Production and Characterization of Reg Knockout Mice: Reduced Proliferation of Pancreatic β-Cells in Reg Knockout Mice

We constructed a targeting vector containing a Reg genomic fragment (9) (GenBank accession number D14010) interrupted in exon 2 by a neomycin resistance gene (neo) cassette with a flanking diphtheria toxin A fragment gene (DTA) cassette (10,11) (Fig. 1A). The construct was electroporated into TT-2 embryonic stem (ES) cells (Lifetech Oriental, Tokyo, Japan), and G418-resistant clones were selected. The ES cell clones were confirmed to contain a Reg-disrupted allele by Southern blot analysis of BamHI-digested genomic DNA. The cell clones were injected into ICR (CD-1; Charles Liver) mouse embryos at the eight-cell stage, and aggregates that developed from a blastocyst to the morula stage were then transferred into the uteri of pseudopregnant ICR mice as described (11). Chimeric mice were identified by their mosaic coat color and eye pigmentation. Chimeras were subsequently mated to ICR mice for germ-line transmission of the Reg replacement. Southern blot analysis of genomic DNA from the tip of the tail was used to identify mice heterozygous for the Reg mutation (Reg^+/−). Brother-sister mating of Reg^+/− mice was carried out to generate both Reg^−/− and Reg^+/+ mice. One of two Reg^+/− ES clones contributed to produce chimeric mice. The chimera transmitted the disrupted Reg allele through the germ line. Brother-sister mating of Reg^+/− animals produced viable Reg^−/− offspring. As shown in Fig. 1B, Southern blot analysis of genomic DNA using the 3′ probe, flanking the targeted region, detected a 12-kbp BamHI restriction fragment for Reg^+/+ mice. An 8-kbp fragment was detected for the targeted locus. Reg deficiency was transmitted in the Mendelian fashion. All animals were maintained in a 12-h light/12-h dark photoperiod in a humidity- and temperature-controlled room (24°C). Water and food were available ad libitum. The animals used in this study were treated in accordance with the guiding principles for care and use of research animals promulgated by Tohoku University Graduate School of Medicine, Sendai, Japan.

Genomic DNA was prepared as described elsewhere (9,11) from mouse tail, and RNA was extracted as described previously (9). Southern and Northern blot analyses were carried out as described elsewhere (9) using a ³²P-labeled mouse Reg gene probe (EcoRI fragment of mouse Reg gene: nucleotides 1361–2331 in D14010) and a 40-bp oligodeoxyribonucleotide probe complementary to the sequence of 21–25 plus 303–337 of Reg gene (D14010), respectively. Hybridization signals were scanned with a bioimage analyzer (BAS 2000; Fuji Photo, Tokyo) and compared with those of the endogenous Reg gene.

Western blot analysis was carried out as described (12) using an anti-Reg monoclonal antibody (4,13) as primary antibody.

Approximately one-third of the pancreas adjacent to the duodenum was used for histology. Tissue was fixed in 4% paraformaldehyde per 0.1 mol/l phosphate buffer overnight at 4°C and then placed in 30% sucrose. After embedding in paraffin, 10-μm sections were cut and collected on polylysine-coated slides. The sections were incubated with a monoclonal antibody directed against Reg or an antiserum directed against insulin, followed by visualization with the avidin-biotin peroxidase method (Vector Laboratories) using diaminobenzidine as a chromogen. Primary antibodies were used at the following dilutions: Reg at 1:500 (4), insulin at 1:500 (Dako), glucagon at 1:100 (ICN Pharmaceuticals, Costa Mesa, CA).

[³H]thymidine incorporation of islet DNA was measured as described elsewhere (7). In brief, pancreatic islets from Reg^+/+ and Reg^−/− mice were isolated in parallel from 8- to 9-week-old litters using density separation on a dextran gradient and cultured free-floating in RPMI-1640 medium containing 10% FCS, 2.7 mmol/l glucose, and antibiotics (100 μg/ml penicillin G and 100 μg/ml streptomycin) for 24 h to allow recovery from the isolation procedure. After this initial period, islets were transferred to 24-well culture dishes in groups of 50 islets. The islets were cultured in RPMI-1640 medium containing 2% FCS, 2.7 mmol/l glucose, and antibiotics for 48 h. During the last 24 h, the islets were cultured in the above medium, to which [methyl-³H]thymidine (Amersham Biosciences) at 10 μCi/ml was added. The islets were washed as described after the culture period and sonicated in 10 mmol/l Tris-HCl/5 mmol/l EDTA. DNA was precipitated by the addition of 7% ice-cold trichloroacetic acid and trapped by filtration on a glass-fiber disc (GF/C; Whatman). The discs were dried, and radioactivity was counted after the addition of scintillation fluid. The DNA content of the islets was measured by a fluorometric DNA assay using Hoechst 33258.

Male ICR mice received intraperitoneal injections of goldthioglucose (GTG) at a dose of 0.5 g/kg of body weight in the 6th and 8th weeks after birth. At 24 weeks after GTG injection, pancreata were fixed in 4% paraformaldehyde per 0.1 mol/l phosphate buffer overnight at 4°C for histological analysis. The islet volumes of the specimens were estimated by the point-count method as previously reported (1,7).

A rat insulin II promoter previously reported to be active in pancreatic β-cells of transgenic mice (14,15) was used. The 0.7-kbp BamHI-XmaI fragment (14,15) of the rat insulin II promoter (nucleotides −695 to +22 in ref. 16) and the 3.0-kbp PstI-HincII fragment of the entire mouse Reg gene (D14010) were ligated using XmaI-PstI linker in the correct orientation (see Fig. 3A). Three copies of BamHI-SalI fragment of the insulin-Reg hybrid gene (3,739 bp) were tandemly ligated three times with BamHI-SalI linker containing HincII site and subcloned into pBS (Stratagene, La Jolla, CA), resulting in the construction of a 14-kbp plasmid. The resultant 11.3-kbp insulin-Reg hybrid gene (Ins-Reg) was separated from the plasmid vector by BamHI and SphI digestion and was purified as described previously (14,15). The production of transgenic mice was performed as described elsewhere (14,15). In brief, a DNA solution (2 μg/ml) was microinjected into male pronuclei of fertilized mouse eggs taken from superovulated C57BL/6J × CBA/J females. The injected eggs were surgically transferred to the oviducts of ICR pseudopregnant female mice. Identification of transgenic mice was performed by PCR together with the genomic DNA isolated from mouse blood and two oligonucleotide primers (5′-AAGGAGAGATAGAAGAGAAGGGACC-3′ and 5′-CACCTTGGCTTGGAGACAGGACGAT-3′) based on the nucleotide sequence of the rat insulin II promoter (nucleotides −486 to −462) and mouse Reg gene (nucleotides 388 to 412), respectively. Strains 280 and 284 were mainly used in this study, and essentially similar results were obtained using strain 282.

NOD/Shi mice were obtained from Cler Japan (Tokyo) and were maintained by continued intercrossing under specific pathogen-free conditions in the Institute for Experimental Animals, Tohoku University Graduate School of Medicine. Ins-Reg transgenic mice were outcrossed to NOD mice, and an F1 male was backcrossed into NOD. In each generation, transgene (Ins-Reg) transmission was confirmed by PCR as described above. Ins-Reg transgenic NOD and control NOD mice were generated by intercrossing nine backcrossings (N10).

Diabetes was monitored once or twice a week by detecting the presence of urine glucose using Testape A (Eli Lilly, Indianapolis, IN) as previously reported (8,15). The onset of diabetes was determined after the appearance of glucosuria on at least two consecutive determinations. Only female mice were used for experiments on diabetes incidence.

We produced Reg-knockout mice by homologous recombination (Fig. 1A and B). Homozygote Reg^−/− mice were yielded in a distribution following Mendelian rules. Northern and Western blot analyses (Fig. 1C and D) showed no detectable Reg mRNA and Reg protein in pancreata from Reg^−/− mice, indicating that gene disruption resulted in a null Reg mutation. Reg^−/− mice, both male and female, showed normal growth and fertility. Litter sizes, urine glucose levels, and blood glucose levels were similar to those of Reg^+/+ mice. Histological analyses showed islets of normal size. The morphology of the islets in Reg^−/− mice was undistinguishable from that of the normal control mice maintained in the standard conditions described in research design and methods (Fig. 1E).

Isolation of pancreatic islets from Reg^+/+ and Reg^−/− mice showed that islet size was independent of Reg genotype. To study islet cell replication, we incubated islets in medium containing [³H]thymidine and compared [³H]thymidine incorporation into β-cell nuclei. As shown in Fig. 2A, [³H]thymidine incorporation was lower in Reg^−/− islets, as compared with Reg^+/+ islets, in both male and female mice. GTG injections were performed in Reg^−/− and Reg^+/+ mice to induce Reg expression and hyperplastic islets (2,9,17), and paraffin sections of pancreata from the GTG-treated mice were examined. The islet volume of GTG-treated Reg^−/− mice was significantly smaller than that of Reg^+/+ mice (Fig. 2B, C, and F). However, islets from both Reg^+/+ and Reg^−/− mice were densely and almost entirely stained for insulin (Figs. 2D and G). Distributions of insulin-positive/glucagon-positive cells in Reg^+/+ and Reg^−/− islets were comparable (Fig. 2D, E, G, and H). These results indicate that the insulin-producing cell mass in GTG-treated Reg^−/− mice was significantly smaller than that in GTG-treated Reg^+/+ mice. The results in Reg knockout mice together with our previous observations (7,8,13,18) suggest that Reg promotes the cell cycle progression in pancreatic β-cells.

We next designed the rat insulin II promoter/mouse Reg hybrid gene (Ins-Reg) (Fig. 3A) to target the overexpression of Reg in pancreatic β-cells in transgenic mice. The hybrid gene was microinjected into the fertilized eggs of (C57BL/6J × CBA/J) F1 mice. Five newborn mice were found to carry the Ins-Reg transgene, as detected by PCR. Founder mice were crossed with ICR mice to obtain littermates. The transgenic strains 280, 282, and 284 transmitted the transgene into their offspring and were further bred onto an ICR mouse background.

The number of copies of the integrated Ins-Reg gene was examined by Southern blot analysis (Fig. 3B) and estimated by comparison with the endogenous Reg gene in strain 280 (four copies), strain 282 (four copies), and strain 284 (30 copies). Northern blot analysis revealed that lines 280 and 284 expressed transgenic Reg gene in pancreatic islets (Fig. 3C). The Reg gene expression in other tissues such as brain, liver, kidney, and skeletal muscle was not changed in transgenic mice as compared with the nontransgenic mice (data not shown), indicating that the expression of the transgene was limited to islets. Paraffin sections of the pancreas were examined after treatment with anti-Reg and anti-insulin antibodies. The islets of the transgenic mice were densely and almost entirely stained for Reg (Fig. 3D, upper left). On the other hand, islets of the nontransgenic mice showed no immunoreactivity for Reg (Fig. 3D, upper right). Enlarged islets, which were seen in depancreatized rats that received Reg protein administration (7), were not observed in the transgenic mice. In contrast to the islets, immunostaining of the pancreatic exocrine cells of transgenic and nontransgenic mice showed a similar appearance (Fig. 3D). Both transgenic and nontransgenic islets were well stained for insulin (Fig. 3D, lower). In transgenic mice, the staining pattern of Reg was almost the same as that of insulin (Fig. 3D, left), indicating that Reg expression was limited to β-cells. The production of Reg in islets did not appear to be deleterious to the health of transgenic mice; fertility and body weight were indistinguishable from those of normal mice. The islets of transgenic mice at 21 weeks of age appeared morphologically normal and well stained for insulin (not shown).

We isolated pancreatic islets from transgenic and nontransgenic mice and cultured them in vitro for 48 h. As shown in Fig. 3E, Reg was secreted from the transgenic islets and accumulated in the medium as a 14-kDa protein. We then isolated pancreatic islets from transgenic and nontransgenic siblings and evaluated the [³H]thymidine incorporation. [³H]thymidine incorporation of the islets from both male and female transgenic mice (lines 280 and 284) was significantly higher than that of nontransgenic islets (Fig. 3F), indicating that Reg secreted from the transgenic islets acts through an autocrine mechanism to stimulate β-cell DNA synthesis. It has been consistently observed that rat Reg protein shows β-cell proliferation in culture (7,13,18) and that human pancreatic thread protein (PTP)/REG and bovine PTP/Reg show mitogenic activity to the pancreatic β-cell–derived RIN 1046-38 cells and to pancreatic duct-derived ARIP cells (19).

We next studied Ins-Reg NOD mice to test the availability of the transgene to delay or ameliorate the development of diabetes on the NOD background. The NOD mice carrying the Ins-Reg transgene showed a significantly delayed development of diabetes (Fig. 3G). In 22- to 24-week-old female mice, the islet volumes in NOD mice carrying the Ins-Reg transgene (1.98 ± 0.42 μm³, n = 6) were increased in NOD mice without the transgene (0.52 ± 0.21 μm³, n = 9), whereas the lymphocyte infiltration in islets remained unchanged. These results indicate that Reg protein, expressed in β-cells after introduction of the Ins-Reg transgene, increased the regeneration of β-cells to compensate the β-cell mass and, as a possible consequence, delayed the development of autoimmune diabetes.

In the present study, islets of Ins-Reg transgenic mice were not enlarged in vivo, whereas the isolated islets showed increased [³H]thymidine incorporation in vitro (Figs. 3D and F). Increased apoptosis coinciding with increased proliferation may occur in the Ins-Reg transgenic islets in vivo (18). The significant delay in diabetes development in the NOD mice carrying the Ins-Reg transgene (Fig. 3G) as well as the previous report that human REG protein injection inhibited the onset of overt diabetes in NOD mice (8) suggest the possible therapeutic use of Reg gene and/or Reg protein in diabetes treatment.

Studies using Reg-disrupted mice and transgenic mice clearly showed that Reg is a β-cell growth/regeneration factor. Nevertheless, despite the growth/regeneration factor deficiency, the development of pancreatic β-cells in Reg knockout mice appeared normal. Recently, several Reg and Reg-related genes have been isolated and revealed to constitute a multigene family, the Reg family (9,20–25). Based on the primary structures of the Reg proteins, the members of the family gene are grouped into four subclasses: type I, II, III, and IV (24). The mouse Reg gene described in the present article is the type I Reg gene, which is expressed in regenerative processes of islets (2,4–6,9,17,22,24,25). It remains to be determined whether the lack of Reg I protein is compensated by the expression of other Reg family members. Because accumulating evidence suggests Reg is involved in the regenerative processes in many tissues and cells other than pancreatic β-cells (22,24–30), further studies using Reg knockout mice may reveal the functions of Reg in the regeneration of a variety of tissues and cells, and such work is under way in our laboratory.

We are grateful to Dr. Satoshi Sizuta, Dr. Tomoko Tomioka-Kumagai, and Dr. Michiaki Abe for kind help; to Keiko Inaba, Emiko Shibuya, Hideo Kumagai, and Yuya Shichinohe for technical assistance; and to Brent Bell for critical reading of the manuscript. This work was supported in part by grants-in-aid from the Ministry of Education, Culture, Sports, Science, Sports and Technology, Japan, and the Research Fund for Digestive Molecular Biology and Public Trust Haraguchi Memorial Cancer Research Fund, Tokyo, Japan. T.A. is a recipient of a fellowship from the Japan Society for the Promotion of Science.