Insulin and glucagon release from the isolated pancreas of foetal and newborn mice

Maturation of the insulin-releasing mechanism in foetal and newborn animals has been studied in several species. It seems to be established that glucose is unable to promote insulin release in several foetal mammals although glucagon and leucine are effective insulin secretagogues (see Grodsky, 1970). The fact that glucagon is able to stimulate insulin release from foetal β-cells (Milner, 1969; Milner, Barson & Ashworth, 1971; Espinosa de Los Monteros, Driscoll & Steinke, 1970) raises the following question: when do the glucagon-producing α₂-cells acquire functional competence? Is endogenous glucagon of significance for the release of insulin in response to arginine in foetal and newborn mice?

Adult mice were obtained from a local colony carrying the gene ob (Hellman, 1965). The homozygous occurrence of this gene gives rise to a syndrome of obesity and hyperglycemia (Ingalls, Dickie & Snell, 1950). However, in the present study only mice of normal phenotype were used. Female mice were caged with males for 2 days. The second day was counted as the first day of the pregnancies that resulted. The length of gestation in our colony is 22–23 days. All animals were fed ad libitum and were killed by decapitation. The pancreatic glands of 18- and 21-day-old foetuses and of 1- and 3-day-old mice were rapidly excised under a stereomicroscope.

The pancreases from all foetuses (5–6) of each pregnant mouse were immediately dropped into isopentane chilled to its freezing-point (–160°C) by liquid nitrogen. After freeze-drying, the specimens were weighed and extracted by the acid-ethanol method of Kenny (1955). The proteins precipitated with ethanolether were freeze-dried. Appropriate dilutions of the freeze-dried precipitate were assayed for GLI (glucagon-like immunoreactivity). When [¹²⁵I]glucagon was added to samples of the mouse-pancreas extract, 20–30% was recovered in the extraction procedure.

For determination of pancreatic IRI (immunoreactive insulin), aliquots of the above mentioned acid-ethanol extract were transferred to polyethylene microtubes and protein was precipitated by the addition of acetone. The precipitate was collected by centrifugation, washed with 2 ml of acetone and centrifuged once again. The supernatant was discarded and the precipitate freeze-dried. Dilutions of the freeze-dried extracts were assayed for IRI. When [¹²⁵I]insulin was added to samples of the mouse pancreatic extract, 60–70% was recovered in the extraction procedure.

Whole pancreatic glands were incubated in vitro using the method of Coore & Randle (1964) and the amounts of GLI and IRI released into the medium were measured by radioimmunoassay (see below). The basal incubation medium consisted of Krebs-Ringer bicarbonate (KRB) medium (Umbreit, Burris & Stauffer, 1964) supplemented with 500 Kallikrein Inactivator Units (KIU) of Trasylol ® and 5 mg human serum albumin per ml. The KRB media were equilibrated with 95% O₂ and 5% CO₂. All incubations were performed at 37°C with shaking (120 strokes/min with an amplitude of 1·25 cm). Each incubation vessel contained 5–8 pancreases from 18-day-old foetuses, 3–5 pancreases from 1-day-old mice or 2 pancreases from 3-day-old mice. The total wet weight of pancreas included in each vessel was 5–11 mg. A preincubation period of 30 min was followed by five successive incubation periods, each of 30 min duration: (1) medium without glucose, (2) medium with 17 mM glucose, (3) medium without glucose, (4) medium with 5 mM L-arginine and (5) medium with 5 mM L-arginine plus 17 mM glucose. After incubation, the pancreatic glands were quickly removed from the incubation vessels and weighed. The incubation media were immediately frozen and stored at – 90°C until assayed for GLI and IRI within 3 weeks.

Glucagon-like immunoreactivity (GLI) was measured in triplicate for the standards and in duplicate for the unknown samples by the procedure of Heding (1970). Anti-glucagon serum (Batch K 37) and [¹²⁵I]glucagon were kindly donated by L. G. Heding, Novo Research Institute, Copenhagen, Denmark.

To determine the GLI release in vitro, all incubation media and the standards (crystalline porcine glucagon) had to be assayed directly in the KRB medium. The crystalline human serum albumin (AB Kabi, Stockholm, Sweden) used was found not to degrade the [¹²⁵I]glucagon (Heding, personal communication). The difference between standard curves for crystalline porcine glucagon prepared in 0·04 M phosphate buffer, pH 7·4 (Heding, 1970), or in basal KRB medium, both containing 500KlU/ml Trasylol, is shown in Fig. 1. The amounts of GLI measured when the incubation media were serially diluted in the KRB medium were exactly proportional to the degree of dilution (Fig. 2).

Immunoreactive insulin (IR1) was determined in duplicate on each sample using the radioimmunoassay technique of Heding (1966). Crystalline mouse insulin was used as standard.

Student’s t test was used for comparison of the pancreatic contents of insulin and glucagon. Effects observed in the in vitro experiments were calculated from the mean difference between control and test incubations in each series of independent experiments.

The pancreatic concentrations of extractable immunoreactive glucagon and insulin in 16- to 21-day-old foetal and 1-day-old newborn mice are shown in Table 1. The concentration of IRI was almost twice as high in the 1-day-old mice as in the foetal specimens. The amounts of GLI measured in the pancreatic extracts were 0·4–2% of the insulin values. A significant increase in GLI content was noted between the 18- and 21-day old foetuses (t = 2·-76; P < 0·05) as well as between the 21-day-old foetuses and the 1-day-old animals (t = 3·97; P < 0·005).

The effects of glucose on the release of GLI and IRI from isolated pancreases of foetal and newborn mice are shown in Tables 2 and 3. During 30 min of incubation 17 mM glucose had no significant effect on either IRI or GLI release from the pancreases of 18-day-old foetal mice regardless of whether 5 mM arginine was also present. Glucose (17 mM) stimulated the release of IRI from pancreases of 1- and 3-day-old mice but did not affect the release of GLI (Table 2). In the presence of 5 mM arginine the release of IRI was again markedly stimulated by 17 mM glucose whereas the simultaneous release of GLI from pancreases of 1-day-old mice was significantly depressed (Table 3).

Electron-microscopic studies on the development of the mouse pancreatic islets have disclosed granulated β-cells on the 11th–13th day of gestation (Munger, 1958; Wessels & Evans, 1968). The mouse β-cells do not stain with aldehyde-fuchsin until the 18th day of gestation (Munger, 1958; Denffer, 1969) and a-cells were not found until after birth (Munger, 1958). Data on the content and release of insulin or glucagon from the foetal mouse pancreas seem, how ever, to be lacking. In the rat, the amounts and synthesis of immunoreactive insulin have been studied in the pancreas from the 12th day of gestation onwards (Clark, 1968). In a recent report, typical a-cells were readily demonstrated in rats on the 18th day of gestation (Orci et al. 1969). Earlier gestational ages were not studied. Orci et al. also measured the pancreatic glucagon concentration in amounts comparable to those found for the mouse in the present study. It therefore is conceivable that glucagon-producing cells are also present in the late foetal mouse pancreas.

The results of the present study in the mouse are consistent with previous reports that glucose is unable to stimulate insulin release from foetal β-cells in various species (cf. Grodsky, 1970). A rapid maturation of the recognition system for glucose in the mouse is obvious from the stimulatory effect of glucose on insulin release from the pancreas of 1-day-old newborn mice. In this respect the mouse is similar to the rabbit (Milner, 1969) but somewhat different from the rat, whose β-cells do not respond to glucose until the second day after birth (Asplund, Westman & Hellerstrom, 1969).

Glucose had no effect on the release of GLI from foetal or 1- to 3-day-old animals. Similar results were obtained in a recent study on the release of glucagon from the isolated pancreas of newborn rats (Edwards, Asplund & Lundquist, 1972). In the present experiments, however, 17 mM glucose significantly depressed the release of GLI from pancreases of 1-day-old mice incubated in the presence of 5 mM arginine. This indicates that the α₂-cells of newborn mice may be sensitive to glucose. However, full glucose-sensitivity of the α₂-cells is apparently acquired later in postnatal life, since high concentrations of glucose alone inhibit the release of glucagon from isolated islets of adult mice (Chesney & Schofield, 1969) and rats (Vance, Buchanan, Challoner & Williams, 1968).

Since it was difficult to isolate islets from the foetal mouse pancreas, the present experiments had to be performed with whole pancreatic glands. This technique is handicapped by the fact that lytic factors from the exocrine parenchyma destroy insulin and glucagon released into the medium (Malaisse, Malaisse-Lagae & Wright, 1967). The degradation of insulin can be inhibited by adding either anti-insulin serum (Malaisse et al. 1967) or Trasylol ® (Malaisse, Malaisse-Lagae & King, 1968) to the incubation medium. Since the latter proteolytic enzyme inhibitor also counteracts glucagon degradation in vitro (Vance et al. 1968; Unger, Ketterer, Dupré & Eisentraut, 1967; Hazzard et al. 1968; Heding, 1970) it was added to all incubation media and immunoassay buffers. In addition, a human serum albumin without [¹²⁵I]glucagon-degrading properties was used throughout. In spite of these precautions further experiments with pancreas taken from progressively older animals revealed an almost total degradation of released GLI and, with increasing age, a less significant increase in the amounts of IRI released after glucose stimulation. The fact that glucose apparently failed to significantly inhibit the release of GLI from pancreas of 3-day-old mice incubated with arginine (inhibition was noted in 8 of 11 experiments) might be explained by increased activity of the exocrine tissue cells after birth.

Glucagon is an effective stimulus for insulin release from the isolated foetal rabbit and human pancreas (Milner, 1969; Milner et al. 1971) as well as from isolated foetal human islets (Espinosa de Los Moteros et al. 1970) and cultivated pancreatic explants of foetal rats (Lambert, Jeanrenaud & Renold, 1967). These observations led Milner (1969) to suggest that glucagon plays a role in the regulation of insulin release in utero in analogy to what has previously been suggested for adult mammals (Hellman & Lernmark, 1970). Provided that the amount of GLI measured after incubating foetal mouse pancreatic glands reflect the intrainsular release of this hormone, the present failure to alter the release of GLI by arginine and glucose does not suggest that endogenously released glucagon is important for foetal β-cell function in our strain of mice.

This work was supported by the Swedish Research Council (12 × -3923), the Swedish Diabetic Association and the Medical Faculty of Umeå. Thanks are due to Dr Lise Heding, the Novo Research Institute, Copenhagen, Denmark, for a generous supply of labelled hormones and antisera.