Main

The pdx-1 -encoded homeodomain protein in mammals (STF-1, IPF-1, IDX-1)4,5,6 was isolated as a transcriptional regulator of insulin and somatostatin7,8,9. The protein was first detected in the embryonic pancreatic and duodenal endoderm. But in the pancreas, pdx-1 expression becomes progressively restricted to islets, where it is produced in more than 90% of β-cells, and in substantially fewer δ-cells (15%) and α-cells (3%)10,11.

Mice that are heterozygous (+/−) for pdx-1 develop normally, but in pdx-1 homozygotes (−/−), the branching outgrowth of the pancreas that usually occurs is arrested at an early stage11,12. The relevance of these findings is underscored by a description of a human phenotype lacking a pancreas and associated with a mutation in the pdx-1 gene13,14. Maturity-onset diabetes occurred in humans heterozygous for this mutation13; this prompted us to examine whether pdx-1 is important for glucose homeostasis in an adult mouse model.

We fasted pdx-1 wild-type and +/− mice for 14-16 hours and injected them intraperitoneally with 20% glucose (2 grams per kilogram body weight). The blood glucose levels of the wild type underwent a threefold increase within 15 minutes but returned to baseline two hours later (Fig. 1a). By contrast, +/− mice showed a 7-10-tenfold increase in blood glucose levels after 30 minutes. These mice remained hyperglycaemic even after 2 hours (Fig. 1a).

Figure 1: Comparison of wild-type and pdx-1 -deficient mice.
figure 1

a, Glucose tolerance tests on black Swiss mice that are either heterozygous for pdx-1 (+/−, red) or wild type (green). Mean blood glucose levels over time are shown with standard error (n = 8 for wild-type mice, n = 11 for +/− mice). Animals were fasted for 14-16 hours before testing. Blood glucose levels were monitored with a glucometer. b, Representative islets from 15-week-old wild-type (top) and +/− (bottom) mice. Islets were immunostained with combined antibodies raised against glucagon, somatostatin and pancreatic polypeptide to show the mantle of islet non-β-cells around the core of unstained β-cells. Magnification bar, 50 μm.

The levels of plasma insulin following glucose administration (not shown) did not differ appreciably between wild-type and +/− animals, but these insulin levels were inappropriately low for the degree of hyperglycaemia observed in the +/− mice.

In histological evaluation, the pancreatic islets from the +/− mice appeared somewhat smaller, with a thicker mantle devoid of β-cells, than those of the wild type (Fig. 1b). The mass of β-cells was reduced, but to an extent that was not significant (wild-type mice: 2.80 milligrams ± 0.44; n = 8; +/− mice: 1.76 milligrams ± 0.13; n = 4). However, the mass of non-β-cells was almost doubled in +/− compared with wild-type mice (wild-type mice: 0.47 mg ± 0.06; +/− mice: 0.82 mg ± 0.10). This suggests that a deficiency in the pdx-1 gene may skew the islet cell lineages towards developing into non-β cells.

Our results support the idea that, as well as acting as a regulatory protein for pancreatic development, the protein encoded by pdx-1 is required for glucose homeostasis in the adult pancreas. Thus a deficiency in pdx-1 may predispose certain individuals to the development of late-onset diabetes, particularly in the context of other genetic mutations within the insulin-signalling cascade.