Short-term low-protein diet during pregnancy alters islet area and protein content of phosphatidylinositol 3-kinase pathway in rats

The phosphatidylinositol 3-kinase and mitogen-activated protein kinase pathways mediate β cell growth, proliferation, survival and death. We investigated whether protein restriction during pregnancy alters islet morphometry or the expression and phosphorylation of several proteins involved in the phosphatidylinositol 3-kinase and mitogen-activated protein kinase pathways. As controls, adult pregnant and non-pregnant rats were fed a normal-protein diet (17%). Pregnant and non-pregnant rats in the experimental groups were fed a low-protein diet (6%) for 15 days. Low protein diet during pregnancy increased serum prolactin level, reduced serum corticosterone concentration and the expression of both protein kinase B/AKT1 (AKT1) and p70 ribosomal protein S6 kinase (p70), as well as the islets area, but did not alter the insulin content of pancreatic islets. Pregnancy increased the expression of the Src homology/collagen (SHC) protein and the extracellular signal-regulated kinases 1/2 (ERK1/2) independent of diet. ERK1/2 phosphorylation (pERK1/2) was similar in islets from pregnant and non-pregnant rats fed a low-protein diet, and was higher in islets from pregnant rats than in islets from non-pregnant rats fed a normal-protein diet. Thus, a short-term, low-protein diet during pregnancy was sufficient to reduce the levels of proteins in the phosphatidylinositol 3-kinase pathway and affect islet morphometry.


INTRODUCTION
Although pancreatic islet mass is relatively constant under normal conditions, it has a remarkable ability to adapt to metabolic changes.An example is pregnancy, when islet mass can increase markedly due to both β-cell hyperplasia and hypertrophy (Sorenson and Brelje 1997).An increase in β-cell proliferation is first observed around day 10 of pregnancy, peaks around day 14 and then returns to control levels by the end of pregnancy at approximately 21 days (Parsons et al. 1992).Another example of the adaptation of pancreatic islets occurs upon malnutrition, which is associated with important structural changes in the endocrine pancreas.Protein deficiency leads to reduced β-cell mass due to a decrease in the proliferation rate and an increase in apoptosis (Swenne et al. 1992, Petrik et al. 1999).Additionally, when the data are corrected for age and weight, a reduction in β-cell size and total islet volume, a lower proportion of β-cells per islet, a degranulation of β-cells and β-cells with a preponderance of abnormal pale granules have been observed (Heard and Stewart 1971, Pimstone 1976, Weinkove et al. 1977).
In growing rats receiving a low-protein diet, serum corticosterone concentrations are increased, the plasma concentration of IGF-1 and the number of GH and PRL receptors (GH-R and PRL-R) are reduced, at least in the liver (Herbert and Carrillo 1982, Prewitt et al. 1982, Maes et al. 1984).In rodents, the PL, PRL, E2 and P levels increase during the second half of pregnancy (Nadal et al. 2009), and elevated concentrations of corticosterone are observed during the later stages of pregnancy (Dupouy et al. 1975).Increases in the mRNA expression of GH-R and PRL-R (Moldrup et al. 1993) and the expression of PRL-R in the pancreas (Brelje and Sorenson 1997) have also been observed during pregnancy.Dexamethasone, a synthetic glucocorticoid, reverses the PRL-induced upregulation of islet mass by inhibiting cell proliferation and increasing apoptosis (Weinhaus et al. 2000).
Changes in the protein expression and/or activation/phosphorylation of a protein common to these pathways could result in alterations in proliferation and apoptosis.These two events could be associated with the expansion and reduction of pancreatic islet mass that are observed during pregnancy with a low-protein diet.Thus, we investigated the effects of short-term protein restriction during pregnancy on the morphometry and the expression and phosphorylation of several proteins involved in the PI3K and MAPK pathways in pancreatic islets.Studies that evaluate the role of protein restriction on molecular and cellular determinants involved in β-cell adaptation during pregnancy are particularly important to identify possible causal factors of gestational diabetes and could contribute towards its prevention.

ANIMALS AND DIET
The animal experiments were approved by the Institutional Ethics Committee in Animal Experimentation (Universidade Federal de Mato Grosso).Virgin female Wistar rats (90 days of age) were obtained from the University's breeding colony.Mating was achieved by housing males with females overnight, and pregnancy was confirmed by the examination of vaginal smears for the presence of sperm.Pregnant and non-pregnant females were each randomly assigned to two diet groups: control and low -protein.The control nonpregnant (CNP) and pregnant (CP) groups were fed a 17% protein diet, and the low-protein nonpregnant (LPNP) and pregnant (LPP) groups were fed a 6% protein diet from days 1 to 15 of pregnancy.The diets were isocaloric, as previously described (Filiputti et al. 2008).During the experimental period, rats had free access to food and water and were housed at 22°C with a 12h light:dark cycle.At the end of this experimental period, the rats were weighed and killed by decapitation.Blood samples were collected and allowed to clot; sera were stored at -20°C for the subsequent measurement of 17β-estradiol and progesterone by EIA (BioChem ImmunoSystems, Italy), prolactin by EIA (ALPCO Diagnostics, USA) and corticosterone by an enzyme-linked immunosorbent assay (Assay Design, USA).Serum albumin levels (Doumas et al. 1971) were determined immediately after decapitation.Since it was not possible to evaluate all variables in the same animals, the number of individual experiments varied among groups.

MORPHOMETRY AND IMMUNOHISTOCHEMISTRY
For morphological analyses, the pancreases of rats were kept for approximately 8h in 10% buffered formalin and then processed by routine methods for histological analysis.After histological sectioning (5-µm-thick sections) and mounting onto positivized slides (ImmunoSlide-EasyPath ® ), the anti-insulin immunohistochemistry technique was performed.The sections were de-waxed in xylene and rehydrated in decreasing concentrations of alcohol (in distilled water).The blocking of endogenous peroxidase was performed by incubating the slides in a 3% solution of hydrogen peroxide in distilled water for 15 min at room temperature.For antigen retrieval, the samples were microwaved for 2 min in a citrate buffer of pH 6.0.Non-specific binding was minimized with a 15 min incubation in 5% skimmed milk (Molico ® ) diluted in distilled water.The sections were covered with a solution containing the guinea pig anti-swine insulin polyclonal antibody (A0564 -Dako, Corp., Carpinteria, CA, dilution 1:150) and were incubated in a moist chamber at 37°C for 60 min.The samples were then washed in distilled water and treated with an anti-IgG guinea pig biotinylated secondary antibody (61-4620, ZYMED, Invitrogen USA) for 30 min in a moist chamber at room temperature.Next, the sections were washed in distilled water and treated with a streptavidin-peroxidase conjugate (DAKO Corp., Carpinteria, CA) for 30 min in a moist chamber at room temperature.Staining was subsequently visualized with the application of red chromogen (VECTOR ® NovaRED) for 5 min.The sections were then washed in distilled water and counterstained with Harris hematoxylin for 30s.Next, the samples were washed in water for 1-2 min, dehydrated in increasing concentrations of alcohol, cleared in xylene and mounted with Entellan (Merck, Darmstadt, Germany Sigma Chemical Co., Saint Louis, USA).Approximately 20 islets were observed in one field of view for each slide.The slides were evaluated by light microscopy, and images were captured with a digital camera (Sony Cyber-shot DSC-H5) and measured using the Motic ® Images Plus 2.0 ML software.In the morphometric technique the observer was blinded as to the groups and the pancreatic tissue was assessed for the area of the islets.

STATISTICAL ANALYSES
The results are expressed as the means ± SDs for the number of rats (n) indicated.For islets, n refers to the number of experiments performed.Levene's test for the homogeneity of variances was used initially to check the fit of the data to parametric ANOVA assumptions.When necessary, the data were log-transformed to correct for variance in heterogeneity or non-normality (Sokal and Rohlf 1995).Data were analyzed by a two-way ANOVA test (nutritional status and physiological status).When necessary, these analyses were followed by LSD's honestly significant difference test to determine the significance of individual differences.The level of significance was set at P<0.05.The data were analyzed using the Statistic Software package (Statsoft).
Detection of IR-β, IRS-1, IRS-2 and PI3K-p85 obtained through immunoprecipitation showed no significant differences in the levels of all four proteins observed among the groups (Fig. 1A to 1D).As shown in Figure 1E, the protein levels of AKT1 were 1.5-fold higher in islets from the LPNP group compared with the CNP group.The pancreatic islets from LPP rats had lower levels of AKT1 than did the pancreatic islets from LPNP rats.No significant difference was detected in the level of AKT1 expression between CP and CNP rats.The phospho [Se473] AKT1 level (Fig. 1F) and the ratio of phospho [Se473] AKT1/AKT1 (Fig. 1G) were lower in islets from low protein rats (LPP and LPNP) than in those of the control rats (CNP and CP) (F 1,8 =25.9, P<0.001, F 1,8 =6.4,P<0.05, respectively).In contrast, the variables were 1.3and 1.7-fold higher in the islets from pregnant rats (LPP and CP) compared with non-pregnant rats, respectively, (LPNP and CNP) (F 1,8 =10.24,P<0.02, F 1,8 =7.7, P<0.05, respectively).Levels of p70 S6K (Fig. 1H) and phosphor p70 S6K (Fig. 1I) were detected from Western blot signals from immunoprecipitated samples.As depicted in Figure1H, the pancreatic islets from LPP rats expressed lower levels of p70 S6K than did those from LPNP rats, whereas the islets from CP rats expressed higher levels of p70 S6K than did those from CNP rats.The phosphor p70 S6K levels were 1.4-fold higher in the islets from pregnant rats (LPP and CP) in relation to non-pregnant rats (LNP and CNP) (F 1,8 =5.32, P<0.05) (Fig. 1I).No significant differences were detected in the phosphop70 S6K /p70 S6K ratio among groups (Fig. 1J).
As shown in Figure 2A, the expression of SHC increased 1.5-fold in the islets from rats fed a lowprotein diet in relation to rats fed the control diet (F 1,8 = 16.25,P<0.01).Similar results were observed in the islets from pregnant rats in relation to those from non-pregnant rats (F 1,8 =15.63,P<0.01).The protein levels of both ERK1 (Fig. 2B) and ERK2 (Fig. 2C) were decreased in the islets from rats fed a low-protein diet (F 1,10 =11.64,P<0.01 and F 1,8 = 11.97,P<0.01, respectively) compared with those from rats maintained on the control diet.However, the expression of ERK1 and ERK2 was 1.4-fold (F 1,10 =10.18,P<0.01) and 1.3-fold (F 1,8 = 7.91, P<0.05) higher in the islets from pregnant rats compared with non-pregnant rats, respectively.The phosphoERK1/2 levels were similar in islets from the LPNP, LPP and CP groups and were significantly higher in relation to those from the CNP group (Fig. 2D).

DISCUSSION
After the short period of low-protein diet consumption (15 d) used here, the classical alterations observed in malnourished animals (weight loss and hypoalbuminemia) were not yet present.We verified an increase in E2 and P in both CP and LPP rats, which is normally observed during pregnancy (Nadal et al. 2009).In disagreement with the observation that the maternal plasma total and free corticosterone concentrations in rats are increased by protein restriction and pregnancy at 15 days (Dupouy et al. 1975, Martin et al. 1977, Herbert and Carrillo 1982), in the present study, the serum corticosterone levels did not increase in CP rats and were decreased in LPP rats.Interestingly, the unexpected decrease in the serum corticosterone levels in LPP rats was accompanied by an early increase in the serum prolactin concentration.This increase is commonly observed in the later stages of pregnancy (Nadal et al. 2009) and suggests an inhibitory role for corticosterone in prolactin secretion (Nishino et al. 1992).Corticosterone and prolactin are involved in the structural adaptations of pancreatic islets to pregnancy.Some studies have attempted to identify the molecular mechanisms underlying these adaptations in response to pregnancy and protein deprivation.Long-term exposure to a low-protein diet during the growth phase produces an increase in IR, IRS2 and PI3K levels and a reduction in IRS1 and p70 S6K expression (Araujo et al. 2004, Filiputti et al. 2008).In this study, a short-term, low-protein diet did not alter the IR, IRS1, IRS2 or PI3K-p85 levels but did increase the expression of AKT1 and p70 S6K in the pancreatic islets of non-pregnant rats.Pregnancy associated with protein deprivation was LOW-PROTEIN DIET DURING PREGNANCY REDUCES ISLET AREA An Acad Bras Cienc (2015) 87 (2) 1015 accompanied by the reduced expression of AKT1 and p70 S6K , whereas normal pregnancy did result in an increase in the expression of p70 S6K , but did not alter AKT1 expression.Although similar in CNP and CP groups, the AKT1 expression tended to increase in CP rats compared with CNP rats.Neither condition resulted in changes in the expression of IR, IRS1/2 or PI3K-p85.Interestingly, normal pregnancy does not affect the protein levels of IRS1 and IRS2; however, it increases IRS1/2 phosphorylation and its association with p85, and also increases AKT1, p70 S6K expression and phosphorylation in pancreatic islets (Amaral et al. 2004).
Several studies indicate that the structure and function of pancreatic islets are influenced by alterations in the expression and/or activity of IR, IRS-1, IRS-2, AKT1 and p70 S6K .A reduced IR content in β-cells strongly impairs insulin secretion in response to glucose (Kulkarni et al. 1999).In contrast, IRS-1 deficiency produces hyperinsulinemia and increases of the β-cell mass, and the absence of IRS-2 produces hyperinsulinemia followed by islet hypoplasia and early diabetes (Kulkarni et al. 1999, Withers et al. 1998).It has been demonstrated that AKT promotes protein synthesis by modulating mTOR and its downstream target, p70 S6K (Hannan et al. 2003).Increased AKT1 activity is accompanied by a marked increase in β-cell mass due to enhanced cell survival and increased cell size (Tuttle et al. 2001).Additionally, p70 S6K deficiency produces hypoinsulinemia, islet insensitivity to glucose, a low insulin content and reduced β-cell size (Pende et al. 2000).In the present study, changes in the AKT1 and p70 S6K levels were not followed by alterations in the insulin content of islets.However, changes in the levels of these two proteins were accompanied by a decrease in islet area in the LPP group.Because the LPP group had lower AKT1 and pAKT1 contents and a lower pAKT1 Ser473/ AKT1 ratio, which are indirect indicatives of AKT1 activity, it is possible that the reduction in islet area resulted from elevated apoptosis and/or reduced AKT1-mediated cellular proliferation and growth.Corroborating with the hypothesis of elevated apoptosis, we observed that islets from LPP group exhibited elevated levels of caspase 3 and caspase 9 (unpublished data).
In several signaling pathways, the expression and phosphorylation of members of the MAPK cascade activate several transcriptional regulators to control cellular growth.Although ERK1/2 modulate insulin secretion (Longuet et al. 2005), the most important role of these proteins in pancreatic β-cells is the regulation of proliferation, cell growth, and survival (Lawrence et al. 2008).
In the present study, pregnancy increased the expression of SHC and ERK1/2 independent of diet and enhanced the level of pERK1/2 only in the islets of normal pregnant rats, as previously shown (Bordin et al. 2004, Amaral et al. 2004).A low-protein diet also elevated the levels of SHC and pERK1/2 and reduced the expression of ERK1 and ERK2.Pregnancy did not increase the pERK1/2 levels in the islets of rats exposed to a low-protein diet.
In pancreatic β-cells, nutrients, hormones and growth factors modulate the expression of proteins in the SHC/ERK and IRS/PI3K/ AKT/ p70 S6K pathways.Prolactin has been considered an important growth factor.Prolactin is involved in the increase in islet mass and increased sensitivity to glucose that occur during pregnancy.Mechanistically, it activates proteins that are downstream of PI3K and MAPK cascades (Amaral et al. 2004).Dexamethasone, a synthetic glucocorticoid, increases the p70 S6K expression and pAKT1 level in pancreatic islets (Rafacho et al. 2007).Curiously, in this study, LPP rats that exhibited an increase in serum prolactin levels and a decrease in serum corticosterone concentrations also had lower p70 S6K and AKT1 levels.These results corroborate the observation that glucocorticoids strongly synergize with prolactin-induced STAT5 activation through distinct mechanisms involving the glucocorticoid receptor (Stocklin et al. 1996, Lechner et al. 1997, Wyszomierski et al. 1999).
Our results show that short-term exposure to a low-protein diet during pregnancy reduces pancreatic islet area that could be related to deficiencies in the AKT1 and p70 S6K levels.We suggest that reduced circulating corticosterone levels altered the PI3K pathway and, consequently, cellular proliferation, growth, survival and death.

ACKNOWLEDGMENTS
This work was partly supported by the Fundação de Amparo à Pesquisa do Estado de Mato Grosso (FAPEMAT, grant number: 0786/2006) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, grant number: 305155/2004-0).Celso Roberto Afonso provided technical assistance for the study.This work is part of a dissertation that was presented by Cristiana dos Santos Barbosa Salvatierra as a partial requirement for her Master's degree in Health Sciences at the School of Medical Sciences, UFMT.

Figure 2 -
Figure 2 -The levels of the SHC (A), ERK1 (B) and ERK2 proteins (C) and phospho-ERK1/2 levels (D) in the islets of CNP, CP, LPNP and LPP rats.Equal amounts of protein (50 µg) were run on SDS-PAGE and blotted with anti-SHC, anti-ERK1, anti-ERK2 and anti-phospho-ERK1/2.The columns represent the means ± SDs of 3-4 independent experiments.# Indicates differences in relation to non-pregnant rats (Two-way ANOVA, P<0.05).*Indicates differences in relation to rats maintained on a control diet (Two-way ANOVA, P<0.05).a,b Mean values with unlike superscript letters were significantly different (LSD test; P<0.05). S6k