GLP-1–oestrogen attenuates hyperphagia and protects from beta cell failure in diabetes-prone New Zealand obese (NZO) mice
- First Online:
- 1.7k Downloads
Oestrogens have previously been shown to exert beta cell protective, glucose-lowering effects in mouse models. Therefore, the recent development of a glucagon-like peptide-1 (GLP-1)–oestrogen conjugate, which targets oestrogen into cells expressing GLP-1 receptors, offers an opportunity for a cell-specific and enhanced beta cell protection by oestrogen. The purpose of this study was to compare the effects of GLP-1 and GLP-1–oestrogen during beta cell failure under glucolipotoxic conditions.
Male New Zealand obese (NZO) mice were treated with daily s.c. injections of GLP-1 and GLP-1–oestrogen, respectively. Subsequently, the effects on energy homeostasis and beta cell integrity were measured. In order to clarify the targeting of GLP-1–oestrogen, transcription analyses of oestrogen-responsive genes in distinct tissues as well as microarray analyses in pancreatic islets were performed.
In contrast to GLP-1, GLP-1–oestrogen significantly decreased food intake resulting in a substantial weight reduction, preserved normoglycaemia, increased glucose tolerance and enhanced beta cell protection. Analysis of hypothalamic mRNA profiles revealed elevated expression of Pomc and Leprb. In livers from GLP-1–oestrogen-treated mice, expression of lipogenic genes was attenuated and hepatic triacylglycerol levels were decreased. In pancreatic islets, GLP-1–oestrogen altered the mRNA expression to a pattern that was similar to that of diabetes-resistant NZO females. However, conventional oestrogen-responsive genes were not different, indicating rather indirect protection of pancreatic beta cells.
GLP-1–oestrogen efficiently protects NZO mice against carbohydrate-induced beta cell failure by attenuation of hyperphagia. In this regard, targeted delivery of oestrogen to the hypothalamus by far exceeds the anorexigenic capacity of GLP-1 alone.
KeywordsBeta cells GLP-1 Liver fat NZO Oestrogen Pomc
Carbohydrate-free high-fat diet
Carbohydrate-containing high-fat diet
Oestrogen receptor α
Insulin sensitivity index
New Zealand obese
To compensate for peripheral insulin resistance and glucose intolerance, pancreatic beta cells start to proliferate and increase the biosynthesis and secretion of insulin . The genetic background and environmental factors limit the capacity of this compensation, however, and beta cells eventually fail, leading to type 2 diabetes. In order to prevent this progression, current research focuses on strategies to protect beta cells against the toxic microenvironment produced by circulating carbohydrates and lipids.
Recent data have implicated a beta cell protective role of oestrogen (17β-oestradiol, E2). E2 has been shown to increase insulin biosynthesis via activation of oestrogen receptor α (ERα) [2, 3] and to protect beta cells against toxic lipid intermediates by promoting cell proliferation and inhibition of lipogenesis and apoptosis [4, 5]. Additionally, systemic E2-mediated effects on food intake and energy expenditure also contribute to beta cell protection . In support of these findings, women generally have a lower prevalence of type 2 diabetes than age-matched males, although this changes after menopause . In line with these observations, the EPIC-InterAct study revealed an inverse correlation between age at menopause onset with the risk of developing type 2 diabetes . Despite all these promising findings, oestrogen has not been evaluated as a glucose-lowering drug due to its mitogenic effects in reproductive tissue .
Similar to E2, the incretin hormone GLP-1 increases insulin biosynthesis and survival of beta cells, lowers food intake and increases glucose uptake in adipose and muscle tissue . In contrast to more widespread action of E2, GLP-1 action is restricted to tissues presenting the GLP-1 receptor at its cell surfaces. Recently, we showed that hybrid molecules of E2 and GLP-1 (GLP-1–oestrogen) boost the weight lowering effects of GLP-1 in C57BL/6 mice by targeted delivery of oestrogen to the hypothalamus . Consequently, lower doses of the steroid could be used, and the tumourigenic potential of E2 was masked. Although the body weight lowering effects of GLP-1-oestrogen were obvious, the question remained whether the hybrid compound would also be sufficient to protect beta cells under diabetogenic conditions.
Similar to humans, New Zealand obese (NZO) mice develop obesity and insulin resistance as a result of hyperphagia, reduced energy expenditure and insufficient physical activity . The progression from insulin resistance to type 2 diabetes in NZO mice is largely driven by dietary carbohydrates, as carbohydrate-free diets fail to induce diabetes in NZO mice . Taking advantage of this, we previously established a dietary regimen of 13 weeks of carbohydrate-free high-fat diet (to induce obesity and insulin resistance) followed by a carbohydrate-containing high-fat diet that rapidly leads to hyperglycaemia and beta cell destruction . In this study, we used the same model system to investigate the glucose-lowering potential of GLP-1–oestrogen under defined glucolipotoxic conditions.
After an overnight 16 h fasting period, mice received 2 mg glucose per g body weight by oral gavage. At the indicated time points (Fig. 3b) blood glucose and plasma insulin were measured, as previously described .
Immunohistochemistry of pancreatic islets
Pancreatic tissue excised immediately after exsanguination was fixed in 4% (wt/vol.) formaldehyde and embedded in paraffin according to standard procedures. For co-staining of insulin and glucagon, mouse monoclonal anti-insulin (clone K36AC10, Sigma-Aldrich, Munich, Germany) and polyclonal rabbit anti-glucagon (Dako, Hamburg, Germany) antibodies were used. Alexa Fluor 546-labelled anti-rabbit (1:200) and Alexa Fluor 488-labelled anti-mouse (1:200; Invitrogen, Karlsruhe, Germany) were used as secondary antibodies. Nuclei were stained with DAPI.
Pancreatic insulin content
For detection of the pancreatic insulin content, whole pancreases were homogenised in ice-cold acidic ethanol (0.1 mol/l HCl in 70% ethanol) and incubated for 24 h at 4°C. After centrifugation (16,000×g, 10 min) insulin was detected in the supernatant fraction using the Mouse High Range Insulin ELISA (Alpco, Salem, USA).
Insulin sensitivity index
Whole body insulin sensitivity was calculated after the method of Matsuda and DeFronzo . Briefly, fasting blood glucose (G0) and insulin (I0) and the mean blood glucose and insulin during OGTT (G and I, respectively) were recorded. Insulin sensitivity index (ISI) was calculated (10,000/square root of [G0 × I0] × [G × I]).
Laser micro dissection of hypothalamic nuclei and gene expression analyses
Dissected brains were immediately frozen on dry ice, and RNA was extracted as described previously .
Gene expression analyses in adipose tissue and liver
Total RNA from visceral adipose tissue and liver tissue of mice was extracted, and cDNA synthesis as well as TaqMan gene expression assays were performed as described previously .
Liver histology and triacylglycerol determination
Histological staining of liver connective tissue was performed using a Masson–Goldner staining kit (Merck Millipore, Darmstadt, Germany). For the quantitative determination of triacylglycerol content, livers were homogenised in 10 mmol/l sodium dihydrogen phosphate, 1 mmol/l EDTA, and 1% (vol./vol.) polyoxyethylene-10-tridecyl ether, incubated for 5 min at 37°C, and the triacylglycerols in the supernatant fraction were detected with a commercial kit (RandoxTR-210, Crumlin, UK).
Islet isolation and transcriptome analysis
Isolation of pancreatic islets was performed by a modified protocol of Gotoh et al . Total islet RNA preparation was performed with the RNAqueous®Micro Kit (Life Technologies, Darmstadt, Germany). RNA integrity was assessed with the RNA6000 nano kit (Agilent, Santa Clara, CA, USA). Microarray analyses were performed by OakLabs (Hennigsdorf, Germany) on a Agilent Mouse 8 × 60 K Chip.
Statistical differences during treatment were determined by two-way ANOVA and Bonferroni posttest. Differences in endpoint measurements were determined by one-way ANOVA and Newman–Keuls Multiple Comparison Tests. Contingency of the expression analyses was calculated by Fisher’s Exact Test. Significance levels were set at *p < 0.05, **p < 0.01 and ***p < 0.001. Data are presented as means ± SEM. For statistical analysis and for graphical presentation GraphPad Prism (5.0; GraphPad Software, San Diego, CA, USA) was used.
GLP-1–oestrogen prevents carbohydrate-induced hyperglycaemia
Pancreatic islets are protected by GLP-1–oestrogen against carbohydrate-induced destruction
Oestrogen improves islet function but not glucose tolerance
Reduced food intake is associated with increased anorexigenic signalling
Attenuation of lipogenic pathways in GLP-1–oestrogen-treated mice
GLP-1 and GLP-1–oestrogen treatment affects the transcriptome of pancreatic islets
Genes selectively upregulated by GLP-1–oestrogen treatment
UDP-N-acteylglucosamine pyrophosphorylase 1-like 1
Vomeronasal 2, receptor 3
C2 calcium-dependent domain containing 4A
Phosphoserine aminotransferase 1
Late cornified envelope 1K
Haematopoietic cell transcript 1
Olfactory receptor 48
Dual adaptor for phosphotyrosine and 3-phosphoinositides 1
Glycine C-acetyltransferase (2-amino-3-ketobutyrate-coenzyme A ligase)
Carbohydrate sulfotransferase 11
Ankyrin repeat domain 22
Olfactory receptor 521
Zinc finger protein 846
Extended synaptotagmin-like protein 3
ADP-ribosylation factor-like 9
Pregnancy induced noncoding RNA
Predicted gene 11213
Genes selectively downregulated by GLP-1–oestrogen treatment
Transient receptor potential cation channel, subfamily M, member 1
Arrestin domain containing 4
F-box protein 34
Coiled-coil domain containing 138
Solute carrier family 22 (organic cation transporter), member 2
Receptor (calcitonin) activity modifying protein 1
Glutathione S-transferase omega 2
Histone cluster 2, H4
Tripartite motif-containing 7
RIKEN cDNA 6330416G13 gene
Olfactory receptor 522
Scavenger receptor class F, member 2
Cytochrome c oxidase subunit VIb polypeptide 2
In this study, we evaluated the potential of oestrogen-coupled GLP-1 to protect beta cell function under glucolipotoxic conditions in diabetes-prone male NZO mice. We show that GLP-1–oestrogen fully prevented the onset of hyperglycaemia and reduced body weight due to a substantially decreased food intake, indicating the hypothalamus to be the main site of GLP-1–oestrogen action. Subsequently, GLP-1–oestrogen protected the mice against carbohydrate-induced beta cell failure, increased glucose tolerance and insulin sensitivity and affected the islet transcriptome. Thus, compared with GLP-1, low-dose GLP-1–oestrogen revealed superior efficacy to preserve beta cell integrity and function under diabetogenic conditions.
The findings of the present study indicate that the combination of GLP-1 and oestrogen in a hybrid molecule possesses a glucose-lowering potential that exceeds the potential of either one of the single molecules. Recently we showed that oestrogen has beta cell protective effects in female NZO mice, as ovariectomised animals displayed elevated blood glucose levels and eventually an increased prevalence of type 2 diabetes compared with sham-operated control mice . Furthermore, treatment of oestrogen-deficient mice with oestrogen resulted in increased protection of pancreatic beta cells against streptozotocin-induced beta cell apoptosis and the collapse of insulin production . In the present study, transcriptome analyses did not indicate direct beta cell protection by GLP-1–oestrogen at the given dose, as neither Trim25 nor Acaca, Fasn and Scd1 were differentially expressed in pancreatic islets of GLP-1–oestrogen-treated mice. The later three lipogenic genes mentioned have previously been shown to be repressed in pancreatic islets of Zucker diabetic fatty rats upon oestrogen treatment . Still, our transcriptome analyses indicated several alterations of the pancreatic expression pattern that could be protective, even as secondary effects of hypothalamic GLP-1–oestrogen action. Particularly, the α-arrestin Txnip is well known as a key player in pancreatic beta cell biology, as it is increased in diabetic islets and induces beta cell apoptosis [21, 22]. The GLP-1–oestrogen-mediated suppression of Txnip could be mediated via GLP-1 as was shown for exenatide . Additionally, oestrogen-mediated repression of Txnip was demonstrated in vitro and in vivo . Still, inhibition of Txnip alone was not sufficient to prevent beta cell failure, as seen in GLP-1-treated animals. Interestingly, a second α-arrestin, Arrdc4, is suppressed in islets after GLP-1–oestrogen treatment. This effect appears to be mediated by oestrogen, because also females, but not GLP-1-treated males, exhibited this lower expression. Whether Arrdc4 has similar adverse effects as Txnip in beta cells is not known; however, our data suggest that inhibition of both genes in GLP-1–oestrogen-treated mice participates in beta cell protection.
GLP-1-bound oestrogen stimulated anorexigenic signalling that was far more effective than GLP-1 alone. The hybrid compound reaches the brain, as GLP-1–oestrogen treatment increased hypothalamic expression of oestrogen-responsive Trim25  to a similar extent as oestrogen alone. Nevertheless, induction of Pomc expression was clearly highest in GLP-1–oestrogen-treated animals, indicating that oestrogen only affects appetite in NZO mice when combined with GLP-1. Brain-targeted oestrogen not only affects Pomc expression but also has been shown to increase the firing rate of Pomc-expressing neurons, resulting in substantial reduction in food intake, and subsequently body weight . Therefore, our data are in line with these published observations and suggest that reduced caloric intake via pro-opiomelanocortin (POMC) activation is the major mechanism leading to improved glycaemia in GLP-1–oestrogen-treated NZO mice. Indeed, several studies have proven that caloric restriction is sufficient to improve glucose tolerance and insulin sensitivity in humans and animal models [26, 27]. Especially, liver fat decreases within days upon caloric restriction and contributes to improved glucose homeostasis . In contrast to Pomc, expression of anorexigenic Leprb was elevated in GLP-1-, GLP-1–oestrogen- and oestrogen-treated mice. In previous studies, NZO mice have been shown to be severely leptin resistant and this leptin resistance might be due to the presence of several polymorphisms in the Lepr gene [29, 30]. Therefore, the absence of any anorexigenic signalling in GLP-1- or oestrogen-treated NZO mice could be illustrative of an impaired leptin signalling in NZO mice, despite increased Lepbr expression.
GLP-1–oestrogen treatment did not alter Trim25 expression in liver or visceral adipose tissue. This is opposite to the findings in the hypothalamus and similar to that in pancreatic islets, and could be explained by the lack of GLP-1 receptor expression in these tissues [31, 32]. Interestingly, carbohydrate feeding also suppressed Trim25 expression in the hypothalamus, but not in the periphery. The mechanism behind this observation, as well as the tissue-specific counter regulation by oestrogen, is not known and data on a metabolic function of Trim25 are missing. Still, both liver and adipose tissue displayed a substantial reduction in their lipid content, indicating a secondary effect through central actions of GLP-1–oestrogen that ultimately influences whole body metabolism. Furthermore, we observed an inhibition of lipogenic genes in the liver. This is in line with previous studies showing that oestrogen treatment of ovariectomised mice and high-fat diet (HFD)-fed mice resulted in inhibition of lipogenic gene expression in liver and adipose tissue [33, 34]. These published data also indicated enhanced lipolytic response and β-oxidation, which was not assessed in our study and would need further investigation to clarify. However, both studies assumed direct effects of oestrogen on both tissues. The data in the present study suggest that a major part of the oestrogen action is via the central nervous system, as a direct action of GLP-1–oestrogen on liver and adipose tissue are more likely to be excluded. This concept is supported by the finding that specific deletion of hypothalamic ERα is sufficient to reduce whole body energy expenditure and induce hyperphagia in female mice, resulting in obesity and impaired glucose tolerance [35, 36].
In summary, oestrogen-coupled GLP-1 displays superior efficacy in preventing the onset of diet-induced diabetes than GLP-1 alone. In NZO mice, this protective effect is due to central attenuation of hyperphagia, resulting in systemic improvement of glucose tolerance and insulin sensitivity. Therefore, hybrid compounds like GLP-1–oestrogen might be the basis for novel therapeutic options for treating type 2 diabetes mellitus more efficiently.
The authors thank C. Gumz, A. Teichmann and K. Warnke of the German Institute of Human Nutrition Potsdam-Rehbruecke for their skilful technical assistance. Furthermore, we thank S. Morin of the Institute for Diabetes and Obesity at the Helmholtz Center (Munich, Germany) for editing the manuscript, and M. Jähnert and G. Schulze of the German Institute of Human Nutrition Potsdam-Rehbruecke for data analysis.
This work was supported by the German Ministry of Education and Research (BMBF, DZD, grant 01GI0922) and the German Research Foundation (GK1208).
Duality of interest
RDD was a cofounder of Marcadia Biotech. All other authors declare that there is no duality of interest associated with their contribution to this manuscript.
RWS was responsible for study conception and design, performed data acquisition and analysis, and drafted the article. CBa, BF, OK and CBr performed data acquisition, data analysis and contributed to the writing of the manuscript. HGJ advised on the study concept and critically revised the manuscript. RDD, MHT and AS made substantial contributions to the study conception and critically revised the manuscript. RWS is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All listed authors approved the final version of the manuscript.
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.