Differentiation, reduction, and proliferation of pancreatic β-cells and their regulatory factors

Abstract

The prevalence of diabetes has increased rapidly in recent years, and many types of therapeutic agents have been developed. However, the main purpose of these drugs is to lower blood glucose levels, and they are not fundamental solutions. In contrast, our research has been aimed at stimulating and inducing β-cell proliferation in vivo and replenishing β-cells. We demonstrated that pancreatic ductal cells are a source of β-cells both after birth and during regeneration after partial duct ligation: cell lineage tracing showed that 39% of growing islets and 50% of adult islets during tissue regeneration contained β-cells differentiated from duct cells. We also examined the factors contributing to β-cell depletion. Insulin and cyclin A genes are tightly regulated by transcriptional activators and repressors, and we found that imbalanced and excessive levels of repressors result in a drastic reduction of insulin and β-cell numbers, leading to severe diabetes. Thus, we searched for factors that induce β-cell proliferation in vivo. In our transgenic (Tg) mice, there was a sex difference in the progression of diabetes and sex steroid hormones were shown to contribute to this. Surprisingly, in diabetic male Tg mice, modulation of sex steroid hormones under certain conditions resulted in a marked increase of β-cells. We identified Greb1 as a factor inducing β-cell proliferation in response to a rapid elevation of E2 levels. This series of studies has demonstrated that islet cells exhibit plasticity and indicates that changes of islet cell mass and function are dynamic and recoverable.

Introduction

The prevalence of diabetes has increased rapidly in recent years, and more than 500 million people (aged 20–79 years) worldwide have diabetes [1]. Many types of therapeutic agents have been developed (see Reference [2] for a review). For example, some approaches aim to stimulate insulin secretion from remaining pancreatic β-cells, some aim to increase the insulin sensitivity of muscle and fat cells, and some aim to inhibit glucose absorption and production in the kidney, intestine, and liver. However, all of these are primarily aimed at lowering blood glucose levels, and are not fundamental solutions. There is an urgent need for new treatment strategies because diabetes is a progressive disease leading to serious complications, resulting in ballooning medical costs and greatly reduced quality of life for patients [3–7].

Characteristic changes in the quantity, phenotype, and function of β-cells occur during the progression of diabetes [8]. Indeed, in islets of diabetic patients, there is a decrease in the number of β-cells (insulin-producing, leading to lower glucose levels) due to cell death and oxidative stress, while the number of α-cells (glucagon-producing, leading to higher glucose levels) is increased [9–12]. The changes in β-cells and α-cells and their effects have been well studied [13–15].

In recent years, researchers worldwide have attempted to identify β-cell stem cells and to differentiate β-cells from ES cells, other cells, and stem cells in vitro to compensate for the decrease of β-cells in diabetes [16, 17]. Furthermore, research on the regeneration and proliferation of β-cells themselves in vivo holds promise. Both approaches share the same goal of replenishing deficient β-cells (Fig. 1), but the latter approach has the major advantage of not requiring transplantation. However, it is not easy to induce β-cell proliferation in vivo, and the relevant mechanisms are largely unknown. This review outlines the background and the results of research focusing on postnatal pancreatic β-cell differentiation, reduction, and proliferation, as well as the regulatory factors of these processes.

Fig. 1.

Two ways to replenish β-cells. (1) “Replenish β-cells from outside the body”. Differentiation and proliferation of β-like cells in culture (in vitro). Islets can also be isolated from the pancreas of humans or genetically modified animals (ex vivo). These cells are transplanted using conventional islet transplantation methods or new techniques such as cell sheets. (2) “Replenish β-cells in the body”. Stimulation and induction of β-cell proliferation in vivo. Replication, neogenesis and trans-differentiation can be included. Both research fields share the same goal of replenishing deficient β-cells, but the latter approach has the major advantage of not requiring transplantation. However, it is not easy to replenish β-cells by inducing proliferation in vivo, as the mechanisms involved are largely unknown

Tissue stem cells and differentiation of β-cells

In postnatal humans, mice, and rats, β-cells increase with body weight gain [18, 19]. However, the estimated number of β-cells generated by cell division alone is much lower than the number of β-cells actually present. This indicates that β-cell proliferation during postnatal growth is not due to cell division alone, but may be due in part to formation of β-cells from other cells, i.e., stem cells [20]. Candidate stem cells include not only cells within the pancreas (pancreatic duct cells, adenocytes, α-cells, β-cells, and Nestin-positive cells), but also bone marrow cells, liver cells and small intestine cells [16, 17]. In particular, pancreatic ductal epithelial cells have long been considered to be stem cells, serving as a progenitor cell pool [20]. Indeed, islets exist adjacent to pancreatic ductal epithelial cells, and budding (neo genesis) of insulin- and glucagon-positive cells from pancreatic ductal epithelial cells has been observed [21]. It has also been reported that long-term culture of human pancreatic ductal epithelial cells, ligation of mouse pancreatic ducts, and gene transfer into pancreatic ducts can yield insulin-positive cells [22–24]. These findings suggest that stem cells acting as a source of β-cells are likely to be present among pancreatic ductal epithelial cells, but so far direct proof is lacking due to technical limitations of cell collection methods.

We hypothesized that some β-cells might be supplied and replenished by pancreatic duct epithelial cells after birth. To directly prove the differentiation of pancreatic ductal epithelium cells into β-cells, we set out to genetically label the ductal epithelial cells themselves and track them (cell lineage tracing) (Fig. 2). In this experiment, the Cre-LoxP system was used to ensure that the labeling is not lost even after the cells have differentiated into β-cells and lost the characteristics of pancreatic ductal epithelial cells. To ensure this, we first searched for genes that are expressed in pancreatic ductal epithelial cells from fetal to adult stages, but not in β-cells. We focused on carbonic anhydrase (CAII) and determined the location and onset of expression of the CAII gene in the pancreas [25]. As shown in Fig. 2, we generated two types of transgenic (Tg) mice expressing Cre or Cre-ER under the human CAII gene promoter and crossed them with Rosa reporter (R26R) mice to obtain double Tg mice. In these double Tg mice, all pancreatic ductal epithelial cells and cells derived from pancreatic ductal epithelium are labeled with β-galactosidase (β-gal). If labeled cells are present among β-cells, this would show that these β-cells are of pancreatic ductal epithelial origin. Conversely, if labeled cells are not present in the β-cells, then the pancreatic ductal epithelial cells do not serve as β-cell stem cells. We found that 39% of growing islets and 50% of adult islets during tissue regeneration (after partial duct ligation) contained labeled β-cells derived from pancreatic duct epithelial cells [26–28]. Furthermore, when PDX-1, a transcription factor that promotes β-cell maturation, was deficient in pancreatic ductal epithelial cells, some β-cells in islets remained immature [29]. After the publication of our work, several groups reported similar studies using the Sox9 and HNF1β genes and concluded that new β-cells derived from the pancreatic duct are found only in the fetal period, but not after birth or partial duct ligation [30–32]. However, in 2016, another group using the Sox9 gene as a marker showed that new β-cells are in fact derived from the pancreatic duct under certain conditions after birth and after partial pancreatic duct ligation [33]. Furthermore, inducible-Sox9 and HNF1β mice have been reported to have very low recombination efficiency of the Cre-loxP system during the neonatal period (E18.5 to postnatal day 5), and expression of Sox9 and HNF1β is heterogeneous in the pancreatic ducts depending on the size of the duct, suggesting that the expression characteristics of these genes used for genetic labeling of pancreatic ductal epithelial cells are very different [34]. Overall, therefore, the evidence supported our hypothesis that pancreatic ductal epithelial cells can give rise to β-cells after birth.

Fig. 2.

Cell lineage tracing of pancreatic duct cells. To directly establish the differentiation of pancreatic ductal epithelium into β-cells, we generated two types of Tg mice expressing Cre or Cre-ER™ under the human CAII gene promoter and crossed them with Rosa reporter mice (R26R) to obtain double Tg mice. In double Tg mice, all pancreatic ductal epithelial cells and cells derived from duct cells are labeled with β-galactosidase (β-gal). We found that 39% of growing islets and 50% of adult islets during tissue regeneration (after partial duct ligation) contained labeled β-cells derived from duct epithelial cells

β-Cell decrease and development of diabetes

Per investigation, one of the causes of pancreatic β-cell decrease might be dysregulation of transcription of the insulin gene. Four CREs (cAMP-responsive elements) are located in the upstream region of the human insulin gene, and disruption of this region significantly reduces the transcriptional activity of the insulin gene [35]. Therefore, we studied transcription factors that directly bind to this region and regulate transcription, and focused on the highly active CREM family. At that time, activators were the focus of attention and being cloned, but it was thought that repressors might also be important, just as a car needs both a gas pedal and a brake. We identified several transcription activators and repressors, all of which are expressed in pancreatic islets, and clarified the mechanisms by which these factors compete and interact directly with basic transcription factors IID components, such as TAF_II130 and TATA-box-binding protein (TBP), to regulate insulin gene transcription [36]. Among them, the repressor ICER in particular competes with activators and strongly represses transcription of the insulin and cyclin A genes [37]. Notably, ICER expression was increased in diabetic islets [38]. Thus, we hypothesized that increased repressor levels in β-cells may contribute to the decreased insulin production and β-cell decrease. To test this idea, we generated transgenic (Tg) mice that highly express ICER specifically in pancreatic β-cells (ICER-Tg). Indeed, ICER-Tg mice had significantly reduced insulin production and β-cell proliferation, developed severe diabetes early in life [39], and later developed diabetic nephropathy [40–44]. These studies confirmed that the combined action of repressors and activators strictly regulates insulin production and β-cell mass.

Environment and β-cell regeneration

In experiments with diabetic mice, blood glucose levels fluctuate greatly depending on diet and environmental factors, even when the mice have the same genetic background. Small differences in the amounts of nutrients in the standard diet can increase blood glucose levels in healthy mice and have long-term effects on their pups. In addition, diabetic mice consume different quantities of food depending on their living environment (housed alone or in groups), so any experiment requires careful consideration of both diet and environment [45].

When blood glucose levels recover from hyperglycemia to normoglycemia, i.e., when the β-cell environment returns to normal, do β-cells increase from their depleted state? To answer this question, we conducted long-term intervention experiments with Insulin Detemir (a long-acting insulin analog), an SGLT2 inhibitor (canagliflozin), and islet transplantation in hyperglycemic streptozotocin (STZ)-induced diabetic mice and ICER-Tg mice. Both methods were successful in controlling blood glucose and maintaining stable blood glucose levels in the normal range for a long period of time (10 or 48 weeks or more), but insulin analogs and SGLT2 inhibitor did not lead to any increase of β-cells [42, 43, 46]. In contrast, islet transplantation increased β-cells, restored islet structure, and promoted both β-cell neo genesis and proliferation. Thus, the possibility of β-cell regeneration from a state of depletion was demonstrated. So, we next asked, what other stimuli besides insulin can cause an increase of β-cells?

β-Cell proliferation mediated by sex steroid hormone

During our daily observation and study of diabetic ICER-Tg mice since 2000, we noticed a marked gender difference in the development of diabetes and diabetic nephropathy in these mice [47]. Only males maintained severe hyperglycemia and developed diabetic nephropathy, while in females, blood glucose levels gradually decreased to near normal with age and the animals did not develop diabetic nephropathy [40–47]. Similarly, in STAT3- and IRS-2-deficient mice, hIAPP-Tg mice, and various other rodent models (Akita, NSY, yellow-KK, or ob/ob mice, or Wistar-fatty, Zucker-fatty, or OLETF rats), females develop normoglycemia or milder diabetes with age [48–52]. This may be the reason why most experimental diabetes studies to date have used male animals. In humans, on the other hand, the difference is less pronounced, and it has generally been assumed that the prevalence and incidence of diabetes are similar in males and females, which would imply that the ratio of the number of adult male and female diabetic patients should be approximately 1. However, several studies have analyzed changes of the ratio by age group in detail, and have found that there is actually a rapid increase in the incidence of diabetes in postmenopausal female patients [53–58]. In other words, there are gender differences, but these are masked when the total numbers alone are compared.

Based on these facts, we speculated that the amount and ratio of circulating sex hormones (17β-estradiol (E2) and androgens) might influence susceptibility to diabetes. To test this idea, we conducted experiments with different amounts and ratios of sex steroid hormones. Surprisingly, when the sex steroid hormones of diabetic male ICER-Tg were regulated in a certain range, β-cells increased markedly, and the β-cell area of male ICER-Tg became comparable to that of female ICER-Tg mice (Fig. 3) [59]. In addition, there was a marked increase in insulin granules in individual β-cells, leading to normalization of blood glucose levels, increased glucose uptake in skeletal muscle cells [60], and recovery from diabetic kidney damage [42]. Based on these results, we hypothesized that altering the ratio of sex hormones by supplementation with high levels of E2 would be effective in lowering blood glucose levels in severely diabetic male mice. To test this idea, islets undergoing proliferation were isolated from the treated mice and their gene expression profiles were analyzed. As a result, we finally identified Greb1 as a factor that induces β-cell proliferation [61]. Greb1 was expressed in response to rapid elevation of E2 levels and induced proliferation of depleted β-cells, leading to complete normalization of blood glucose levels (Fig. 3). This series of studies revealed that sex steroid hormones play an important role in β-cell proliferation.

Fig. 3.

β-Cell proliferation mediated by sex steroid hormone. In a series of experiments, we found that altering the ratio of sex hormones by supplementation of high concentrations of E2 was effective in lowering blood glucose levels of severely diabetic male mice. Greb1 was expressed in response to rapid elevation of E2 concentration, and induced β-cell proliferation

Summary

Our findings can be summarized as follows as illustrated in Fig. 4. 1) The pancreas contains tissue stem cells, and some pancreatic ductal epithelial cells can give rise to β-cells, which are supplied to islets as needed. 2) In a healthy state, insulin and cyclin A genes are tightly regulated by transcriptional activators and repressors. However, imbalanced and excessive levels of repressors drastically reduce insulin levels and β-cell proliferation (β-cell mass), leading to the development of severe diabetes. 3) There are sex differences in the onset and progression of diabetes, and sex steroid hormones are involved. Greb1 is expressed in response to a rapid elevation of E2 levels, and induces proliferation of β-cells. Overall, our findings indicate that islet cell plasticity, islet cell mass and islet cell function are dynamic and recoverable. These findings open up the possibility of new treatment strategies to cure diabetes. However, many features of islet and β-cells and their regulation remain unexplored. Our research in this area is continuing.

Fig. 4.

Summary of our findings. (i) The pancreas contains tissue stem cells, and some pancreatic ductal epithelial cells can give rise to β-cells, which are supplied to islets as needed. (ii) In a healthy state, insulin and cyclin A genes are tightly regulated by the combined action of transcriptional activators and repressors. However, an imbalanced and excessive levels of repressors dramatically decrease insulin production and β-cell proliferation (β-cell mass), and severe diabetes develops. (iii) There are sex differences in the onset and progression of diabetes, and sex steroid hormones are involved. Greb1 is expressed in response to a rapid elevation of E2 levels and induces proliferation of depleted β-cells. Overall, our findings indicate that islet cell plasticity, islet cell mass and islet cell function are dynamic and recoverable

Data availability

Data availability is not applicable to this article as no new data were created or analyzed in this review.

Declarations

Conflict of interest

AI declares no conflict of interest.

Human or animal rights

This article does not contain any studies with human or animal subjects.