By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Sanford Children's Health Research Center, Sanford-Burnham Medical Research Institute, La Jolla, California, USA
Correspondence: Fred Levine, M.D., Ph.D., Sanford Children's Health Research Center, Sanford-Burnham Medical Research Institute, La Jolla, California 92037, USA. Telephone: 858-795-5179; Fax: (858) 795-5387; e-mail: firstname.lastname@example.org
Author contributions: E.H.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript, analyzed data, wrote the manuscript, and carried out experiments; S.-H.L.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript, analyzed data, and wrote the manuscript; F.L.: conception and design, data analysis and interpretation, manuscript writing, final approval of manuscript, analyzed data, and wrote the manuscript.
Achieving efficient β-cell regeneration is a major goal of diabetes research. Previously, we found that a combination of β-cell ablation and pancreatic duct ligation led to β-cell regeneration by direct conversion from α-cells. Here, we studied the effect of surgical reversal of the duct ligation, finding that there was a wave of β-cell replication following reversal. The combination of β-cell neogenesis prior to reversal of the duct ligation and β-cell replication following reversal resulted in efficient β-cell regeneration and eventual recovery of function. This provides an important proof of principle that efficient β-cell regeneration is possible, even from a starting point of profound β-cell ablation. This has important implications for efforts to promote β-cell regeneration. Stem Cells2013;31:2388–2395
Because decreased β-cell number is a central feature of both type I and type II diabetes, inducing β-cell regeneration has been a focus of efforts to develop a definitive treatment for diabetes. A prevalent approach to studying β-cell regeneration is to induce damage to the pancreas and then dissect the process by which regeneration occurs. Common damage models have included partial pancreatectomy [1, 2], pancreatic duct ligation (PDL) [3-5], and chemical β-cell ablation [4, 6].
An issue of considerable importance that has been inadequately addressed is whether β-cell regeneration can fully reconstitute a normal number of β cells, particularly starting from a condition of severe β-cell deficiency as occurs in type I diabetes. In large part, efforts to address that question have been confounded by the considerable controversy that exists about the mechanism and the extent to which β-cell regeneration occurs in the various damage models [1, 4, 5, 7, 8]. In fact, most damage models, including PDL and partial pancreatectomy, do not involve β-cell ablation, and in most studies in which β-cell ablation has been a feature, it has not been efficient, making it impossible to determine the maximum extent of β-cell regeneration from the very low starting point that exists in some cases, particularly type I diabetes. Efficient β-cell ablation using a transgenic mouse in which the diphtheria toxin receptor was expressed under the control of the insulin promoter has been studied, with the finding that β-cell neogenesis from α-cells occurred, but only partial β-cell reconstitution was achieved and reversion to normoglycemia took 5–10 months .
Previously, we developed a new model of pancreatic damage that combined PDL and severe β-cell ablation using high-dose alloxan . While it is now clear that β-cell neogenesis from duct cells or Ngn3-positive intermediates does not occur following PDL alone [5, 8], when we combined PDL with β-cell ablation, β-cell neogenesis from α cells was rapid and efficient. α- and β-cell replication occurred to some extent, but neogenesis and replication were completely unlinked, consistent with their being under distinct regulatory control. Because of the severe and continuing injury from PDL, it was not possible to maintain the mice for a long enough period to study the long-term outcome. In particular, we could not determine the full extent of recovery in the number of β cells or the extent to which the increased number of β cells would translate into increased β-cell function.
We hypothesized that reversing the effects of the inflammatory milieu in which the neogenic β cells existed following PDL might allow us to answer questions about the extent of β-cell recovery following PDL plus alloxan. One possibility was that β-cell regeneration would cease once the damage stimulus of PDL ended. Another possibility was that β-cell regeneration, whether in terms of β-cell number or maturation, would continue. To distinguish between those possibilities, we surgically reversed the PDL. PDL reversal (Re-PDL) resulted in rapid normalization of the exocrine pancreatic architecture and allowed us to follow the mice for a longer period to determine the ultimate effect on β-cell number and function.
Following Re-PDL, the number of β cells increased enormously. Surprisingly, this was due to a wave of β-cell replication. Over a long period, newly formed immature β cells acquired characteristics of mature β cells, including the expression of MafA, membrane-localized GLUT2, and PDX-1. Islet architecture, which was severely disrupted following PDL plus alloxan, gradually reverted to the normal state, with β cells internal to α cells. The combination of β-cell neogenesis following PDL plus alloxan in combination with the β-cell replication that occurred following Re-PDL resulted in a normal number of β cells. We conclude that, at least in young mice, β-cell regeneration from a starting point of profound β-cell deficiency is sufficiently robust to reconstitute a normal number of β cells and lead to a major improvement in β-cell function. This finding offers encouragement for efforts to induce effective and efficient β-cell regeneration.
Materials and Methods
All animal experiments were approved by the Institutional Animal Care and Use Committee of the Sanford-Burnham Medical Research Institute in accordance with national regulations.
PDL and PDL Reversal
Four-week-old male Imprinting Control Region (ICR) mice were purchased from Harlan Sprague Dawley, Hayward, CA (www.harlan.com). All Re-PDL procedures were done with ICR mice, in contrast with our previous study  that used primarily C57BL/6J mice. The choice of ICR mice was dictated by the fact that Re-PDL was technically difficult and associated with high mortality. ICR mice are larger than C57BL/6J at the same age (26–30-g vs. 22–25-g at 4 weeks of age) and gain body weight faster under normal conditions and postoperatively. On average, mice lost approximately one-third of their body weight in the postoperative period, making it advantageous to use larger mice to start. In terms of the surgical procedure, we found that C57BL/6J had more connective tissue between the abdominal organs than ICR mice, making surgery on C57BL/6J more difficult.
Mice underwent PDL plus alloxan treatment as previously described [4, 10]. One week following PDL plus alloxan, mice with blood glucose levels >500 mg/dl were divided into two groups: PDL (n = 10) and PDL followed 1 week later by Re-PDL (n = 26). Six of the ten PDL alone mice died within 2 weeks of surgery. Pancreases of sufficient quality for analysis were harvested from two mice that died or became sick and so were sacrificed per the terms of our animal care protocol at 17 and 19 days following PDL plus alloxan injection—analyzed as the PDL alone early group. Two PDL alone mice survived for 35 and 52 days following PDL and were analyzed as the PDL alone late group.
Of the 26 mice that survived for at least 1 week following Re-PDL, five had tissue harvested between days 15 and 18 following PDL and were analyzed as the Re-PDL early group. Six Re-PDL mice had tissue harvested at late times following PDL and were analyzed as the Re-PDL late group. One mouse in the Re-PDL group exhibited partial recovery of the exocrine pancreas and was analyzed separately. The pancreases from the remaining mice could not be harvested due to degradation. The day of sacrifice for the mice in the late groups is presented in the figure legends.
Re-PDL was done as follows: After anesthesia with ketamine/xylazine, sutures that were used to close the skin and abdominal muscles were removed and the incision was opened by gently stretching skin and abdominal muscle on both sides. The pancreas distal to the ligation exhibited severe exocrine atrophy, with islets being visible surrounded by translucent connective tissue and blood vessels. The suture used to ligate the duct was removed by gently untying the knot. The duct distal to the ligation was gently massaged until fluid was observed flowing antegrade through the duct. The wound was then closed.
To determine which cells replicated following PDL or Re-PDL, BrdU (1 mg/ml, Sigma-Aldrich, St. Louis, MO (www.sigmaaldrich.com)) was administered in the drinking water beginning on day 1 after Re-PDL or day 8 after PDL and continuing for 5–6 weeks. Diabetic mice were injected subcutaneously once daily with insulin glargine (Sanofi-Aventis). Insulin doses were adjusted depending on blood glucose level. After the mice had four consecutive blood glucose measurements <200 mg/dl within 1 week, an intravenous glucose tolerance test (IVGTT) was performed to test for β cell function. After IVGTT, the mice were sacrificed and the pancreas harvested. When the pancreas were harvested for histology in the Re-PDL group, the section of pancreas distal to the ligation had greatly increased in size but still appeared to be slightly smaller than prior to PDL. Compared with the proximal pancreas, it was whiter in color, allowing it to be clearly distinguished from the proximal pancreas. Using that morphological difference, the pancreas was separated into two parts, one proximal and one distal to the PDL.
For IVGTT, the mice were fasted for 12 hours, followed by penile vein injection of 0.5 g/kg Dextrose (Hospira, Lake Forest, IL (www.hospira.com)). Blood was collected via tail vein at 0, 1, 5, 10, 15, 30, 60, 90, 120, 150, and 180 minutes after dextrose injection. A small drop of blood was withdrawn from the tail vein and glucose measured using a OneTouch UltraMini glucometer.
Histology and Immunochemical Staining
Tissue was fixed in 4% paraformaldehyde (Santa Cruz Biotechnology, Inc, Santa Cruz, CA (www.scbt.com)) for 12 hours at 4°C, washed in phosphate buffered saline, followed by overnight incubation at 4°C in 30% sucrose (Sigma-Aldrich). Samples were then embedded in OCT compound and frozen at −80°C. Cryosections of 5-µm thickness were incubated with primary antibody (Table 1) overnight at 4°C. Secondary antibodies (Table 1) were added and incubated for 45 minutes. Antigen retrieval was used for BrdU and MafA staining. Nuclei were visualized by 2-(4-amidinophenyl)-1H-indole-6-carboxamidine (DAPI) staining (Sigma-Aldrich).
Table 1. Antibodies used in this study
Glucagon (Mouse, K79bB10)
Insulin (Guinea Pig)
Insulin (Rabbit, H-86)
Somatostatin (Goat, D-20)
Jackson Immuno Research
Jackson Immuno Research
Quantification of β-Cell Area and Number
All slides were scanned and analyzed using the Aperio ScanScope XT system (version 10, Aperio Technologies, Vista, CA (www.aperio.com)). Insulin-positive area per section was calculated by dividing the area of insulin-positive area by the pancreas area. The number of insulin-positive cells was calculated by counting the number of DAPI-positive nuclei surrounded by insulin-positive cytoplasm. The number of insulin-BrdU double-positive cells was calculated by counting the number of BrdU stained nuclei surrounded by an insulin-positive cytoplasm.
Data are presented as mean ± SEM. The statistical significance of the differences between groups was analyzed by Student's t test. The Kaplan-Meier curve was analyzed by log-rank test.
Reversal of PDL Led to Rapid Recovery of the Exocrine Pancreas
PDL reversal (Re-PDL) was performed by untying the suture used to ligate the pancreatic duct 1 week after PDL plus alloxan. While PDL led to complete (compare Fig. 1A, 1B) and irreversible (Fig. 1D) elimination of the exocrine enzyme amylase expression,, Re-PDL led to rapid (Fig. 1F) and almost complete (Fig. 1H) recovery of the exocrine pancreas (quantified in Fig. 1M). In neither PDL nor Re-PDL animals was there any effect on the exocrine pancreas proximal to the ligation (Fig. 1A, 1E, 1C, 1G).
To monitor for cellular replication following Re-PDL, BrdU was administered in the drinking water beginning 1 day after reversal (or the analogous time for the controls) and continuing for 5–6 weeks. Unexpectedly, the acinar cells present after Re-PDL were composed of a mixture of BrdU-positive and BrdU-negative cells. This indicates that many acinar cells did not arise by replication of an acinar cell precursor, but rather from pre-existing acinar cells that had lost amylase expression.
β-Cell Neogenesis Following PDL and Re-PDL
Consistent with our previous study of β-cell neogenesis following PDL plus alloxan , large numbers of cells coexpressing α- and β-cell markers, most notably glucagon and insulin, were found within 2–3 weeks following PDL plus alloxan (yellow cells in Fig. 2B, 1F, Supporting Information Fig. S1). There was almost no colocalization of insulin and glucagon at later times following PDL or Re-PDL (Fig. 2D, 2H), indicating that β-cell neogenesis occurred in a brief wave.
Also consistent with our previous results, β-cell regeneration did not occur proximal to the ligation in either PDL or Re-PDL animals (Fig. 2C, 2G) (Fig. 3C, 3G, 3K, 3O) (quantified in Fig. 3R, 3S). This indicates that the stimulus for β-cell neogenesis was local, that is, restricted to the pancreatic tail, the area distal to the PDL. This is in contrast to β-cell replication following partial pancreatectomy, where the effect of partial pancreatectomy on β-cell replication was equal throughout the pancreas, regardless of the distance from the site of the surgery .
β-Cell Replication and Normalization of β-Cell Area Following Re-PDL
At later times following Re-PDL, there was a dramatic increase in the number of β cells compared with PDL (Fig. 2H vs. 2D) (Fig. 3H vs. 3D, 3P vs. 3L, quantified in 3Q, 3R). Since there was no evidence of continuing β-cell neogenesis in the form of cells coexpressing multiple hormones, we examined β-cell replication as a possibility. Most of the new β cells shortly after Re-PDL were negative for BrdU (Fig. 3N, quantified in 3S), consistent with their having arisen by neogenesis rather than by replication of the rare β cells that survived alloxan treatment. However, at late times following Re-PDL, there was an enormous increase in the number of BrdU-positive β cells (Fig. 3H, 3P, quantified in 3S). Thus, an unexpected wave of β-cell replication occurred following Re-PDL.
Islet Structure and β-Cell Area Returned to Normal Following Re-PDL
A conspicuous feature of islets immediately following β-cell ablation and regeneration is a disorganized structure in which the different islet cell types are intermingled rather than segregated with β cell internal to the α and δ cells (Fig. 4A, 4B). With the large increase in the number of β cells following Re-PDL and the opportunity to follow the animals for a prolonged period, it was of interest to determine whether the islet structure returned to normal. In contrast with the disorganized structure early after PDL and Re-PDL, by 10 weeks following Re-PDL the islet structure had begun to normalize (Fig. 4C), and essentially normal islets were found at later times (Fig. 4D).
To determine the extent to which a combination of β-cell neogenesis and β-cell replication could reconstitute a normal number of β cells, we measured the insulin-positive area as a fraction of the distal or proximal pancreatic area on each section. Strikingly, the β-cell area following Re-PDL was indistinguishable from normal mice that had not undergone surgery or β-cell ablation (Fig. 4E), demonstrating virtually complete β-cell regeneration from almost complete absence following high-dose alloxan.
β Cells Gradually Matured Following Re-PDL
Previously, we found that β cells that were present following PDL plus alloxan had an immature phenotype . This appeared to be due to selective survival of a pre-existing population of immature cells that survive high-dose alloxan due to the absence of GLUT2, which is required for alloxan to enter the cell . Similar to our previous results , insulin-positive cells in the early period following PDL or Re-PDL lacked detectable levels of GLUT2, MafA, or PDX-1 (Fig. 5). Following Re-PDL, the β cells gradually matured, acquiring all three of those markers of functional β cells (Fig. 5H, 5P, 5X).
A mouse in which Re-PDL was only partially successful as determined by limited exocrine regeneration (Supporting Information Fig. S2A--S2C) was instructive. It failed to exhibit maturation of the surviving β cells, which remained negative for PDX-1 (Supporting Information Fig. S2D, S2E), MafA (Supporting Information Fig. S2F, S2G), and GLUT2 (Supporting Information Fig. S2H, S2I). β-Cell regeneration was restricted to areas where acinar regeneration took place (compare Supporting Information Fig. S2B and S2C, high-power view of regenerating islets inset), suggesting that the pancreatic milieu following PDL is unfavorable for β-cell maturation.
Mice with Re-PDL Exhibited Improved Glucose Homeostasis and Survival
Given the maturation of β cells as determined by the acquisition of GLUT2, MafA, and PDX-1 in mice that underwent Re-PDL, we performed a time course analysis of the blood glucose concentration. As we found previously, the control PDL plus alloxan mice remained severely diabetic (red line in Fig. 6A) and none survived past day 50 (red line in Fig. 6B). However, the blood glucose concentration in mice that underwent Re-PDL began to drop immediately following Re-PDL (black line in Fig. 6A). Mice that underwent Re-PDL had a highly statistically significant survival advantage (Fig. 6B). The mouse that had incomplete Re-PDL (Supporting Information Fig. S2) survived until day 150, but did not revert to normoglycemia (green line in Fig. 6A). No long-term surviving mice were found in which acinar cell recovery did not occur. To determine the extent to which β-cell function was completely restored, mice were subjected to IVGTT. Glucose tolerance was substantially improved but not normal (Fig. 6C).
The principal finding of this study is that a combination of neogenesis and replication is capable of completely reconstituting a normal number of β cells from a baseline in which there were very few pre-existing β cells. This is an important proof of principle that β-cell regeneration can be an effective therapy for diabetes, potentially even in settings where the number of β cells is close to or at zero .
There are relatively few studies of β-cell regeneration in animals that underwent severe β-cell ablation  and none that we are aware of achieved complete β-cell reconstitution. Previously studied damage models induced β-cell neogenesis or β-cell replication, but not both. In our model, regeneration appears to have occurred in two distinct phases; an initial burst of β-cell neogenesis from α cells followed by a wave of β-cell proliferation following Re-PDL. The initial burst of β-cell neogenesis in response to PDL plus alloxan recapitulated our previous study , but the increased replication of islet endocrine cells following Re-PDL was surprising.
The initial impetus for trying to reverse PDL was to determine whether that would lead to a return to normoglycemia, and if so in what time frame. A previous study of severe β-cell ablation found that return to normoglycemia took 5–10 months, but it was unclear whether that was because continuing regeneration was required or because maturation was slow. Following Re-PDL, the blood glucose concentration began to drop immediately but slowly, along with the acquisition of markers of mature β-cells. Because of the wave of β-cell replication, we could not determine the extent to which maturation versus replication accounted for the return to normoglycemia.
In summary, beginning from a starting point of profound β-cell ablation, and in which the few surviving β cells were immature, we have succeeded in achieving complete restoration of β-cell number and islet morphology and good restoration of β-cell function. This finding is encouraging and should provide further impetus for efforts to induce β-cell neogenesis and replication with pharmacologically relevant approaches that can be clinically translated.
We thank Dr. Ron Piran of SBMRI for providing sections of normal mouse pancreas and for helpful discussions, and Robbin Newlin of the SBMRI Histology Core for help with slide preparation and Aperio imaging. This work was supported by the Sanford Children's Health Research Center and BetaBat (in the Framework Program seven of the European Community). This article does not involve the products or services of a commercial interest.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.