Osteoprotegerin promotes islet β cell proliferation in intrauterine growth retardation rats through the PI3K/AKT/FoxO1 pathway

Abstract

Osteoprotegerin (OPG) is a critical factor involved in bone metabolism. The level of OPG is increased in the serum of diabetic patients; however, there is no consensus in prior studies on the role of OPG in regulating the function of islet β cells. A rat model of intrauterine growth retardation (IUGR) was established in the present study to investigate whether OPG could enhance the proliferation of β cells; and an in vitro culture model of rat islet β cell line INS-1 was used, to confirm the effect of OPG supplementation and reveal the possible mechanism. The results showed that endogenous OPG expression was reduced and normal proliferation of β cells was impaired in the IUGR islets. Exogenous supplement of OPG restored β cell proliferation to an extent in the IUGR rats, possibly associated with regulation of the PI3K/AKT/FoxO1 signalling, as evidenced by the changes of protein expression in the pathway. Furthermore, treating rat islet INS-1 cells with a PI3K inhibitor, LY294002, blunted the effects of OPG supplement in promoting cell cycle and suppressing cell apoptosis. Taken together, the present work demonstrated that OPG supplementation could improve the proliferation of islet β cells in IUGR, and the PI3K/AKT/FoxO1 pathway is involved in the underlying mechanism.

Introduction

Intrauterine growth retardation (IUGR) refers to a condition in which a fetus fails to reach the appropriate size due to effects of genetic factors. Maternal conditions for example, obesity and hypertension, could result in decreased nutritional flow through the placenta, resulting in IUGR [1], which is associated with a high perinatal mortality rate (responsible for 30% of stillbirths) and increases vulnerability to other diseases including gestational diabetes mellitus (GDM) [1,2]. Many animal experiments and clinical studies have suggested that IUGR is closely related to the occurrence of metabolic syndrome in adulthood [3-5] such as insulin resistance and diabetes later in life, but the mechanisms remain unclear. Currently, a “thrifty phenotype hypothesis” is generally accepted. Hales and Baker reassessed this hypothesis [6] and indicated that fetal and early postnatal malnutrition can lead to islet β cell dysplasia and islet function impairment. Given that the adult human β cell has a low rate of proliferation and is highly refractory to external stimuli by pregnancy, obesity or insulin resistance [7,8], the β cell homeostasis in younger life could be important to set the foundations for normal, functional and regenerative β cells in later life [9].

Osteoprotegerin (OPG), a member of the tumour necrosis factor receptor (TNFR) superfamily, is a soluble decoy receptor for the receptor activator of nuclear factor κB ligand (RANKL) [10,11], whose role in bone metabolism has been widely studied. In recent years, the role of OPG in the pancreas has gradually attracted attention, but no consensus has yet been reached [9,12,13]. Different groups of investigation have suggested elevated serum OPG levels in type 1 and type 2 diabetic patients [14-16] and Toffoli et al. have shown intraperitoneal supplementation of OPG might impair insulin secretion [13]. However, other studies in animal models [9,12] and human islet cells [9,12] suggested an opposite role of OPG in promoting β cell function and preventing cell death in young, old and diabetic mice. There are a few studies trying to decode the relationship between IUGR and OPG regulation [17-19]. Serum OPG levels were similar among IUGR and healthy group (both the mothers and the infants) at postnatal day 1-4 [17,18]. However, by using an animal model, serum OPG levels significantly differed, being much lower in IUGR pups compared to control [19]. None of the studies investigated the OPG levels in detail at the site of pancreas, and maternal supply of OPG in infants was not excluded in these studies.

Many existing studies have indicated that the PI3K signalling pathway is involved in regulating the proliferation and function of islet β cells, and that AKT as an important downstream substance of PI3K plays a crucial role in the growth of islet β cells [20-23]. As recent studies have shown that OPG can activate the PI3K/AKT pathway in many organs [24-26], we speculate that the PI3K/AKT pathway is likely an important pathway for OPG-mediated islet β cell proliferation. FoxO1 is specifically expressed in islet β cells in adult islets, and several studies have indicated that it is involved in cell proliferation, apoptosis, and metabolism [22,27-30]. However, how FoxO1 is regulated and the functional outcome of FoxO1 locational and post-translational dynamics in infants remains elusive.

In this study, we aim to utilize an in vivo rat IUGR model and replenish the new-born pups with OPG simulating a dose of endocrine OPG to investigate the role of OPG in regulating infant β cell proliferation. At the same time, we aim to combine the in vitro cell culture model to dissect the underlying molecular mechanism. The results from this research would shed light on potential new therapeutics in IUGR infant to prevent metabolic disorders in an individual’s adult life.

Materials and methods

Animals and grouping

Virgin female Wistar rats (weighting 250~300 g, obtained from Changsheng Biotechnology Co., Ltd., Benxi, Liaoning, China) were mated overnight. Rats that were confirmed to have conceived were randomly divided and received two diet regimens during the full pregnancy period, respectively: a low-protein diet and a standard-feed diet [31] (Table S1). The newborn offspring of the mothers with standard-feed diet were used as the control group (n = 30/group). The newborn offspring of the mothers with low-protein diet, whose birth weight was below the 10th percentile of the normal birth weight, were defined as IUGR rats. The IUGR offspring rats were randomly divided into the IUGR+OPG and IUGR groups (n = 30/group).

The rats in the IUGR+OPG group received the first intraperitoneal injection of rhOPG (3 μg/g [32-34]; Peprotech, Rocky Hill, NJ, USA) within 24 hours after birth, followed by injections every other day until 3 weeks after birth; and the rats in the IUGR and control groups received the first intraperitoneal injection of water (with the same volume as that of the rhOPG solvent) within 24 hours after birth, followed by injections every other day until 3 weeks after birth. The selected time points of this study included the first week after birth (1 w), the third week after birth (3 w) and the 12th week after birth (12 w).

Cell culture and grouping

The rat islet β cell line (INS-1 cells) was cultured in RPMI 1640 medium (containing 10% foetal bovine serum, 5.6 mmol/l glucose, 100 U double antibiotics, and 50 mol/l 2-mercaptoethanol; Bioind, Israel) under culture conditions of 37°C and 5% CO₂. Cells were divided to CON, OPG, LY and OPG+LY groups. The cells in the OPG group were treated with 0.1 μg/ml rhOPG [9] (Peprotech), cells in the LY group were treated with 10 μM LY294002 for 30 min (CST, Boston, MA, USA), cells in the OPG+LY group were treated with 10 μM LY294002 for 30 min and then 0.1 μg/ml rhOPG was added. The control group were treated with vehicle. The cells were treated for 48 h before further experiments.

Immunohistochemistry

3 μm-thick paraffin of pancreas were sequentially deparaffinized in xylene and gradient ethanol. Antigens were retrieved using sodium citrate retrieval solution. The sections were incubated by hydrogen peroxide for 45 min. Goat serum was dropped onto the tissues for blocking for 40 min. Then, the sections were incubated with OPG antibody (1:200, Abcam) overnight at 4°C. The sections were incubated with secondary antibody for 30 min. The sections were incubated with horseradish peroxidase-labeled streptavidin working solution for 15 min. Next, DAB Substrate Kit was used to perform the chromogenic reaction. Then sections were stained with hematoxylin for 2 min and washed in running water for bluing. Sections were dehydrated sequentially in gradient ethanol and xylene.

Immunofluorescence

3 μm-thick paraffin of pancreata were subjected to immunofluorescence staining with anti-Ki67 antibody (1:200, Abcam, Cambridge, MA, USA) and anti-insulin antibody (1:50, Wuhan Boster Biological Engineering Co., Ltd., Wuhan, Hubei, China) according to manufacturer’s instructions. The donkey anti-mouse Alexa Fluor 488 (1:400, Abcam) and donkey anti-rabbit Alexa Fluor594 (1:400, Abcam) secondary antibody were added in dark. DAPI staining for the nucleus was performed for 5 min. The slides were examined and photographed under an immunofluorescence confocal microscope (Zeiss Ism880). Five rats in each group were used to prepare the slices for immunofluorescence. One field of view per section was observed. The number of Ki67⁺ beta cells/total number of beta cells was calculated for each group.

Cell viability analysis (CCK-8 assay)

INS-1 cells in the logarithmic growth phase were evenly plated in 96-well plates, with n = 3 for each condition. For each well, 100 μl of cell suspension solution was added at the density of 1000-2000 cells per well. After treatment, the cell viability was measured according to manufacturer’s instructions (EnoGene Biotech Co. Ltd., Nanjing, Jiangsu, China). The OD value at the 450-nm wavelength for each measurement well was determined with a microplate reader (BioTek Instruments, Inc., USA).

Cell cycle analysis

The cells were harvested and fixed in 70% ethanol at 4°C for more than 2 h, and then the cells were washed by PBS. Each group of cell samples was added to 500 μl of staining buffer, and then stained with PI staining solution (25 μl PI and 10 μl RNase A) at 37°C away from light for 30 min. Last, flow cytometry (BD Calibur) was performed. The rate of cellular proliferation was determined using a formula as follows: %P = (S+G2-M)/G0-G1+S+G2-M) × 100%.

Apoptosis analysis (Annexin-FITC/PI staining and flow cytometry)

Post treatment, the cells were harvested. Then, 1 × Annexin V Binding Solution (Dojindo Laboratories, Japan) was added to prepare a cell suspension at the concentration of 1 × 10⁶ cells/ml. For each group, 100 μl of 1 × 10⁵ cell suspension was collected, and 5 μl of FITC and 5 μl of PI were added; the resultant sample was then allowed to stand for 15 min at room temperature in the dark, and apoptosis was detected by flow cytometry (BD Calibur) within 1 h.

Western blot

Rat pancreatic tissues or cells were extracted of total protein (or cytosol protein and nuclear protein) using RIPA lysis buffer. The proteins were quantified by BCA method. About 40 μg proteins were loaded in each lane and separated by 10% SDS-PAGE gel, and then transferred to a PVDF membrane. The primary antibodies against AKT (1:1000, CST), p-AKT (Ser473; 1:1000, CST), FoxO1 (1:1000, CST), p-FoxO1 (Ser256; 1:1000, CST), cyclin D1 (1:500, Wanleibio Co. Ltd., Shenyang, Liaoning, China), cyclin E (1:750, Wanleibio), Bcl2 (1:500, Wanleibio), Bax (1:500, Wanleibio), GAPDH (1:6000, Proteintech, Rosemont, IL, USA), Tubulin (1:6000, Proteintech), and Histone H3 (1:300, Wanleibio) were incubated overnight at 4°C, and then goat anti-rabbit secondary antibody (1:5000, Proteintech) was added followed by incubation at room temperature for 2 h. After ECL development, Image J software was used to analyse the protein density.

Statistics

The statistical analysis was performed using SPSS21.0 software. The data was expressed as the mean ± standard deviation (SD). Comparisons between two groups were performed using independent-samples t-test, and comparisons among multiple groups using one-way analysis of variance (ANOVA) with Fisher’s least significant difference (LSD; homogeneity of variance) or Dunnett’s T3 (heterogeneity of variance) post-hoc test. P < 0.05 was considered significant.

Results

The growth of the pancreas was impaired in the IUGR rats

As shown in Table 1, the body weights of the rats in the IUGR group at 1 week after birth and 3 weeks after birth were significantly lower than those of the CON group, and the body weights of the IUGR rats injected with OPG were not significantly different from those of the IUGR rats without OPG injection; at 12 weeks after birth, no significant difference in body weight was observed between the three groups of rats.

Table 1.

Body weights, pancreas weights, and pancreas weight/body weight ratios of the rats

Group		Body weight (g)	Pancreas weight (mg)	Pancreas weight/body weight (‰)
1 W	CON	14.583 ± 0.372	36.183 ± 1.788	2.480 ± 0.075
	IUGR	11.417 ± 0.691**	22.620 ± 1.431**	1.984 ± 0.116**
	IUGR+OPG	12.081 ± 0.903**	28.315 ± 1.419**,##	2.348 ± 0.109*,##
3 W	CON	53.122 ± 2.085	260.658 ± 13.556	4.912 ± 0.229
	IUGR	44.187 ± 1.003**	185.967 ± 10.984**	4.206 ± 0.163**
	IUGR+OPG	46.083 ± 2.242**	215.077 ± 9.275**,##	4.685 ± 0.210##
12 W	CON	459.342 ± 7.712	2567.308 ± 138.151	5.586 ± 0.207
	IUGR	451.373 ± 6.730	2158.135 ± 143.059**	4.778 ± 0.295**
	IUGR+OPG	457.925 ± 7.168	2433.642 ± 139.282##	5.311 ± 0.238##

^*indicates P < 0.05, compared to the CON group;

^**indicates P < 0.01, compared to the CON group;

^##indicates P < 0.01, compared to the IUGR group.

At each time point, the pancreas weights of the rats in the IUGR group were lower than those of the rats in the CON group; the pancreas weight of the IUGR+OPG group was higher than that of the IUGR group. Furthermore, at each time point, the ratio of pancreas weight/body weight in the IUGR group was lower than that in the CON group; the ratio of pancreas weight/body weight in the IUGR+OPG group was higher than that in the IUGR group, and the results were significantly different.

OPG was down-regulated in IUGR islets after birth

To investigate whether the development of islet cells was impaired in IUGR rats, we performed immunohistochemistry (IHC) against OPG in control (CON) vs. IUGR newborn litters at postnatal (p) day 1, week 1, week 3 and week 12 (Figure 1A). We found that OPG was specifically expressed in the pancreatic islets of both CON and IUGR groups. However, the size of islets was much smaller by observation (Figure 1A) and the intensity of staining was statistically lower in IUGR vs. CON islets, as quantified in Figure 1B, which suggested that the expression of OPG was attenuated in IUGR islets as early as p1 and lasted for at least 3 months in postnatal growth.

Figure 1.

The OPG expression was examined by immunohistochemistry (IHC) in the pancreas islets of IUGR pups vs. control. A. Representative IHC results showing OPG expression in pancreas islets at postnatal day 1, week 1, 3 and 12. Magnification is 400×. B. The mean density of OPG staining using ImageJ analysis. Control pups were term pups by wild type Wistar rats under standard diet. IUGR pups were born from wild type mothers on low-protein diet and birth weight below 10% of normal. N = 6. *P < 0.05 and **P < 0.01.

Administration of OPG restored the proliferation of islet β cells in IUGR rats

The reduced islet size in IUGR rats could be due to impaired expression of OPG, which could cause a decrease of cell volume or numbers. To test this hypothesis, we performed immunofluorescence (IF) staining for a well-known proliferation marker, Ki67. The cell size was not obviously different among CON, IUGR and IUGR+OPG groups, defined with the longest length times the longest width for insulin labelled cells (Data not shown) at 1 w (Figure 2A, 2D), 3 w (Figure 2B, 2D) and 12 w (Figure 2C, 2D) after birth. However, proliferating islet β cells (Insulin⁺Ki67⁺) in the IUGR group were significantly fewer than that in the CON group (Figure 2A-D). With replenishment of recombinant human OPG (rhOPG) through intraperitoneal injection, the proliferation of islet β cell in the IUGR rats was restored to a level comparable to 50-70% of control (Figure 2D). Taken together, these results indicated that the smaller islets in IUGR rats were due to a reduction in β cell numbers due to loss of proliferation, which could be rescued partially by supplementing injected rhOPG.

Figure 2.

Administration of OPG alleviated the defects of β cell proliferation in IUGR pancreas. Pancreatic sections were used for immunofluorescent (IF) staining for insulin and Ki67 at postnatal day 1 (A), postnatal week 3 (B) and postnatal week 12 (C). The ratio of proliferative islet cells was calculated as Ki67⁺ beta cells/total number of beta cells in (D). Data were analyzed from n = 5 pups for each condition. N = 5 for postnatal week 12, and N = 6 for postnatal week 1 and 3. **P < 0.01 versus the CON group at the same time point; #P < 0.05 and ##P < 0.01, versus the IUGR group at the same time point.

Administration of OPG restored the level of p-AKT and p-FoxO1 in IUGR rat pancreas

Given that OPG could activate the PI3K/AKT/FoxO1 in many organs [24-26] and the activation of PI3K promotes islet β cell growth [20-23], we next examined whether OPG regulates islet β cell proliferation by activating the PI3K/AKT/FoxO1 pathway in vivo. The results showed that compared with the CON group, the levels of activated form, p-AKT (Figure 3A) and pre-degraded p-FoxO1 (Figure 3B), in the pancreatic tissues of IUGR rats were significantly decreased. We also evaluated the cellular localization of FoxO1 as a transcription factor. We identified that FoxO1 protein translocated to and was venriched in the nucleus for IUGR rats at a higher level compared to CON (Figure 3C, 3D), indicating augmented FoxO1 signalling that impairs the β cells. After treatment with rhOPG, the phosphorylation levels of p-AKT and p-FoxO1 in the pancreatic tissues of IUGR rats were significantly increased (Figure 3A, 3B), and the protein amount of FoxO1 was decreased in the nucleus (Figure 3C) and increased in cytosol (Figure 3D). Taken together, these results suggested that the PI3K/AKT/FoxO1 pathway functioned downstream of OPG and the growth defect of IUGR β cells could be due to mitigated PI3K/AKT/FoxO1 pathway in the presence of inadequate OPG expression.

Figure 3.

Administration of OPG restored the attenuated PI3K/AKT/FoxO1 pathway in IUGR pancreas at postnatal week 12. A. Phosphorylation levels of AKT in each group of rat pancreatic tissues. B. Phosphorylation levels of FoxO1 in each group of rat pancreatic tissues. C. FoxO1 protein expression in the nucleus. D. FoxO1 protein expression in the cytosol. N = 6. **P < 0.01 versus the CON group at the same time point; #P < 0.05 and ##P < 0.01, versus the IUGR group at the same time point.

Inhibition of PI3K using LY294002 offset the effects of OPG treatment in promoting cell cycle and reducing apoptosis in INS-1 islet cells

Due to the fact the rhOPG administration could improve the proliferation of IUGR islet cells and activate the PI3K/AKT/FoxO1 pathway in vivo, we next performed in vitro experiments specifically in rat islet β cells using an established cell line INS-1, because it provides large numbers of pure cells for biochemical and molecular analysis. First, we treated the cultured cells with OPG to test whether the in vitro model mimics the molecular events in vivo and found that addition of OPG in the culture medium would cause up-regulation of p-AKT and p-FoxO1 (Figure 4A, 4B). Next, we investigated the effect of OPG in re-establishing p-AKT and p-FoxO1 levels in the presence of LY294002 (LY), an inhibitor for PI3K (Figure 4A, 4B). Consistent with the in vivo data, supplementation of OPG in INS-1 cells could elevate the levels of p-AKT and p-FoxO1 and application of LY brought down the p-AKT and p-FoxO1 levels to normal. Notably, the cellular location and enrichment of FoxO1 were also consistent; FoxO1 tended to reduce in the nucleus and accumulate in the cytosol with OPG, whereas LY294002 caused FoxO1 to translocate from cytosol to the nucleus. OPG+LY showed similar nuclear and cytosolic levels of FoxO1 as CON (Figure 4C, 4D). These results indicated that INS-1 cells could faithfully recapitulate the molecular regulation underlying OPG in vivo and inhibition of PI3K would function against OPG.

Figure 4.

Inhibition of PI3K attenuated the elevation of p-AKT and p-FoxO1 by OPG. A. Phosphorylation levels of AKT in each group of cells. B. Phosphorylation levels of FoxO1 in each group of cells. C. FoxO1 protein expression in the nucleus. D. FoxO1 protein expression in the cytosol. Cells were treated with OPG (0.1 μg/ml), LY (PI3K inhibitor, 10 μM) and OPG+LY in the culture media. N = 3. **P < 0.01 versus the CON group at the same time point; ##P < 0.01 versus the OPG group at the same time point.

Then we performed CCK8 viability test on INS-1 cells to see whether OPG and LY294002 would affect cell proliferation in vitro. Similar to the in vivo results, administration of OPG promoted the growth of INS-1 cells while LY294002 prevented the proliferation of INS-1 cells, and LY treatment on top of OPG could offset the improvement in cell viability by OPG (Figure 5).

Figure 5.

Inhibition of PI3K reduced OPG-promoted cell viability. N = 3. *P < 0.05 versus the CON group; #P < 0.05 versus the OPG group.

Another possibility causing a small islet would be a stall in cell cycle on top of reduced proliferation in IUGR rats, which are difficult to evaluate due to methodological limitations in vivo. Therefore, we went ahead to evaluate the effects of OPG, LY and OPG+LY on cell cycle regulation in INS-1 cells. Compared with the CON group, OPG treated INS-1 cells showed reduced proportion in the G0/G1 phase, therefore meaning more proliferative cells in S/M/G2 phase. Inhibition of PI3K by LY294002 resulted in blockage in G0/G1, and similar with the LY group, OPG+LY still showed stall in G0/G1 compared to CON or OPG (Figure 6A). With regard to cyclin D1 (marking G0/G1 transition) and cyclin E (marking G1/S transition) expression levels, we observed a similar trend with the p-AKT and p-FoxO1 expression and cell growth (Figures 4 and 5). The expression of cyclin D1 and cyclin E were increased in OPG, decreased in LY and maintained comparable to CON in OPG+LY (Figure 6B, 6C). This suggested that LY could cause an increase in G0/G1 cell proportions, which suggested fewer proliferative cells, even in the presence of OPG, although the levels of Cyclin D1 and Cyclin E were similar between OPG+LY and CON.

Figure 6.

Inhibition of PI3K offset the increase of INS-1 cell proliferation and cyclin expression induced by OPG treatment. A. Cell cycle assay results for each group of cells. B. Expression of Cyclin D1. C. Expression of Cyclin E. N = 3. *P < 0.05 and **P < 0.01, versus the CON group; ##P < 0.01 versus the OPG group.

To test another hypothesis that reduced cell death by OPG treatment would be counteracted by LY, we analysed apoptosis in these cells by staining and flowing with Annexin V, a marker for surface phosphatidylserine (PS) expression on cells. Compared with the CON group, the proportion of apoptosis was significantly decreased in the OPG group and significantly increased in the LY group, while the proportion of apoptosis in the LY+OPG group were still much higher than that in CON or OPG group (Figure 7A). To confirm that cell apoptosis was indeed significantly present in OPG+LY, we further detected the protein expression of BCL-2 and BAX, the classic mediators in apoptosis. BCL-2 protein expression was increased, and BAX protein expression was decreased in the OPG group vs. CON. The LY group showed the opposite trend whereas OPG+LY group showed similar expression levels as CON (Figure 7B, 7C). Taken together, these results led to the conclusion that LY treatment would neutralize the lessening of apoptosis due to exogenous OPG in rat INS-1 cells, which was also confirmed by the western blot results.

Figure 7.

Inhibition of PI3K upon the OPG treatment resulted in an increase of INS-1 cell apoptosis. A. Results of the cell apoptosis assay in various groups. B. The expression of apoptotic protein Bcl-2. C. The expression of apoptotic protein Bax. All annexin V positive cells were included as apoptotic including DAPI⁺ (dead) and DAPI^- (pro-apoptotic). N = 3. *P < 0.05 and **P < 0.01, versus the CON group; #P < 0.05 and ##P < 0.01, versus the OPG group at the same time point.

Discussion

OPG has been implied to play a role in regulating the development of pancreas and pancreatic functional cells and suggested as a biomarker for IUGR pancreatic function. Given the difficulty of studying OPG in IUGR human subjects and lack of evidence in studying OPG in IUGR animal models, we performed studies using OPG as a potential therapy for treating IUGR rat pancreas and at the same time scrutinized the underlying molecular mechanism regulating islet β-cell numbers. We have showed that OPG could help in maintaining normal levels of PI3K/AKT/FoxO1 signalling, resulting in normal proliferation of islet β cells in vivo using a rat IUGR model. In addition, we further confirmed this result in vitro and showed inhibiting PI3K with LY treatment would offset the appealing effects of OPG, suggesting that OPG indeed functions at least partially through the PI3K/AKT/FoxO1 pathway to housekeep normal cell growth and viability. Our study has revealed that OPG is a key regulator in foetal and neonatal islet β cell development and intraperitoneal supplement of OPG is a potential treatment for IUGR related pancreatic dysfunction.

The risk of metabolic syndrome increasing in IUGR individuals is mainly caused by islet β-cell dysplasia. In recent years, the role of OPG in the pancreas has been gradually discovered by researchers: studies have shown that OPG injection into diabetic mice resulted in a protective effect on islet β cells [9]. Injecting certain drugs into IUGR rats during the time window of β cell remodelling (2-3 weeks after birth) could prevent β cell failure and reduce the chance of getting diabetes in IUGR adults [35,36]. Therefore, in this study, to detect the effect of OPG on the proliferation of islet β cells in IUGR rats, IUGR pups were injected with rhOPG inconsecutively from birth to 3 weeks after birth. The proliferation level of islet β cells in IUGR offspring rats receiving rhOPG injections was found to be lower than that in the normal offspring rats but significantly higher than that in the IUGR pups without rhOPG injection. This pattern was observed not only at 1 week and 3 weeks after birth but also at 12 weeks after birth. Therefore, in the islet β cell remodelling period, rhOPG injection into IUGR rats could promote islet β cell proliferation, which may presumably reduce the likelihood of T2DM development later in life. Further studies would be required to test this hypothesis.

In addition, the translocation of FoxO1 from the nucleus to the cytosol may also inhibit FoxO1-mediated transcription [37,38]. In our results we found that IUTR pancreas and LY-induced INS-1 cells displayed less p-FoxO1 but more nuclear total proteins. This could be due to cellular compensatory effects trying to elevate p-FoxO1 levels by supplying more substrates or failure to down-regulate FoxO1 transcription by enabling translocation to the cytosol. AKT might be phosphorylated upon OPG stimulation, and activated AKT could promote the phosphorylation of FoxO1, which in turn causes FoxO1 to translocate from nucleus to cytosol and thus disable its activity, ultimately protecting islet β cells.

This study has also addressed the cell cycle and apoptosis of islet β cells under the condition of OPG and OPG+LY treatment, which is usually difficult to approach in human subjects. We showed that the key cell cycle modulators Cyclin D1 and Cyclin E were affected with OPG, LY and OPG+LY treatment. The related apoptosis was closely involving the Bcl-2/Bax classic pathway, consistent with previous studies that FoxO1 can regulate cell apoptosis by regulating the protein expression of Bcl-2 and Bax [27,39]. LY294002 is a strong and specific PI3K inhibitor containing a morpholine group [40]. Application of LY could offset the promoting effect of elevated p-AKT and p-FoxO1, suggesting the binding of LY directly to PI3K was strong enough to block induced activation of PI3K by transcriptional or physical regulation of OPG. Yet, the mechanism between OPG activation and LY inhibition could be partially or totally independent since adding LY on top of OPG did not fully block the PI3K/AKT/FoxO1 pathway.

Some controversial studies have been reported around studying the relationship between OPG and IUGR [17-19]. Using human subjects, circulating OPG was evaluated between appropriate for gestational age (AGA) and IUGR pregnancies both for the mothers and foetuses at postnatal day 1 to 4 and no difference was observed [17,18]. However, in rat IUGR models, the serum OPG levels were significantly lower in IUGR pups compared to controls at postnatal week 1, 3 and 8 [18]. The differences between human and rat IUGR newborns in terms of serum OPG concentrations might be attributed to different timings of examination. Maternal OPG could be present in newborns within 4 days after birth, and OPG synthesis deficiency would be manifested later in postnatal life. Moreover, these above studies mainly focused on the bone development of IUGR fetuses and mainly examined the serum OPG together with other bone-forming biomarkers. The direct analysis of OPG levels in the pancreas or even in the islets would be the most ideal to reveal the relationship between OPG and islet development. A few studies have been performed trying to elucidate the role of OPG in β cell function [9,12,13]. Toffoli et al. suggested OPG-treated mice showed increased islet macrophage infiltration and impaired function [13]. In contrast, Schrader et al. showed OPG could respond to pro-inflammatory cytokines and prevent cell death in human and rat β cells [12]. Kondegowda et al. showed that induced OPG could induce β cell mass in young, aged, and STZ (Streptozotocin)-treated diabetic mice [9]. Our study agreed with Schrader et al. and Kondegowda et al. that OPG provides a positive effect in β cell proliferation. The difference against Toffoli et al. could be non-specific TUNEL staining on the islet cells or an effect resulted from different doses of OPG intake (1 μg/mouse in Toffoli et al.). Therefore, the fine-tuning of OPG might be important when considering clinical therapeutic strategies in treating IUGR related islet failure.

In summary our study has performed both in vivo and in vitro studies to elucidate the role of OPG in modulating the normal proliferation and cell death of β cells with a potential to treat islet dysplasia in IUGR individuals. Our research has suggested a role of OPG acting at least partially through the PI3K initiated cell growth pathway and noted that the fine-tuned dosage of OPG should be taken into great consideration for future clinical exploration.

Disclosure of conflict of interest

None.