Kulkarni, R.N., Bruning, J.C., Winnay, J.N., Postic, C., Magnuson, M.A., Kahn, C.R.  (1999) Tissue Specific Knockout of the Insulin Receptor in Pancreatic b Cells Creates an Insulin Secretory Defect Similar to that in Type 2 Diabetes.  Cell, 96, 329-339.

This paper investigates one of the aspects of b cell function that may be responsible for some of the biochemical dysfunctions associated with non-insulin dependent (type 2) diabetes.  The method of Cre-loxP mediated recombination was used to assess the role of the b cell insulin receptor in type 2 diabetes. By deleting one of the exons of the insulin receptor gene through the Cre-loxP method, Kulkarni et. al tested the effects of removing the insulin receptor from b cells.

The investigators created mice that lacked the b cell insulin receptor through homologous recombination and the Cre-loxP method.  Figure 1a displays the constructs used to surround exon4 of the insulin receptor to create a “floxed gene” in embryonic stem cells.  The chimeric mice were bred with wild type mice, resulting in heterozygous IRlox (exon4 surrounded by loxP sites) mice.  Their construct is represented by the “Type 2 Deletion” in Figure 1a.  Separate Cre transgenic mice were created to express Cre only in b cells of the pancreas, by inserting nuclear localization sequence-modified Cre downstream from the rat insulin 2 promoter.  Figure 1b shows immunofluorescence of two islets from pancreas cells.  The islets appear green with red circles in the center.  The green color indicates insulin-producing b cells, while the red color indicates that Cre is being expressed.

Mating of IRlox mice carrying the Cre transgene with heterozygous IRlox mice produced the experimental group, bIRKO.  These mice contain the Cre gene inserted at the rat insulin 2 promoter and are homozygous for the floxed insulin receptor exon4.  The mating also produced three groups of littermates to serve as controls:
        homozygous IRlox – rule out any effects of loxP sites on the insulin receptor gene
        wt-Cre+ – rule out effects of Cre on b cell function
        wt – wild type controls
Although statistics on birth ratios are not included, the survival of all four genotypes suggests that Mendelian ratios were probably seen in the offspring.

Figure 2 represents some of the tests to confirm that exon4 was deleted in bIRKO mice.  Panel A shows the results of PCR analysis of DNA from bIRKO and IRlox islets that was tested for the presence of exon4.  IRlox samples from the pancreas and other tissues showed only the band at 300 bp.  The bIRKO mice showed normal bands in outside tissue, but islet samples showed bands at 220 bp.  Although the base pair lengths are not given for the insulin receptor gene, we are told that 300 bp corresponds to intact receptor DNA and 220 bp corresponds to receptor DNA that lacks exon4.  A band is also visible for the bIRKO group at 300 bp when 25 islets were sampled.  Panel C explains that this band is due to the presence of other cells mixed in with b cells in the large islet sample.  When the cells are separated by flow cytometry into b and non-b, the non-b band is at the same location as the IRlox band, while the b band for bIRKO mice is located lower on the gel.  RT-PCR was also used to confirm that the cell was producing mRNA for the knockout gene (Panel B).  Both groups exhibit a band at 480 bp, probably due to insulin receptor mRNA from non-b cells.  Only the bIRKO mice demonstrate a band at 220 bp, which probably represents truncated mRNA that does not include the exon4 sequence.  Thus, the bIRKO mice are undergoing normal transcription of the gene.  The band at 220 in the bIRKO1 sample may fainter than that for bIRKO2 because of mRNA degradation.  The Western blot seen in panel D indicates that insulin receptors are expressed normally in tissue besides pancreatic b cells.  Thus, the loxP sites do not affect expression, nor is Cre being expressed in any of these tissues.

Figures 3-7 give the results of various physiological tests performed to determine if deletion of part of the insulin receptor gene from b cells can serve as an adequate model of type 2 diabetes.

The data in Figure 3 suggests that blood glucose levels were not significantly different for 6 month old wild type, homozygous IRlox, Cre or bIRKO mice after 14 hour fasting or random feeding for males or females.  These results suggest that glucose homeostasis is maintained in bIRKO as compared to normal mice.  Also, the presence of lox or Cre in the DNA does not affect glucose homeostasis (negative controls).  However at 6 months, the fasting levels of insulin are significantly increased (hyperinsulinemia) in both male and female bIRKO mice as compared to the other three control groups.

Figure 4 shows the results of insulin release from pancreatic b cells after intraperitoneal (IP) injections of either glucose or arginine into 2 month old male and female mice.  Panels A and B show the effects of glucose injections.  Injections into the three control groups produced an initial spike after 2-3 minutes, followed by a drop and then a gradual increase.  The bIRKO mice, however, did not demonstrate the peak at 2-3 minutes, but did show a gradual increase in insulin levels.  The levels were significantly lower for the first 5 minutes after injection of glucose, but approached the same levels as the control groups after 30 minutes.  Panels C and D show the insulin response after IP injections of arginine.  No significant difference was found between groups at any time after injections.  The insulin levels spiked at 2-3 minutes and declined to slightly higher than the initial level after 30 minutes.

Figure 5 shows the glucose levels in 2, 4, and 6 month old mice in the minutes after IP glucose injections.  At all ages, the bIRKO mice showed significantly increased glucose levels compared to the three control groups.

Figure 6A shows pictures of immunostained pancreatic sections at 2 and 4 months.  The islets for bIRKO mice are similar in size at 2 months to control animals, but appear to have decreased in size after 4 months.  Figure 6B contains plots of pancreatic insulin concentration in control and bIRKO mice at 2 and 4 months.  Both groups have similar concentrations at 2 months, but at 4 months, the levels have increased for control animals, while they stayed the same for bIRKO mice.  The difference between the bIRKO mice and the controls was significantly lower.

Figure 7a is an electron microscopic picture of b cells in isolated islets.  No apparent differences can be seen between IRlox and bIRKO mice.  Figure 7b shows the results of immunostaining for Glut2, a glucose transporter.  The region stained is smaller in bIRKO than in IRlox mice.  Ob/Ob mice, however, show only a few, faint spots.

The results of this study indicate that deletion of exon4 of the b cell insulin receptor by the Cre-loxP method results in abnormal insulin response to a glucose challenge.  This response is indicated by lower-than-normal insulin levels (Figure 4) and higher-than-normal glucose levels (Figure 5) following IP injections of glucose.  According to the authors, the loss of first-phase insulin secretion seen in the bIRKO mice is similar to early symptoms in type 2 diabetes. Immunostaining demonstrated that the receptor deletion did not affect any visible aspects of pancreas islet morphology.  The experiments in this study contained good controls by using used wild type, IRlox and Cre mice to make sure that these recombinants do not have any effects on b cell activity.

Future experiments could be carried out to determine how the interaction between glucose and the insulin receptor affect insulin secretion.  A phage display peptide library kit or the yeast two hybrid approach could be used to find the sequence of the next protein that interacts with the insulin receptor.  Cre-loxP deletion of the coding sequence of the next protein could indicate whether the insulin receptor itself is responsible for the physiological changes, or if it is the failure to activate a protein further along in the pathway.  Similar tests to the ones described in this paper could be performed to assess insulin and glucose levels.  The Cre-loxP method could also be used to delete the immune receptor in other pancreatic cells to see if similar physiological effects are observed.

back to my home page

© Copyright 1999, Department of Biology, Davidson College, Davidson, NC 28036
Send comments, questions, and suggestions to: lafreeman@davidson.edu