Lett. 188 (1995) 93 – 96. T. Okumura, T.N. Pappas, I.L. Taylor, Pancreatic polypeptide microinjection into the dorsal motor nucleus inhibits pancreatic secretion in rats, Gasteroenterology 108 (1995) 1517.
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19 [24] F.A. Van Assche, Quantitative morphologic and histoenzymatic study of the endocrine pancreas in nonpregnant and pregnant rats, Am. J. Obstet. Gynecol. 118 (1974) 39–41. [25] S. Bonner-Weir, L.A. Baxter, G.T. Schuppin, F.E. Smith, A second pathway for regeneration of adult exocrine and endocrine pancreas. A possible recapitulation of embryonic development, Diabetes 42 (1993) 1715–1720.
.
19
[26] R.N. Wang, G. Kloeppel, L. Bouwens, Duct- to islet-cell differentiation and islet growth in the pancreas of duct-ligated adult rats, Diabetologia 38 (1995) 1405. [27] K. Holemans, L. Aerts, F.A. Van Assche, Absence of pregnancy-induced alterations in tissue insulin sensitivity in the offspring of diabetic rats, J. Endocrinol. 131 (1991) 387 – 393.
The endocrine pancreas in virgin and pregnant offspring of diabetic pregnant rats L. Aerts *, L. Vercruysse, F.A. Van Assche Laboratory of Obstetrics and Gynaecology, U.Z. Gasthuisberg, Herestraat 49, B3001 Leu6en, Belgium Received 17 February 1997; received in revised form 14 May 1997; accepted 9 July 1997
Abstract Diabetes of the mother during pregnancy induces structural and functional adaptations in the fetal endocrine pancreas. We have previously shown in our experimental rat model, that the impact of this abnormal intra-uterine milieu leads, in the adult offspring, to a disturbance of the glucose homeostasis and to the development of gestational diabetes. The aim of the present work is to investigate wether these functional differences can be explained by structural differences at the level of the endocrine pancreas. Therefore the size and the structure of the endocrine pancreas, as well as the contribution of the insulin-, glucagon-, somatostatin- and PP-cells, were investigated morphometrically in the adult youngsters of mildly and of severely diabetic mothers, since both display a disturbed glucose tolerance but with divergent characteristics. Also the adaptation of their endocrine pancreas to pregnancy was measured and compared to that of a control pregnancy. In the offspring of mildly diabetic mothers, the size of the endocrine pancreas and the distribution of the islets of Langerhans are normal. Also the doubling of the endocrine mass during pregnancy is similar to controls. The high proportion of A-cells, especially in relation to a normal B-cell mass and the low amount of PP-cells, might play a role in the impairment of the insulin response in these animals and in the development of gestational diabetes. In the offspring of severely diabetic mothers a clear hypertrophy of the endocrine pancreas is noted, which is mainly due to the presence of numerous small islets and which does not increase further during pregnancy. In these animals, the size of the endocrine pancreas and of the B-cell mass have reached ‘pregnant’ values without pregnancy, which coincides with an exaggerated insulin output and peripheral insulin resistance, as during normal pregnancy. No further increase in islet mass is seen during pregnancy, which is associated with gestational diabetes. © 1997 Elsevier Science Ireland Ltd. Keywords: Endocrine pancreas; Islets of Langerhans; Morphometry; Glucose tolerance; Gestational diabetes; Insulin; Glucagon; Somatostatin; Pancreatic polypeptide
* Corresponding author. 0168-8227/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved. PII S 0 1 6 8 - 8 2 2 7 ( 9 7 ) 0 0 0 8 0 - 6
10
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
1. Introduction Diabetes during pregnancy induces adaptations in the fetal metabolism, which result in persistent consequences in later life. Glucose homeostasis is impaired in the adult youngsters of diabetic mothers and the characteristics of this impairment are divergent according to the degree of the maternal diabetes: deficient insulin to glucose response in the youngsters of mildly diabetic mothers, exaggerated response associated with peripheral insulin resistance in the youngsters of severely diabetic mothers. When pregnant, all animals develop gestational diabetes. The characteristics of these alterations in experimental work have been extensively described by our [1 – 5] and other [6 – 8] laboratories. Epidemiological studies have confirmed similar effects in humans [9 – 13]. The occurrence of gestational diabetes in these offspring is of crucial importance, since it presents yet again an abnormal intra-uterine milieu to the developing fetuses and so transmits the effects to the next generation. This experimental design might therefore present a model for the ontogeny of disturbed glucose tolerance and gestational diabetes. The aim of the present study in the rat is to investigate whether the metabolic alterations in these youngsters from diabetic mothers can be explained by structural characteristics of their endocrine pancreas and of its adaptation during pregnancy. An extensive morphometric analysis has therefore been performed on pancreatic biopsies of control rats (CO) and of adult female youngsters of both mildly (MD) and severely (SD) diabetic mothers, non-pregnant and at term pregnancy. In each group, the mass of endocrine tissue is evaluated, the number and size of the islets of Langerhans and the islet size distribution. Also the contribution of the different endocrine celltypes, insulin-, glucagon-, somatostatin- and pancreatic polypeptide-cells, is determined.
by a single intravenous injection of streptozotocin (Upjohn) into the tail vein. According to the injected dose (30 or 50 mg/kg body weight), the animals developed mild (MD) or severe (SD) diabetes, with plasma glucose levels, at day 20 of gestation, of 9.8 mM9 0.4 (nine) and 28 mM9 1.6 (ten), versus 7.3 mM9 0.2 (ten) in the controls. These animals were allowed to deliver spontaneously and to raise their pups in standard laboratory conditions. Only litters with eight or more pups were included in the study. Part of the female offspring of these MD- and SD-mothers, as well as control animals (CO), were killed at adulthood (80 days) by cervical dislocation. The others were caged overnight with a control male and examined for copulation plug in the morning. Pregnant animals from each group were killed at day 20 of gestation. General features of these animals have been described previously [4]. Blood was taken from the tip of the tail of the rats without anesthesia for determination of plasma glucose and insulin concentrations. The ‘Principals of Laboratory Animal Care’ (NIH publication No. 85-23, revised 1985) were followed.
2.2. Biopsies The complete pancreas of each animal was removed, coiled into a homogeneuos mass without any orientation, and immediately fixed in Bouin’s solution. The biopsies were embedded in paraffin in random position. From each pancreas three series of ten consecutive sections (3 mm) were made, each series being at a distance of at least 1 mm from the previous one. From each animal, three sections, one from each series, were stained immunohistologically for insulin. Similarly, three consecutive sections from each biopsy were stained for glucagon, somatostatin and pancreatic polypeptide.
2.3. Immunohistological staining 2. Material and methods
2.1. Animals Diabetes was induced in adult female Wistar rats at day 1 of gestation (day of copulation plug)
After dewaxing and blocking endogenous peroxidase with 0.5% hydrogen peroxide in absolute methanol for 15 min, irrelevant binding sites were blocked by a 30 min incubation with either normal goat serum (Dako: 1:30, containing 0.5% Tween-80) in the staining for pancreatic polypep-
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
tide or with bovine serum albumin (BSA fraction V 2% w/v) in all other staining sequences. Subsequently, serial sections were incubated at 4°C overnight with one of the following primary antibodies: guinea-pig anti-porcine insulin (MilesYeda: 1:32 000), rabbit anti-porcine glucagon (Milab: 1:40 960), rabbit anti-somatostatin (kindly provided by W. Gepts, Brussels, Belgium: 1:10 000) rabbit anti-bovine pancreatic polypeptide (Cappel-Organon Technica: 1:2000). After a second protein blocking step with normal goat serum or BSA, anti-insulin and anti-pancreatic polypeptide were detected with peroxidase-conjugated goat anti-guinea pig and goat anti-rabbit immunoglobulins respectively (Jackson Immunoresearch Laboratories: 1:100); anti-glucagon and anti-somatostatin were detected with goat anti-rabbit immunoglobulins (kindly provided by F. Van de Sande, Leuven, Belgium: 1:80) and subsequent incubation with peroxidase – antiperoxidase complex (Dako: 1:300). In order to avoid cross-reaction with tissue immunoglobulins all secondary antibody dilutions were mixed with 4% complement-inactivated normal rat serum, 30 min before use. The peroxidase enzyme activity was demonstrated with diaminobenzidine by standard methods and sections were counterstained lightly with Mayer’s hematoxylin, dehydrated and mounted.
11
each islet included in the 255 randomly chosen fields, the total area of endocrine pancreas, the mean area of islet profiles and the volume density of insulin-positive tissue in the islets were calculated and the number of islets was noted. These data were collected per animal, and expressed as mean9 S.D. per group. The size distribution of the islet profiles was also calculated from these data, but for this purpose all islets (9 500) were pooled per experimental group. On the slides stained immunohistologically for glucagon, somatostatin and pancreatic polypeptide, 100 islets were examined from each animal for each hormone, equally distributed at random over three distended sections. Also here the area of each islet was manually delineated and the positivelystained area within this surface was quantitated automatically. From these data, the volume density of positive tissue in the islet was calculated for each hormone. All data on the endocrine pancreas, as measured in this standard area, can be used to compare absolute amounts of endocrine tissue between the different groups, since all measurements were performed on a similar area of pancreatic tissue for each animal and since pancreatic size is not different between the three experimental groups, nor changes during pregnancy [14].
2.5. Statistics 2.4. Morphometry Sections were examined with a ZEISS Axioskoop microscope with an objective lens magnification of ×20, connected to a VIDAS Automatic Image Analyzer System (KONTRON, Germany). Measurements were performed on histological images as projected on the video screen with a final magnification of ×600, each examined field representing 0.08 mm2 of pancreatic tissue. On each of the three insulin-stained sections per animal, 85 consecutive fields (255 in total) were examined, representing a total surface of 20 mm2. On each field the area of endocrine tissue was measured by manually delineating the border of each islet profile, while the area of insulin-positive tissue in the islet was registered automatically. From these data, registered for
All data were expressed as mean9 S.D. Differences between the six groups, between the three non-pregnant groups and between the three pregnant groups were tested by the Kruskall-Wallis test and if appropriate, further evaluated by the non-parametric Mann-Witney U-test for comparison of groups with non-Gaussian distribution. Differences were considered significant if P5 0.05.
3. Results In the non-pregnant rats plasma insulin and glucose levels were similar in controls and in offspring of mildly and of severely diabetic mothers. During pregnancy insulin levels rise and glu-
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
12
Table 1 Plasma glucose and insulin levels in non-pregnant and pregnant offspring of control, mildly diabetic and severely diabetic mothers Controls
MD
SD
Insulin (mU/ml) Non-preg. Pregnant
1492 (7) 57 97 (5)*
1592 (7) 43911 (9)°*
159 2 (10) 489 14 (6)°*
Glucose (mM) Non-preg. Pregnant
6.790.5 (7) 6.99 0.2 (7) 6.89 0.2 (10) 4.490.2 (5)* 6.09 0.2 (9)°* 5.39 0.2 (6)°*
Mean9S.D. (n); ° PB0.05 versus controls, * PB0.05 versus non-pregnant.
cose levels drop in the youngsters of MD and SD mothers, but not to the same extent as in the controls and at day 20 of gestation both parameters are significantly different from normal (Table 1). In the non-pregnant animals the total area of endocrine pancreas (Fig. 1) is higher in the SDoffspring than in the controls (P B0.01). During normal pregnancy the area of endocrine tissue in the pancreas doubles (P=0.001). A similar in-
Fig. 1. Total area of endocrine pancreas measured on the investigated pancreatic biopsy in non-pregnant (N) and pregnant (P) control rats (CO) and offspring of mildly (MD) and of severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean9 S.D. per group; the number of animals is six for each group; °, significant difference (P5 0.05) of MD or SD versus CO; *, significant difference (P5 0.05) of P versus corresponding N.
Fig. 2. Number of islets on the investigated pancreatic biopsy and mean area of the islet profiles in non-pregnant (N) and pregnant (P) control rats (CO) and offspring of mildly (MD) and of severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean9 S.D. per group; the number of animals is six for each group; °, significant difference (P5 0.05) of MD or SD versus CO; *, significant difference (P5 0.05) of P versus corresponding N.
crease occurs in the offspring of the MD-group (PB0.005), but in the SD-group the difference is not significant. However, the total area occupied by endocrine tissue at the end of gestation is similar in the three groups. The amount of endocrine tissue is determined by two factors: the number of islets and the size of the islets (Fig. 2). In the non-pregnant animals the number of islets is not different in the MDyoungsters but is significantly higher (P= 0.02) in the SD-youngsters as compared to the controls. During normal pregnancy, there is a significant increase (P= 0.02) in the number of islets, but this increase is smaller (× 1.2) than the increase in endocrine area (×2). The increase in islet number during pregnancy is similar to the controls in the
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
13
Fig. 3. Islet size distribution histogram of all the islet profiles (upper panel, A) and of the small islet profiles (B10.000 mm2; lower panel, B) measured on the investigated pancreatic biopsies in non-pregnant (N) and pregnant (P) control rats (CO) and offspring of mildly (MD) and of severely (SD) diabetic mothers. Data are sampled per group; number of islets between 450 and 550 per group. Number of islets on the y-axis.
MD-offspring (×1.2, P =0.03) but is absent in the SD-youngsters. In the pregnant offspring the number of islets is not significantly different between the three groups. The mean area of the islet profiles in the nonpregnant MD- and SD-youngsters is not significantly different from control values. During pregnancy the increase in mean area of islet profiles is approximately ×1.5 in control- and MDyoungsters (P = 0.03 and P =0.07 versus non pregnant, respectively), but is very small (× 1.2) in the SD-youngsters (NS). In the pregnant animals the mean area of islet profiles is similar in the three groups. In control non-pregnant animals the size distri-
bution of the islet profiles ranks from 50 mm2 up to 60.000 mm2 (Fig. 3A). From this islet population 87% consists of ‘small islets’ with an islet profile smaller than 10.000 m 2 (Fig. 3B), constituting together about 15% of the total islet mass. More than half of the islets (55%) is even smaller than 2.000 mm2. For the MD- and SD-youngsters the shape of the histogram for islet size distribution is quite similar to that of the controls, but larger islets do occur, although rarely, and the number of small islets is increased, especially in the SD-group. During normal pregnancy, there is an increase in number in all islet classes and the maximal size of the largest islets increases considerably (up to 100.000 mm2), but still 83% of the
14
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
islet profiles are smaller than 10.000 mm2 and 50% smaller than 2.000 mm2. A similar pattern is seen in the MD-youngsters, but in the SD-group the difference with non-pregnant rats is minimal. Comparing the three groups of pregnant animals, we see very large islets in the CO- and MD-groups but not in the SD-youngsters, where on the contrary the number of small islets is higher. The volume density of the insulin-positive tissue in the endocrine pancreas (Fig. 4) remains remarkably constant in the three groups of nonpregnant animals: 955%. It remains at the same level during pregnancy in the controls and the MD-youngsters, but it is significantly higher (P = 0.05) in the pregnant SD-offspring The total area of insulin-positive tissue, as measured on the standard pancreatic area (Fig. 5), is not significantly different in the MD-youngsters but markedly higher (P =0.004) in the SD-youngsters. During normal pregnancy, there is a doubling (× 2.1) of the insulin-positive area (P= 0.001); in the MD-offspring this increase is somewhat lower (× 1.6) but still very significant (P= 0.005), while in the SD-offspring it is much smaller (× 1.4) and not significant. The difference in insulin-positive area between the groups of pregnant animals is therefore minimal, and not significant. The volume density of glucagon-positive tissue in the islets (Fig. 4) is considerably higher in the non-pregnant MD- and SD-youngsters than in the controls (PB0.0005 and P B0.0001, respectively). It has a tendency to decrease during pregnancy, which is only significant in the SD-group. Comparing the three groups of pregnant animals, the volume density of glucagon-cells within the islets appears to be significantly higher (P B0.05) in the MD-youngsters. The absolute area of glucagon-positive tissue (Fig. 5) is increased in both experimental groups (× 1.7 for the MD youngsters, P B0.05 and × 2.2 for the SD, P= 0.001). The glucagon area does increase during pregnancy in the controls (P=0.04) and in the MD-youngsters (P = 0.05), but not in the SD-animals where its value was already very high before gestation. No significant difference is seen between the three groups of
pregnant animals, although the value is high in the MD-group. The ratio between the area of insulin-containing tissue and the area of glucagon-containing tissue (Fig. 6), although lower than normal in both experimental groups, is only significantly different for the MD-youngsters (P= 0.02). The insulin/ glucagon ratio has a tendency to increase during
Fig. 4. Volume density of insulin-, glucagon-, somatostatinand pancreatic polypeptide-positive tissue in the islets of Langerhans in non-pregnant (N) and pregnant (P) control rats (CO) and in offspring of mildly (MD) and severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean 9 S.D. per group; °, significant difference (P5 0.05) of MD or SD versus CO; *, significant difference (P5 0.05) of P versus corresponding N.
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
15
value of the controls, P=0.001). During normal pregnancy, the relative contribution of PP-cells to the islets decreases (P= 0.05), due to the fact that the absolute amount of PP-cells remains quite stable within the increased islet mass. In the offspring of the MD-mothers the low volume density of PP-tissue in the islets does not change significantly during pregnancy and remains low as compared to normal pregnant values (P= 0.05). In the offspring of the SD-mothers, no changes occur. When expressed as absolute values (Fig. 5), the amount of PP-cells is very low in the non-pregnant MD-youngsters.
4. Discussion
4.1. The endocrine pancreas in non pregnant adult rats
Fig. 5. Area occupied by insulin-, glucagon-, somatostatinand pancreatic polypeptide-positive tissue in the standard area of pancreatic tissue in non-pregnant (N) and pregnant (P) control rats (CO) and in offspring of mildly (MD) and severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean9S.D. per group; °, significant difference (P 5 0.05) of MD or SD versus CO; *, significant difference of P versus corresponding N.
pregnancy (only significant in the SD-group, P B 0.005), and is not different between the three pregnant groups. The volume density of somatostatin-positive tissue in the islets (Fig. 4) is remarkably constant in all groups with or without pregnancy. The somatostatin-containing area (Fig. 5) is similar in all non-pregnant animals. It increases during pregnancy in the controls (P =0.05) and in the MD-group (P =0.004), but not in the SDgroup, and the final area of somatostatin tissue is not different between the three groups of pregnant animals. The volume density of pancreatic polypeptide positive tissue in the islets (Fig. 4) is very low in the non-pregnant MD-youngsters (only half the
Although previous studies have shown a morphologically normal endocrine pancreas in adult offspring of diabetic mothers [15], a more extended quantitation of the islets of Langerhans, performed on insulin-stained material in order to detect even the smallest clusters of islet cells and
Fig. 6. Ratio of the area of insulin-positive tissue and the area of glucagon-positive tissue in non-pregnant (N) and pregnant (P) control rats (CO) and in offspring of mildly (MD) and severely (SD) diabetic mothers. Data are sampled per animal and expressed as mean 9S.D. per group; °, significant difference (P5 0.05) of MD or SD versus CO; *, significant difference of P versus corresponding N.
16
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
Fig. 7. Histological images of islets of Langerhans from a control pregnant rat, stained immunohistologically for insulin (A), glucagon (B), somatostatin (C) and pancreatic polypeptide (D). (final magnification of × 85).
including quantification of the different endocrine celltypes (Fig. 7), reveals significant differences with control animals. These alterations must result from the perinatal influences induced by the maternal diabetes. In the fetuses of mildly diabetic mothers, the permanent mild hyperglycemia has stimulated the fetal endocrine pancreas to islet hyperplasia at term gestation [16]. During suckling, a period of poor carbohydrate supply, the involution in the mass of endocrine pancreas [17] is more pronounced in the pups of diabetic mothers than in the controls and the islet mass in these youngsters at weaning is deficient [15]. By adulthood however, their endocrine pancreas appears to have restored from the perinatal influences. Also the proportion of insulin-containing cells in the islets and the total mass of B-cells are normal. The deficient insulin response in these adult youngsters [1,3] can therefore not be attributed to a deficient islet- or B-cell mass. The high proportion of Acells, especially in relation to the normal B-cell mass, might play a role in the disturbed equi-
librium of glucose handling, although basal glucagon levels are normal in these MD- as in the SD-youngsters [4]. Glucagon, as an antagonist of insulin, increases plasma glucose levels by glycogenolysis and gluconeogenesis from aminoacids in the liver. From our previous data, we know that exactly in these MD-youngsters the livers are heavier than normal [2], the circulating levels of amino-acids are lower and the total nitrogen excretion in the urine is increased [2,18]. These combined data suggest a higher turnover of amino-acids, in compensation for a lower mobilization of hepatic glycogen. The resulting low plasma amino-acid levels as such might be involved in the deficient insulin response in these animals, since normalization of the amino-acid levels normalizes also the stimulated insulin output [2]. No changes are found in the amount of somatostatin-containing D-cells, known to modulate the function of both insulin and glucagoncells by paracrine secretion. The size of the somatostatin cell population in the endocrine pancreas is indeed little or not affected by experimen-
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
tal conditions [19–21]. Pancreatic polypeptide is known to inhibit the pancreatic secretion by modulating the vagal tone on the pancreas, via PP-receptors in the dorsal vagal complex of the brainstem [22,23]. Quantitation of PP-cells is complicated by an unequal distribution of these cells in different parts of the pancreas, which can be divided in a PP-rich (head of the pancreas) and a PP-poor (tail of the pancreas) zone [14]. In the present study, which was mainly focused on insulin- and glucagon-cells, the whole pancreas was collected and randomly examined, as was necessary for that purpose. Results on PP-cells therefore give global data, without regard for zonal differences. The low amount of PP cells in these MD-youngsters suggests an involvement of the nervous system in their deficient insulin to glucose response. In the fetuses of severely diabetic mothers, the permanent severe hyperglycemia has induced islet hyperplasia, overstimulation and exhaustion of the B-cells, resulting in hypoinsulinemia and microsomia [16]. During suckling, the mass of endocrine pancreas in these pups drops in a similar way as in the MD-youngsters [15] and equally recuperates afterwards. By the time of adulthood the endocrine pancreas is clearly hypertrophic and this is mainly due to an increase in the number of small and very small islets. The contribution of B-cells to the islet mass remains normal. This abundance of insulin-producing cells allows the adult youngsters of severely diabetic mothers to maintain their basal glucose homeostasis, despite the high insulin requirements due to insulin resistance in their liver and peripheral tissues [5], but it can not explain per se the exaggerated insulin response to glucose stimulation in these animals [3]. Since the insulin/glucagon ratio is not different from control values and since the contribution of the other endocrine cell-types in the pancreas is also normal, their involvement in the impaired glucose tolerance in these animals seems to be neglectable.
4.2. The endocrine pancreas in pregnancy During pregnancy the increased metabolism of the mother requires extension of the endocrine
17
pancreas. Our data confirm previous publications in control rats [14,24], showing a doubling of the mass of endocrine tissue. This hypertrophy of the total islet mass is associated with a mild increase in the number of islets, which is mainly due to an increased number of small and even very small islets, consisting of only a few cells and often situated in the vicinity of pancreatic ducts. It is also evident that islets grow much larger during pregnancy and that large islets are more numerous than before gestation. Our data thus suggest that the extensive growth of the endocrine pancreas during pregnancy is achieved by the same processes as those involved in normal islet maintenance and in major regeneration processes [25,26]: islet neogenesis and B-cell replication. This hypertrophy of the endocrine pancreas parallels the hyperplasia of B-, A- and D-cells, while the contribution of the PP-cells remains rather constant. In the MD-youngsters the hypertrophy of the endocrine pancreas as a result of pregnancy occurs parallel to that of the controls, including a similar hyperplasia of B-, A- and D-cells. The low contribution of PP-cells to the islets confirms the involvement of the vagal nerve system in the deficient insulin secretion of these animals. However, the structural adaptations of the endocrine pancreas to pregnancy appear to be adequate in these MD-youngsters and the origin of their deficient basal insulin secretion must be found rather in a failure to adapt the B-cell sensitivity and responsiveness to the demands of pregnancy. In the non-pregnant SD-youngsters, all parameters have reached ‘pregnant’ values already before pregnancy, including the total islet and B-cell mass. These values seem to have reached a limiting point, since they do not increase further under the influence of pregnancy, as they do in normal animals. It appears that in the SD-youngsters in basal conditions all means are already mobilized up to their maximal capacity in the B-cells as in the peripheral tissues, where an insulin resistance comparable to that of normal pregnancy is present already before gestation [5,27]. Neither of these two aspects of glucose handling, insulin production by the B-cell and insulin handling by the peripheral tissues, adapts further to the demands of pregnancy, resulting in
18
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19
low circulating insulin levels and mild but significant hyperglycemia. It can be concluded that in the adult offspring of mildly diabetic mothers neither the impaired glucose tolerance, nor the gestational diabetes, can be attributed to a lack of B-cells. Differences in the islet composition however might play a role. In the adult offspring of severely diabetic mothers islet hypertrophy and B-cell hyperplasia reach pregnant values before gestation and are associated with increased insulin response and peripheral insulin resistance. They do not adapt further during pregnancy, resulting in gestational diabetes.
[10]
[11]
[12]
[13]
References [14] [1] L. Aerts, F.A. Van Assche, Is gestational diabetes an acquired condition?, J. Dev. Physiol. 1 (1979) 219–225. [2] L. Aerts, K. Holemans, F.A Van Assche, Impaired insulin response and action in offspring of severely diabetes rats, in: E. Shafrir (Ed.), Frontiers in Diabetes Research. Lessons from Animal Diabetes III, SmithGordon and Company Limited (London), 1990, pp. 561 – 566. [3] L. Aerts, F. Sodoyez-Goffaux, J.C. Sodoyez, W.J. Malaisse, F.A. Van Assche, The diabetic intrauterine milieu has a long-lasting effect on insulin secretion by B-cells and on insulin uptake by target tissues, Am. J. Obstet. Gynecol. 159 (1988) 1287–1292. [4] L. Aerts, K. Holemans, F.A. Van Assche, Maternal diabetes during pregnancy: consequences for the offspring, Diabetes Metab. Rev. 6 (1990) 147–167. [5] K. Holemans, L. Aerts, F.A. Van Assche, Evidence for an insulin resistance in the adult offspring of pregnant streptozotocin-diabetic rats, Diabetologia 34 (1991) 81– 85. [6] M.T. Bihoreau, A. Ktorza, M.F. Kinebanyan, L. Picon, Impaired glucose homeostasis in adult rats from hyperglycemic mothers, Diabetes 35 (1986) 979–984. [7] A. Ktorza, D. Gauguier, M.T. Bihoreau, M.F. Berthault, L. Picon, Adult offspring from mildly hyperglycemic rats show impairment of glucose regulation and insulin secretion which is transmissible to the next generation, in: E. Shafrir (Ed.), Frontiers in Diabetes Research. Lessons from Animal Diabetes III, SmithGordon and Company Limited (London), 1990, pp. 555 – 560. [8] D. Gauguier, M.T. Bihoreau, A. Ktorza, M.F. Berthault, L. Picon, Inheritance of diabetes mellitus as consequence of gestational hyperglycemia in rats, Diabetes 39 (1990) 734 –739. [9] J.W. Farquhar, Prognosis for babies born to diabetic
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
mothers in Edinburgh, Arch. Dis. Child 44 (1969) 36 – 47. W. Knowler, D.J. Pettitt, C.I. Kunzelman, J. Everhart, Genetic and environmental determinants of non-insulin diabetes mellitus, Diabetes Res. Clin. Pract. S1 (1985) S309. A.O. Martin, J.L. Simpson, C. Ober, N. Freinkel, Frequency of diabetes mellitus in mothers of probands with gestational diabetes: possible maternal influence on the predisposition to gestational diabetes, Am. J. Obstet. Gynecol. 151 (1985) 471 – 475. D.J. Pettitt, K.A. Aleck, H.R. Baird, M.J. Carraher, P.H. Bennett, W.C. Knowler, Congenital susceptibility to NIDDM role of intrauterine environment, Diabetes 37 (1988) 622 – 628. G. Dorner, E. Steindel, H. Thoelke, Evidence for decreasing prevalence of diabetes mellitus in childhood apparently produced by prevention of hyperinsulinism in the foetus and newborn, Exp. Clin. Endocrinol. 84 (1984) 134 – 142. F.A. Van Assche, W. Gepts, L. Aerts, Immunocytochemical study of the endocrine pancreas in the rat during normal pregnancy and during experimental diabetic pregnancy, Diabetologia 18 (1980) 487 – 491. L. Aerts, F.A. Van Assche, Endocrine pancreas in the offspring of rats with experimentally induced diabetes, J. Endocrinol. 88 (1981) 81 – 88. L. Aerts, F.A. Van Assche, Rat foetal endocrine pancreas in experimental diabetes, J. Endocrinol. 73 (1977) 339 – 346. D.T. Finegood, L. Scaglia, S. Bonner-Weir, Dynamics of B-cell mass in the growing rat pancreas: estimation with a simple mathematical model, Diabetes 44 (1995) 249. L. Aerts, R. Van Bree, V. Feytons, W. Rombauts, F.A. Van Assche, Plasma amino-acids in diabetic pregnant rats and in their fetal and adult offspring, Biol. Neonate 56 (1989) 31 – 39. L.I. Larsson, N. Goltermann, L. De Magitus, J.F. Rehfeld, T.W. Schwartz, Somatostatin cell processes as pathways for paracrine secretion, Science 205 (1979) 1393 – 1395. J.P. Pegorier, P.H. Duee, J. Girard, J. Peret, Development of gluconeogenesis in isolated hepatocytes from fasting or suckling newborn pigs, J. Nutr. 112 (1982) 1038 – 1046. P. Pons, A. Aoki, Differential proliferation of somatostatin and glucagon cells in rat pancreatic islets submitted to various stimuli, Anat. Anz. 177 (1995) 221 – 227. D.M. McTigue, R.C. Rogers, Pancreatic polypeptide stimulates gastric motility through a vagal-dependent mechanism in rats, Neurosci. Lett. 188 (1995) 93 – 96. T. Okumura, T.N. Pappas, I.L. Taylor, Pancreatic polypeptide microinjection into the dorsal motor nucleus inhibits pancreatic secretion in rats, Gasteroenterology 108 (1995) 1517.
L. Aerts et al. / Diabetes Research and Clinical Practice 38 (1997) 9–19 [24] F.A. Van Assche, Quantitative morphologic and histoenzymatic study of the endocrine pancreas in nonpregnant and pregnant rats, Am. J. Obstet. Gynecol. 118 (1974) 39–41. [25] S. Bonner-Weir, L.A. Baxter, G.T. Schuppin, F.E. Smith, A second pathway for regeneration of adult exocrine and endocrine pancreas. A possible recapitulation of embryonic development, Diabetes 42 (1993) 1715–1720.
.
19
[26] R.N. Wang, G. Kloeppel, L. Bouwens, Duct- to islet-cell differentiation and islet growth in the pancreas of duct-ligated adult rats, Diabetologia 38 (1995) 1405. [27] K. Holemans, L. Aerts, F.A. Van Assche, Absence of pregnancy-induced alterations in tissue insulin sensitivity in the offspring of diabetic rats, J. Endocrinol. 131 (1991) 387 – 393.