Introduction

Type 2 diabetes results from pancreatic beta cell failure on a background of insulin resistance. To compensate for the increased insulin demand in insulin-resistant states, beta cells augment their insulin biosynthetic and secretory capacity in part by expanding their mass through hyperplastic and hypertrophic mechanisms [1]. However, when beta cells reach their maximum compensatory response and can no longer overcome insulin resistance, overt diabetes occurs. Hence, understanding how beta cells regulate their growth could lead to the development of therapeutic approaches to prevent and/or delay the onset of this disease.

During the last few years, our group has been investigating the actions of sodium tungstate on beta cell plasticity. We have shown that this agent can normalise blood glucose levels in streptozotocin (STZ)-treated rats and that this normalisation correlates with increased beta cell proliferation and neogenesis [2–4]. Recently, in a global transcriptome analysis, we identified the gene encoding the Raf-1 kinase inhibitor protein 1 (RKIP1) as one of the potential candidate genes involved in the ability of tungstate to modulate pancreatic beta cell mass [5].

RKIP1 is an evolutionarily conserved protein that was first recognised as a regulator of the MAPK signal transduction cascade through its interaction and inhibition of Raf-1 kinase activity [6]. However, upon phosphorylation by protein kinase C, RKIP1 can dissociate from Raf-1 and inhibit G protein-coupled receptor kinase 2 (GRK2) [7], a negative regulator of G protein-coupled receptors. Additionally, RKIP1 has been implicated as a negative regulator of nuclear factor κB (NF-κB)-signalling pathways and an activator of glycogen synthase kinase 3β (GSK3β)-dependent signalling pathways [8, 9]. Therefore, RKIP1 mediates the crosstalk between various important cellular signalling pathways and, consequently, can influence a wide array of cellular functions including cell proliferation, differentiation and apoptosis. Thus, RKIP1 has been implicated in cancer, neurological functioning and reproduction (reviewed in [10–12]).

In the pancreas, RKIP1 is present at high levels in islets, whilst its presence in exocrine cells is somewhat controversial [13, 14]. The role of RKIP1 in beta cells is not fully understood. Because modulation of RKIP1 levels in the beta cell line HIT-T15 influences MAPK signalling and cell proliferation, it has been speculated that the presence of high amounts of RKIP1 in adult beta cells could partially explain the low proliferative capacity of these cells [13]. In line with this, overproduction of Raf-1 kinase in primary beta cells enhances cell proliferation [15]. Conversely, Raf-1 kinase inhibition results in increased beta cell apoptosis, indicating that modulation of this pathway may also contribute to beta cell survival [16]. Taken together, this in vitro evidence supports a possible role of RKIP1 in the modulation of pancreatic beta cell plasticity.

The aim of the present study was to investigate the involvement of RKIP1 in the regulation of beta cell mass using mice with a homozygous germ-line deletion of the Rkip1 gene (Rkip1 −/−). The phenotypes of Rkip1 −/− mice described to date include age-dependent learning and olfaction deficits and fertility defects within the males [12, 17]. Here we have characterised pancreatic morphology and glucose homeostasis in Rkip1 −/− and provide evidence of a role of RKIP1 as a negative regulator of beta cell mass under physiological (normal growth) and pathological (STZ-induced diabetes) conditions in vivo.

Methods

Mouse studies

Rkip1 −/− mice were generously provided by Dr Jan Klysik (Brown University, Providence, Rhode Island, USA), maintained in a C57BL/6 background and genotyped as described previously [17]. Protocols were approved by the Animal Ethics Committee of the University of Barcelona and the Principles of Laboratory Animal Care were followed. Diabetes was induced by intraperitoneal injection of 100 mg/kg STZ (Sigma-Aldrich, St Louis, MO, USA) in adult mice weighing 18–22 g that were fasted for 6 h. Blood glucose levels were measured weekly until the time they were killed.

For intraperitoneal glucose tolerance tests (IPGTTs), mice fasted for 6 h were injected with 2 g glucose/kg body weight. Blood samples were collected at 0, 15, 30, 60 and 120 min after injection and glycaemia was measured using a clinical glucometer and Accu-Chek test strips (Roche, Basel, Switzerland). Plasma was separated by centrifugation and kept at −80°C for insulin determination. For insulin tolerance tests, animals were injected with 0.75 U insulin/kg body weight. Blood samples were collected at 0, 15, 30 and 60 min and glucose levels were measured as indicated above.

Islet studies

Islets were isolated from 28-day-old (P28) or 8-week-old male mice by collagenase digestion. P28 islets were individually handpicked under a stereomicroscope after undergoing collagenase digestion, whereas adult islets were purified using a Histopaque gradient (Sigma-Aldrich) [18]. Insulin secretion was evaluated using freshly isolated islets in static incubation assays [19]. Insulin was measured by ELISA (Mercodia, Uppsala, Sweden).

Immunofluorescence and morphometric analysis

Pancreases from P28 and adult mice were dissected, fixed overnight in 10% formalin neutral buffered solution (Sigma-Aldrich) and embedded in paraffin. Sections (4 μm) were deparaffinised, rehydrated and treated with citrate buffer (10 mmol/l; pH 6.0) in a microwave oven. Primary antibodies used were: rabbit anti-RKIP (1:1,000; Calbiochem, Darmstadt, Germany), guinea pig anti-insulin (1:2,500; Dako, Glostrup, Denmark) and rat anti-cytokeratin 19 (DSHB, Iowa City, IA, USA). Cy2- and Cy3-labelled secondary antibodies (1:200; Jackson ImmunoResearch, Newmarket, UK) were used for immunofluorescence. Hoechst 33258 (Sigma-Aldrich) was used as nuclear marker. Images were taken with a Leica DMR HC epifluorescence microscope and a Leica TCS-SL confocal microscope (Leica Microsystems, Wetzlar, Germany).

For morphometric analysis, serial sections from three different levels for each pancreas block (six sections/animal) were stained with guinea pig anti-insulin antibody as used for immunofluorescence or with mouse anti-glucagon (1:1,000; Dako) using a modified avidin–biotin–peroxidase method and counter-stained with toluidine blue [20]. Slides were analysed using Image J (National Institutes of Health, Bethesda, MD, USA) software.

Proliferation and apoptosis measurements

For proliferation analysis, sections were stained with mouse anti-Ki67 (1:20; Dako, Denmark) as indicated above. A minimum of 3,000 insulin-positive cells per pancreas was counted in adult, P14 and P28 pancreases. For detection of apoptosis, the ApoAlert DNA fragmentation Assay Kit (Clontech, Mountain View, CA, USA) for tissue sections was used following manufacturer’s instructions.

Western blot

Protein extracts from whole pancreas and islets were prepared in lysis buffer (50 mmol/l Tris pH7.5, 5 mmol/l EDTA, 150 mmol/l NaCl, 1% (vol./vol.) Triton X-100, 10 mmol/l sodium phosphate, 10 mmol/l sodium fluoride, 10 mmol/l sodium orthovanadate and protease inhibitors). Antibodies used for western blot were: RKIP (1:6,000), phosphorylated extracellular signal-regulated kinase 1 or 2 (ERK1/2) and total ERK1/2 (1:1,000; Cell Signaling, Beverly, MA, USA), GSK3 (1:1,000; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and, as loading controls, actin (1:5,000; GE Healthcare, Hertfordshire, UK) or tubulin (1:1,000; Sigma-Aldrich).

RNA isolation and quantitative PCR analysis

Total RNA was prepared from whole pancreas using RNAgents Total RNA Isolation System (Promega, Madison, WI, USA) and from isolated islets using RNeasy Micro Kit (Qiagen, Venlo, the Netherlands). RNA was treated with DNase (Qiagen), and integrity was assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). RNA was reverse transcribed using SuperScript VILO (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s instructions. Real-time PCR was carried out in a LightCycler 480 System (Roche) using SYBR Green (Mesa Green, Eurogentec, Seraing, Belgium). The primer sequences used are presented in electronic supplementary material Table 1. Expression levels were normalised to the expression of the gene encoding TATA box binding protein (Tbp) gene and expressed relative to the mean of the values of the same gene found in control wild-type (WT) animals, set at 1.

Microarray analysis

RNA from three WT and three Rkip1 −/− pancreases was reverse transcribed, labelled and hybridised to mouse genome 430A 2.0 arrays (Affymetrix, Santa Clara, CA, USA) according to the manufacturer’s protocol. Raw data was background-adjusted, normalised and summarised by robust multi-array analysis (RMA) using the Affy package [21]. Raw and processed data were submitted to several quality controls and analysed for statistical differences as described previously [5]. Briefly, analysis was performed with the LIMMA software package available from Bioconductor (www.bioconductor.org) [22], adjusting p values using Benjamini and Hochberg’s method to control the false discovery rate at a value <5% [5]. Genes were clustered and represented in a heat diagram with dChip software (http://biosun1.harvard.edu/complab/dchip) [23]. Data have been deposited in NCBIs Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo/), accession number GSE31150.

Statistical analysis

Data are expressed as mean ± SEM and statistical significance determined by Student’s t test. A p value <0.05 was considered statistically significant.

Results

RKIP1 levels in the mouse pancreas

We characterised RKIP1 levels in the adult mouse pancreas by immunofluorescence and found that it was highly abundant in islets, where it co-localised largely, but not exclusively, with insulin (Fig. 1a). RKIP was also found in cytokeratin19+ ductal cells whereas it was not detectable in acinar cells (Fig. 1b). RKIP enrichment in islets relative to total pancreatic tissue was further confirmed by western blot (Fig. 1c). Pancreatic tissue from Rkip1 −/− animals was used to verify that both the staining (Fig. 1a,b) and the immunoblot band (Fig. 1c) detected in pancreases from WT mice corresponded to RKIP1.

Fig. 1
figure 1

Localisation of RKIP1 in the adult mouse pancreas. (a,b) Pancreases were harvested from 8-week-old male WT and Rkip1 −/− (KO) mice, and fixed and stained with the indicated antibodies. In (a), insulin is indicated in green, RKIP in red. In (b), ductal cell marker cytokeratin 19 is indicated in green, RKIP in red. Nuclei are marked with Hoechst 33258 and indicated in blue. Scale bars represent 50 μm. (c) Western blot analysis of lysates from total pancreas and isolated islets obtained from Rkip1 −/− and WT mice. Actin was used as a loading control (20 μg/lane)

Full size image

Glucose homeostasis in adult Rkip1 −/− animals

To examine the role of RKIP1 in endocrine pancreas function, we assessed whole body glucose homeostasis in 8-week-old Rkip1 −/− animals. We found no significant differences in body weight (WT 17.8 ± 2.3 g, Rkip1 −/− 19.0 ± 3.0 g), fasting blood glucose (WT 6.8 ± 0.4 mmol/l, Rkip1 −/− 7.1 ± 0.3 mmol/l) or fasting plasma insulin levels (WT 92.7 ± 10.9 pmol/l, Rkip1 −/− 80.1 ± 15.6 pmol/l) between Rkip1 −/− mice and their littermate controls. However, following an intraperitoneal glucose challenge, Rkip1 −/− mice displayed improved glucose tolerance (Fig. 2a,b), which correlated with increased plasma insulin 15 min after the glucose challenge (Fig. 2c). Rkip1 −/− mice and WT littermates showed similar sensitivity in an insulin tolerance test (Fig. 2d). Altogether, these data indicate that Rkip1 ablation results in improved glucose homeostasis and point to the pancreas as the source for this improvement.

Fig. 2
figure 2

Glucose tolerance and insulin sensitivity in Rkip1 −/− mice. (a) Blood glucose levels of 8-week-old Rkip1 −/− (KO, n = 8) and WT (n = 6) mice were measured at the indicated times after intraperitoneal delivery of glucose. (b) The AUC for glucose was calculated using the trapezoidal rule. (c) Plasma insulin levels after glucose administration in an IPGTT; values are expressed relative to time 0, which is defined as 100%. (d) Insulin tolerance test: blood glucose levels of 8-week-old Rkip1 −/− (n = 3) and WT (n = 3) mice were measured at the indicated times after intraperitoneal injection with insulin. Black squares, WT animals; white circles, Rkip1 −/− animals. All results represent mean ± SEM.*p < 0.05 for KO vs WT

Full size image

Beta cell function and mass in Rkip1 −/− animals

Since Raf-1 activity is critical for insulin secretion [16, 24, 25] and insulin synthesis [25], we examined whether Rkip1 ablation affected these functions. We found that islets isolated from 8-week-old Rkip1 −/− and WT animals exhibited similar basal and glucose-induced insulin secretion and had comparable insulin contents (Fig. 3), thus ruling out a major contribution of differences in islet function to enhanced glucose homeostasis in Rkip1 −/− animals.

Fig. 3
figure 3

Glucose-induced insulin secretion in isolated islets from Rkip1 −/− mice. Islets were isolated from adult WT (black bars) and Rkip1 −/− (white bars) mice and insulin secretion was assayed at the indicated glucose concentrations in static incubation experiments as indicated in the Methods section. (a) Insulin secretion per islet, (b) islet insulin content, and (c) insulin secretion as percentage of total insulin content. Data are the mean ± SEM for n = 3 independent experiments

Full size image

Alternatively, improved glucose tolerance may result from augmented beta cell mass [26]. To test this possibility, we performed morphometric analysis on fixed pancreatic tissue and found that fractional beta cell area (relative to total pancreatic area) was similar in Rkip1 −/− and WT mice (Fig. 4a). Similarly, optical projection tomography [27] failed to detect major differences in beta cell distribution throughout the whole pancreas between Rkip1 −/− and WT animals (ESM Fig. 1). However, Rkip1 −/− animals exhibited bigger pancreases (WT 131.2 ± 7.2 mg, Rkip1 −/− 185.1 ± 7.3 mg; p < 0.001), which resulted in a nearly twofold increase in total beta cell mass in Rkip1 −/− compared with WT mice (Fig. 4b).

Fig. 4
figure 4

Characterisation of the pancreatic beta cell compartment in adult Rkip1 −/− mice. Morphometric analysis of fixed pancreas from adult Rkip1 −/− (KO, n = 8) and WT (n = 8) mice. (a) Beta cell area calculated as insulin-positive islet area (in μm2) as a percentage of the total pancreatic area (in μm2), (b) total beta cell mass, (c) islet cell composition, (d) beta cell volume, (e) islet size distribution, (f) islet density and (g) percentage of islets located at <5 μm from ducts. WT animals are shown in black bars and Rkip1 −/− animals in white bars. Data are the mean ± SEM. *p < 0.05 and **p < 0.01 KO vs WT

Full size image

No differences in islet cell composition or beta cell volume were apparent between WT and Rkip1 −/− mice (Fig. 4c,d). Yet, Rkip1 −/− pancreases displayed significant alterations in islet size distribution, with more small islets and fewer medium and large islets in Rkip1 −/− compared with WT animals (Fig. 4e). Islet density tended to be higher in Rkip1 −/− mice than in controls (Fig. 4f), which could explain why, despite exhibiting a larger proportion of small islets, the relative beta cell area remained similar in Rkip1 −/− mice to that in WT animals. The percentage of islets located near ducts (<5 μm) was significantly increased in Rkip1 −/− animals (Fig. 4g).

Taken together, these data indicate that an augmented beta cell mass:body weight ratio, rather than alterations in beta cell functionality, is responsible for the improvement in whole body glucose control in Rkip1 −/− mice [26]. Furthermore, while the relevance of altered islet size distribution and location in the control of glucose homeostasis remains unclear, our findings unravel additional potential effects of RKIP1 on islet morphogenesis.

Beta cell proliferation in Rkip1 −/− animals

We next examined the mechanisms underlying the increased beta cell mass in Rkip1 −/− mice. Since RKIP1 has been shown to inhibit beta cell proliferation in vitro [13], we investigated whether the augmented beta cell mass in Rkip1 −/− animals was due to increased proliferation. Co-immunostaining for insulin and Ki67 in pancreases of adult animals failed to detect any significant difference in cell proliferation rates between Rkip1 −/− mice and control littermates (Fig. 5a). We then postulated that beta cell mass might have been augmented during young postnatal life, when the greatest expansion of beta cells takes place [28]. Consistent with this idea, we found that Rkip1 −/− animals exhibited a higher number of proliferating beta cells at P14 and P28 than controls of the same age (Fig. 5a). Total pancreatic insulin content was also significantly increased in P28 Rkip1 −/− mice relative to age-matched controls, and this difference was maintained in adult animals (Fig. 5b). Hence, RKIP1 appears to regulate beta cell proliferation in vivo specifically during the early postnatal period.

Fig. 5
figure 5

Beta cell proliferation in Rkip1 −/− mice. (a) Relative levels of proliferating beta cells (Ki67+/insulin+) in P14, P28 and 8-week-old Rkip1 −/− and WT pancreases (n = 3–6 animals per group). For comparison purposes, data are expressed relative to levels found in age-matched WT mice, set at 1. (b) Total pancreatic insulin content was measured by ELISA in P21, P28 and adult Rkip1 −/− mice (n = 5–6) and WT (n = 5–6) littermates. (c,d) quantitative RT-PCR for Rkip1 and Ins1 + 2 mRNAs in pancreases harvested from P14, P21 and P28 WT mice (n = 3–4). Values are expressed relative to Tbp. (e) Percentage of proliferating beta cells (Ki67+/insulin+) in pancreases from control mice at P14 and P28 (n = 3–6 animals per group). (f) Representative immunoblot for phosphorylated ERK1/2 (p-ERK1/2) and total ERK1/2 in isolated islets from P28 Rkip1 −/− (KO) and WT control littermates. Bands were quantified by densitometry and values expressed as the ratio of p-ERK1/2 relative to total ERK1/2 protein for n = 6 per genotype. Black bars, WT animals; white bars, Rkip1 −/− animals. All bars represent the mean±SEM *p < 0.05 between indicated groups

Full size image

Our previous results suggested that RKIP1 might be involved in restricting beta cell expansion specifically around the lactation–weaning transition (P21–P28). To further support this possibility, we assessed whether changes in endogenous Rkip1 expression correlated with inhibition of beta cell proliferation during this transition in control mice. We observed a stepwise rise in the abundance of pancreatic Rkip1 mRNA, accompanied by a significant drop in beta cell proliferation rates between P14 and P28 (Fig. 5c,e). Remarkably, increased Rkip1 expression was not correlated with an increase in insulin (Ins1 and Ins2) gene expression (Fig. 5d). Thus, these findings support the notion that activation of Rkip1 expression could participate in restricting beta cell expansion at the end of lactation.

To elucidate the molecular basis for the enhanced beta cell proliferation in P28 Rkip1 −/− animals we examined the phosphorylation status of the MAPKs ERK1/2, which are the canonical targets of the Raf-1/MAP kinase kinase (MEK) pathway. However, we found no significant changes in basal ERK1/2 phosphorylation in islets freshly isolated from Rkip1 −/− mice relative to those from controls (Fig. 5f). RKIP1 also enhances GSK3β-dependent signalling through stabilisation of this kinase [9]. GSK3β is involved in β-catenin turnover and WNT/β-catenin signalling has been shown to regulate neonatal beta cell expansion [29]. Accordingly, GSK3β deficiency results in expanded beta cell mass [30]. These findings prompted us to examine whether the absence of RKIP1 affected GSK3β protein accumulation. However, we found no differences in GSK3β protein or Axin2 mRNA expression (an endogenous indicator of WNT/β-catenin signalling) between P28 Rkip1 −/− and WT islets (ESM Fig. 2). Thus, although alterations in a small subset of beta cells or transient modulation of these pathways cannot be ruled out, these results indicate that neither MAPK nor GSK3β are constitutively changed in Rkip1 −/− islets.

Pancreatic growth in Rkip1 −/− animals

The finding that Rkip1 ablation affected the pancreatic endocrine and exocrine compartments in a similar manner indicates that RKIP1 might also regulate exocrine cell proliferation. This result was unexpected because of the absence of RKIP1 in adult acinar cells (Fig. 1). To investigate this issue further, we determined the time at which the increase in pancreas size was first evident in KO mice. Pancreas weights (corrected for body weights) were similar between KO and control littermates until P28 (Fig. 6a), at which time Rkip1 −/− mice showed a higher pancreas weight, thus indicating that this phenotype is not determined during embryogenesis but, as for beta cell proliferation, during young postnatal life, around the lactation–weaning transition (P21–P28). Consistent with enhanced pancreatic growth, total pancreas (endocrine + exocrine) proliferation rates measured by Ki67 staining were dramatically increased in P28 Rkip1 −/− pancreases relative to controls (Fig. 6b). In contrast, apoptotic rates as measured by TUNEL assay were negligible and similar between P28 Rkip1 −/− mice and controls (Fig. 6c). In view of these results, we wondered whether the pattern of RKIP1 distribution in the young postnatal pancreas might differ from that of the adult organ. We found that, as in the adult, RKIP1 was preferentially located in islet and ductal cells at P21 and P28 (Fig. 6d and data not shown). However, we also noticed distinctive nuclear staining in many acinar cells at these stages (Fig. 6d), thus indicating that ablation of the Rkip1 gene in postnatal acinar cells might be involved in enhanced exocrine cell proliferation in a cell autonomous manner in Rkip1 −/− mice.

Fig. 6
figure 6

Effects of Rkip1 deletion on pancreatic growth. (a) Relative pancreatic weight (corrected for body weight) at P1, P7, P14, P21, P28 and in 8-week-old Rkip1 −/− (n = 10) and WT (n = 10) mice. (b) Percentage of proliferating pancreatic cells (Ki67+) in pancreases from P28 Rkip1 −/− (KO) and WT mice (n = 6) and a representative image of a P28 Rkip1 −/− pancreas showing immunostaining for Ki67 (green) in non-insulin and insulin-positive (red) cells. Nuclei are marked with Hoechst 33258 (blue). (c) Percentage of apoptotic cells in pancreases from P28 Rkip1 −/− (KO) and WT mice (n = 6). (d) Representative images showing immunostaining for RKIP (red), cytokeratin 19 (green) and insulin (blue) on pancreatic sections from P21 WT and Rkip1 −/− (KO) animals. Black bars, WT animals; white bars Rkip1 −/− animals. All bars represent the mean ± SEM. *p < 0.05 between indicated groups. Scale bars represent 50 μm

Full size image

To elucidate the pathways responsible for increased pancreatic growth in Rkip1 −/− mice, we performed whole genome expression profiling using pancreases from P28 Rkip1 −/− and WT animals. Gene ontology analysis identified the cell cycle as the most significantly represented pathway among RKIP1-regulated genes (ESM Fig. 3). Among the genes most highly upregulated in P28 Rkip1 −/− pancreases relative to controls, we identified those encoding the four members of the superfamily of C-type lectins known as regenerating islet-derived 3: Reg3a, Pap (also known as Reg3b), Ingap (also known as Reg3d) and Reg3g. Interestingly, Reg3 genes have been shown to function as stress response genes, antiapoptotic and growth factors in the digestive system and to specifically promote beta cell proliferation in the pancreas [31]. We confirmed these data by quantitative RT-PCR and found that Reg3 overexpression was only seen after weaning (ESM Fig. 3). However, Reg3 mRNAs were not overexpressed in isolated islets from P28 Rkip1 −/− relative to WT animals (MPJ, RG unpublished observations), pointing to the exocrine pancreas as being responsible for Reg3 gene activation in Rkip1 −/− animals.

Taken together, these data reveal RKIP1 as a negative regulator of postnatal pancreatic cell proliferation. Since cell cycle genes are mainly regulated at the post-transcriptional level, the molecular signature of Rkip1 ablation and its significance in pancreatic growth will need to be further explored through the identification of changes in cell cycle protein production in Rkip1 −/− pancreases.

Rkip1 deletion ameliorates STZ-induced diabetes

Our results thus far have demonstrated that Rkip1 deletion affects beta cell proliferation in early postnatal life. To determine whether RKIP1 might be a potential target to trigger beta cell mass recovery in diabetes, adult mice were rendered diabetic by STZ injection. At 7 days post-injection, both Rkip1 −/− and WT animals treated with STZ became hyperglycaemic (WT/STZ 14.1 ± 1.4 vs WT/control 7.7 ± 0.5 mmol/l; Rkip1 −/−/STZ 11.2 ± 0.83 vs Rkip1 −/−/control: 8.7 ± 0.4 mmol/l) (Fig. 7a) and exhibited >80% depletion of Ins2 mRNA levels (Fig. 7b). The onset of hyperglycaemia and the marked loss of Ins2 mRNA indicate that beta cells are efficiently eliminated by STZ in both Rkip1 −/− and WT groups. One week later, Rkip1 −/− mice exhibited near normalisation of blood glucose levels, whereas WT animals became severely hyperglycaemic and remained diabetic until the end of the study (Fig. 7a). The recovery of Rkip1 −/− mice correlated with increased Ins2 mRNA levels at 14 days after the injection (Fig. 7b). At 28 days post-injection, STZ-treated Rkip1 −/− mice had normalised plasma insulin levels and had an increased pancreatic insulin content (Fig. 7c,d) compared with STZ-treated control mice, pointing to the partial rescue of beta cell mass in Rkip1 −/− animals and implying that beta cell regeneration participates in the restoration of normoglycaemia in these animals. In sum, these findings indicate that RKIP1 can also regulate beta cell mass in an experimental model of diabetes.

Fig. 7
figure 7

Reversal of STZ-induced diabetes in Rkip1 −/− mice. (a) Glucose plasma levels after STZ (white symbols, dotted line) or vehicle (control, black symbols, solid line) injection in WT (squares) and Rkip1 −/− (KO, circles) mice (n = 7 for each group); † p < 0.05, KO control vs KO STZ; *p < 0.05, **p < 0.01 WT control vs WT STZ. (b) Ins2 mRNA expression in whole pancreas from WT and Rkip1 −/− mice 7-days and 14-days post STZ or saline injection. For each time point, Ins2 mRNA expression of WT mice injected with vehicle (WT control) is set at 1 (n = 5). (c) Plasma insulin levels of WT and Rkip1 −/− mice 28 days after injection with STZ or vehicle (n = 5). (d) Pancreatic insulin content of WT and Rkip1 −/− mice 28 days after injection with STZ or vehicle (n = 4). Black bars, WT animals treated with vehicle; white bars, WT animals treated with STZ; horizontal lines, KO control, vertical lines, KO STZ. All data are the mean ± SEM. *p < 0.05, **p < 0.01 between indicated conditions

Full size image

Discussion

How beta cell mass is regulated is a critical issue for understanding the pathophysiology of diabetes, which is characterised by a complete (type 1) or relative (type 2) deficiency in beta cell number. During mouse embryogenesis, new beta cells arise through the proliferation and differentiation of a population of progenitor cells. During postnatal development, between birth and weaning, there’s an additional and massive increase in beta cell mass that mainly involves replication from pre-existing beta cells, even though neogenesis is maintained at least during the first week of life. By the time of weaning, beta cell proliferation decreases and is maintained relatively low throughout adulthood (for reviews, see [26, 32]). In this study we aimed to investigate the physiological role of RKIP1 in the regulation of beta cell mass. Our results indicate that RKIP1 participates in the control of pancreatic growth, particularly at the time of weaning. In addition, we demonstrate that the absence of RKIP1 confers significant protection against STZ-induced diabetes, thus indicating that this protein may negatively influence beta cell regeneration after injury in the adult pancreas.

Rkip1 −/− animals exhibit increased beta cell mass relative to body weight and this augmentation, in the absence of changes in insulin sensitivity, is the likely cause for their improved glucose tolerance. Indeed, beta cell mass expansion and body weight follow a linear correlation throughout life [28], and alterations in this correlation are known to affect glucose homeostasis [26]. A recent study showed that beta cell-specific deletion of Raf-1 kinase results in the downregulation of Ins2 gene transcription and, consequently, in blunted insulin secretion [25]. In contrast, we have found that the insulin content, insulin gene expression and glucose-induced insulin secretion of Rkip1 −/− islets is similar to that of WT islets (Fig. 3 and data not shown), arguing against a major involvement of altered islet function in enhanced whole body glucose control in Rkip1 −/− mutant mice. Yet, it is conceivable that, in adult islets under normal physiological conditions, Raf-1 kinase is already maximally stimulated, so that elimination of RKIP1 has no measurable effects on Raf-1 kinase-dependent processes.

In accordance with prior in vitro experiments [13], we have observed that RKIP1 inhibits beta cell proliferation in vivo. Remarkably, the inhibitory effect of RKIP1 is only evident during early postnatal stages when the beta cell proliferative capacity is maximal. This age-dependent loss of the effects of RKIP1 may simply reflect changes in the repertoire of signalling factors available to the beta cell at any given moment in time. Alternatively, changes in the responsiveness of beta cells to specific signalling molecules, rather than their availability, may underlie the age-dependent effects of Rkip1 ablation. In line with this, Chen et al [33] have recently linked age-triggered loss of beta cell expression of the platelet-derived growth factor (PDGF) receptor to reduced beta cell proliferation. Hence, it is tempting to speculate that the proliferative response to Rkip1 ablation will be lost in parallel with the loss of the ability of the beta cell to respond to specific factors.

Genetic mouse models and in vitro studies have helped unravel the identity of growth factors that stimulate beta cell replication, but their role under normal physiological conditions remains uncertain. Among the pathways implicated, many transduce signals through the Raf-1/MEK/ERK cascade, including insulin, insulin-growth factor, hepatic growth factor, epidermal growth factor and PDGF [34–36], and thus could potentially be targeted by RKIP1. Yet, our data indicate that constitutive ERK1/2 activation may not be linked to enhanced beta cell proliferation in Rkip1 mutants. In a similar way, deletion of Raf-1 does not result in decreased ERK1/2 phosphorylation in pancreatic islets, indicating that other Raf-1-dependent pathways may be operative in beta cells [25]. Additionally, RKIP1 may function through Raf-1 independent mechanisms involving NF-κB [8], GSK3β [9] or G protein-coupled receptors [11]. Apoptotic rates are comparable between Rkip1 −/− and WT P28 pancreases, implying that NF-κB-dependent survival pathways are not affected by Rkip1 deletion. On the other hand, RKIP1 deficiency does not alter levels of GSK3β in islets, as reported in other tissues [9], ruling out this possible mechanism. Future work should be aimed at delineating the signal cascades regulated by RKIP1 and their connection to beta cell mitogens during physiological beta cell growth.

This study reveals that RKIP1 is not only involved in restricting beta cell proliferation but it may also inhibit postnatal pancreatic growth. Gene expression profiling has identified the Reg3 gene family as potentially involved in the stimulatory effect of RKIP1 ablation on pancreatic growth at P28. Interestingly, the Reg3 genes have been previously associated to pancreatic regeneration and islet cell neogenesis [37–41]. Further supporting a role for Reg3 genes in pancreatic postnatal development, a recent study has described the upregulation of Reg3 genes in the mouse pancreas between P7 and P30 followed by their subsequent downregulation in adulthood [42]. In this scenario, our data suggest that RKIP1 may participate in restricting Reg3 gene expression after weaning and, in this way, control pancreatic growth. The question of whether exocrine Reg3 gene activation might contribute to enhanced beta cell proliferation in Rkip1 −/− mice remains unresolved. Double Reg3 and Rkip1 mutants and beta cell-specific Rkip1 −/− mice will be necessary to properly address this issue.

In summary, in this paper we show that RKIP1 inactivation results in enhanced beta cell and pancreatic growth in postnatal stages, leading to an increment in beta cell mass and improved glucose tolerance in adult animals. Furthermore, Rkip1 deletion confers recovery to STZ-induced hyperglycaemia in adult animals. These results have therapeutic implications for the development of novel strategies to trigger beta cell mass expansion in diabetes.