Restoration of Lepr in β cells of Lepr null mice does not prevent hyperinsulinemia and hyperglycemia.

OBJECTIVE: The adipose-derived hormone leptin plays an important role in regulating body weight and glucose homeostasis. Leptin receptors are expressed in the central nervous system as well as peripheral tissues involved in regulating glucose homeostasis, including insulin-producing β cells of the pancreas. Previous studies assessing the role of leptin receptors in β cells used Cre-loxP to disrupt the leptin receptor gene (Lepr) in β cells, but variable results were obtained. Furthermore, recombination of Lepr was observed in the hypothalamus or exocrine pancreas, in addition to the β cells, and Lepr in non-β cells may have compensated for the loss of Lepr in β cells, thus making it difficult to assess the direct effects of Lepr in β cells. To determine the significance of Lepr exclusively in β cells, we chose to selectively restore Lepr in β cells of Lepr null mice (LeprloxTB/loxTB).
MATERIALS AND METHODS: We used a mouse model in which endogenous expression of Lepr was disrupted by a loxP-flanked transcription blocker (LeprloxTB/loxTB), but was restored by Cre recombinase knocked into the Ins1 gene, which is specifically expressed in β cells (Ins1Cre). We bred LeprloxTB/loxTB and Ins1Cre mice to generate LeprloxTB/loxTB and LeprloxTB/loxTBIns1Cre mice, as well as Leprwt/wt and Leprwt/wtIns1Cre littermate mice. Male and female mice were weighed weekly between 6 and 11 weeks of age and fasting blood glucose was measured during this time. Oral glucose was administered to mice aged 7-12 weeks to assess glucose tolerance and insulin secretion. Relative β and α cell area and islet size were also assessed by immunostaining and analysis of pancreas sections of 12-14 week old mice.
RESULTS: Male and female LeprloxTB/loxTB mice, lacking whole-body expression of Lepr, had a phenotype similar to db/db mice characterized by obesity, hyperinsulinemia, glucose intolerance, and impaired glucose stimulated insulin secretion. Despite restoring Lepr in β cells of LeprloxTB/loxTB mice, fasting insulin levels, blood glucose levels and body weight were comparable between LeprloxTB/loxTBIns1Cre mice and LeprloxTB/loxTB littermates. Furthermore, glucose tolerance and insulin secretion in male and female LeprloxTB/loxTBIns1Cre mice were similar to that observed in LeprloxTB/loxTB mice. Analysis of pancreatic insulin positive area revealed that restoration of Lepr in β cells of LeprloxTB/loxTB mice did not prevent hyperplasia of insulin positive cells nor did it rescue Glut-2 expression.
CONCLUSION: Collectively, these data suggest that direct action of leptin on β cells is insufficient to restore normal insulin secretion and glucose tolerance in mice without leptin receptor signaling elsewhere.

Introduction

Obesity is often associated with elevated circulating insulin levels (hyperinsulinemia), glucose intolerance, and reduced insulin sensitivity [1], [2]. The molecular mechanisms underlying the increased insulin secretion and islet hyperplasia that are often concomitant with obesity remain unclear, but the adipose-derived hormone leptin may be a potential link between adipose tissue and β cells. Leptin has an established role in regulating food intake, adipose metabolism, and glucose homeostasis through signaling in the central nervous system (CNS) [3], [4]. In addition, leptin receptors are distributed in several peripheral tissues involved in glucose regulation including the liver, adipose, and pancreas [5], [6]. Within these tissues, the long isoform of the leptin receptor (Leprb) is believed to mediate the glucoregulatory actions of leptin [7].

In vitro studies have demonstrated that Lepr is expressed in murine and human pancreatic β cells, as well as in β cell lines [8], [9], [10], [11]. Previous studies have attempted to assess whether leptin has a direct action on β cell function, but the results are contradictory. While some studies have found that incubation of human or rodent islets with leptin did not alter the expression or secretion of insulin [12], [13], [14], other studies demonstrated reduced insulin expression and secretion from islets treated with leptin [8], [12], [15], [16], [17], [18]. The inhibitory actions of leptin on β cell function are supported by in vivo characterization of glucose homeostasis and circulating insulin levels in mice lacking leptin or its receptor. In leptin deficient mice (ob/ob) and mice with mutated Lepr (db/db), hyperinsulinemia develops prior to increases in body weight, and islet hyperplasia is evident in these mice [19], [20], [21]. When ob/ob mice are treated with leptin, there is a reduction of insulin gene expression and secretion and a reduction of circulating insulin levels within hours [22]. Moreover, doses of leptin insufficient to reduce body weight are sufficient to reverse hyperinsulinemia, indicating that the insulin-lowering effects of leptin can occur independent of weight loss [23].

To elucidate the mechanism by which leptin can regulate insulin secretion and expression, previous studies have employed Cre-loxP technology. Disruption of Lepr in β cells was achieved by Covey et al. using a rat insulin promoter driving Cre expression (RIPCre) [24]. Leprflox/flox RIPCre mice were obese, hyperinsulinemic, glucose intolerant, and experienced islet hyperplasia relative to their Leprflox/flox littermates. Interestingly, a similar approach by Morioka and colleagues to disrupt Lepr using a Pdx-1 (pancreatic duodenal homeobox 1) promoter to drive Cre expression resulted in mice with mildly elevated fasting insulin levels but the absence of obesity [25]. Furthermore, Morioka et al. reported improved glucose tolerance and enhanced glucose stimulated insulin secretion in chow-fed Leprflox/flox Pdx-1Cre mice compared to Leprflox/flox littermates [25]. Though intended to target β cells, the Cre recombinase used in both of these studies was present in other tissues including the hypothalamus [24] and brain stem, as well as the exocrine pancreas [25], [26], [27], confounding interpretation of the direct actions of leptin in pancreatic β cells. More recently, Soedling et al. crossed mice in which Cre was knocked into the insulin I gene (Ins1Cre) with Leprflox/flox mice to selectively delete Lepr in pancreatic β cells [28]. Unlike previous studies, which reported elevated plasma insulin levels following knockout of Lepr [24], [25], knockout of Lepr in β cells using Ins1Cre did not induce elevated plasma insulin levels. Moreover, knockout of Lepr using Ins1Cre did not result in glucose intolerance or obesity. While the studies described above used very similar approaches to assess the role of Lepr in β cells, differences in genetic background and tissue specificity of Cre lines may have contributed to the different phenotypes observed upon deletion of Lepr. Furthermore, these knockout studies are limited by the potential compensation by Lepr expressed either in other peripheral tissues or the central nervous system (CNS) to make up for the life-long loss of direct action of leptin in β cells.

To determine the actions of leptin in β cells, we exploited a murine model in which a transcriptional blockade flanked by two loxP sites is placed between exon 16 and exon 17 of Lepr (LeprloxTB/loxTB), resulting in impaired leptin signaling [29]. To assess the direct actions of leptin in β cells, we crossed these mice with Ins1Cre mice to selectively restore Lepr in pancreatic β cells, while functional leptin receptors were absent in other tissues. This provided a unique model with which to assess whether the direct actions of leptin in β cells are sufficient to prevent hyperinsulinemia.

Methods

Experimental animals

LeprloxTB/loxTB mice (JAX no. 018989, 50% FVB, 50% C57BL/6J, <1% 129) and Ins1Cre mice (JAX no. 026801, C57BL/6J) [29] were obtained from Jackson Laboratory (Bar Harbor, ME, USA). LeprloxTB/wt mice were then bred with Ins1Cre mice for two generations to yield LeprloxTB/loxTB Ins1Cre, LeprloxTB/loxTB, Leprwt/wt Ins1cre, and Leprwt/wt mice on a B6.FVB hybrid background. Offspring were born at the expected Mendelian ratio. To assess the specificity of Ins1Cre mediated recombination, mT/mG reporter mice (JAX no. 007676, B6.129, Jackson Laboratory), which possess a membrane-bound Tomato fluorescent protein (mT) flanked by two unidirectional loxP sites, followed by the open reading frame for membrane bound enhanced green fluorescent protein (EGFP, mG) driven by the chicken β actin promoter, were crossed with Ins1Cre mice to generate homozygous mT/mG mice with or without 1 allele for Ins1Cre. Mice were housed on a 12 h light/12 h dark cycle and had ad libitum access to food (chow diet 2918, Harlan Laboratories, Madison WI, USA) and water. At 13 weeks of age, animals were fasted for 4 h prior to euthanasia. Pancreata were collected and weighed prior to fixation. All procedures with animals were approved by the University of British Columbia Animal Care Committee and carried out in accordance with the Canadian Council of Animal Care guidelines.

Metabolic measurements

Body weight and blood glucose were measured in mice fasted for 4 h. Blood was collected from the saphenous vein to measure blood glucose using a One Touch Ultra Glucometer (Life Scan, Burnaby, Canada). Insulin was measured in plasma (Mouse Ultrasensitive Insulin ELISA, Alpco Diagnostics, Salem, NH, USA) collected at 7 and 11 weeks of age. At 9 weeks of age, LeprloxTB/loxTB Ins1Cre mice and littermate control mice underwent a 6h fast prior to oral gavage of dextrose (1.5 g/kg, 40% solution) followed by measurement of blood glucose and collection of plasma at time = 0, 15, 30, and 60 min to assess glucose tolerance and glucose stimulated insulin secretion.

Islet isolation

Islets were isolated from 11 to 13 week old male and female Lepr Ins1Cre and LeprloxTB/loxTB littermate mice as previously described [30]. Briefly, mice were euthanized by CO2, and 2 ml of Hank's balanced salt solution (HBSS) containing 1000 units/ml of type XI collagenase (Sigma–Aldrich, St. Louis, MO) was immediately injected into the pancreatic duct. Pancreata were digested at 37 °C for 11 min and then washed with ice cold HBSS containing 1 mM CaCl2. Islets where then hand picked into RPMI 1640 (Sigma Aldrich) containing 0.5% BSA, 100 units/ml penicillin and 100 μg/ml streptomycin. Islets were then flash frozen in liquid nitrogen and stored at −80 °C.

Gene deletion

Immediately following euthanasia, mice were decapitated and their brains were isolated. Hypothalami was then dissected on ice and flash frozen. To assess gene deletion, genomic DNA (gDNA) was extracted from flash frozen hypothalamic tissue and islets as previously described [31]. Deletion of the transcriptional blockade upstream of exon 17 in the Lepr gene was assessed using Taqman qPCR probes (IDT) (for primer sequences see Table 1; for location of primers in Lepr gene, see Figure 1A). The degree of recombination was determined using the Pfaffl method in which the deltaCT value of the floxed region of Lepr (target gene) was divided by the deltaCT value of a region upstream of the loxP sites (reference gene).

Table 1

Sequences of primers used for quantitative PCR. Primer names as depicted in Figure 1A. Primer set 1 was used to amplify sequence within the transcriptional blockade region of Lepr while Primer set 2 was used to amplify exon 17 within Lepr.

Primer name	Sequence
1(forward)	5′ GTG AGA TCA TGA GAC CCT AAA 3′
1(reverse)	5′ GGA ACT CAA GAC CAT CTA TCA 3′
1(HEX probe)	5′ TTC TGA ATT GGT GTC CCT GGA GCC 3′
2(forward)	5′ CCT TTC CAG ATA ATG CCT GAT AGA 3′
2(reverse)	5′ GCA CCA CAC TTA GCT CCA ATA 3′
2(FAM probe)	5′ TAG GGC GGA TGA ACC AGC AAA TGT 3′

Figure 1

Lepr was restored in mice with a transcriptional blockade (TB) upstream of exon 17 in Lepr by recombination of loxP sites flanking the TB using Ins1Cre (A). Percent recombination was assessed by measuring the ratio of qPCR product obtained using primers within and outside of the TB region (primers 1 and 2, respectively in A) of gDNA extracted from islets (B) and hypothalamus (C). Mice expressing mT/mG were used to determine the extent of recombination in islets with Ins1Cre. Representative images were collected from pancreata of mT/mG Ins1Cre and mT/mG mice co-stained for insulin (red) and EGFP (green) and DAPI (gray) (D; n = 3–4). Expression of GFP indicates Cre-mediated recombination. Scale bar represents 100 μm. Fasted insulin levels were measured at 5 weeks of age (E), along with fasting glucose levels (F) and body weight (G) between 5 and 12 weeks of age in male mT/mG Ins1Cre and mT/mG mice. Blood glucose was measured following an i.p. injection of insulin (0.75 U/kg body weight) to assess insulin sensitivity in 6-week old mice (H). Plasma insulin levels were measured in 8-week old mice following a gavage of 40% glucose (1.5 g/kg body weight) to assess glucose tolerance and glucose stimulated insulin secretion (I, J). Values are presented as individual data points in B, C, E, and J with a line indicating the median and the shaded area spanning the range of values. A student's t-test with P < 0.05 was used to determine statistical significance. Values in F-I are presented as mean ± SEM and a repeated one-way ANOVA with Tukey post-hoc test was used to assess statistical significance (n = 4–6/group).

Islet morphology and immunohistochemistry

Pancreata were collected from 12 to 14 week old mice, fixed overnight in 4% PFA, rinsed in 70% ethanol, embedded in paraffin, and processed for sectioning by Wax-It Histology Services (Vancouver, Canada). Three sections separated by 200 μm per mouse were immunostained for insulin (rabbit anti-insulin antibody, Cat# C27C9, 1:1000, Cell Signaling) and glucagon (mouse anti-glucagon antibody, Cat# G2654, 1:1,000; Sigma–Aldrich) or Glut-2 (rabbit anti-Glut-2 antibody, Cat# 07-1402, 1:500, Millipore) overnight at 4 °C. Slides were then incubated with AlexaFluor-conjugated secondary antibodies (AF488, AF594; Life Technologies, Burlington, Ontario, Canada) for 1 h at room temperature followed by counterstaining with DAPI to identify nuclei. Sections were scanned and quantified as previously described [32]. The insulin and glucagon positive areas are expressed relative to the whole pancreas area and averaged across three pancreas sections per mouse (n = 3–5 per group). Islet sizes were approximated by measuring insulin positive area of each islet in three pancreas sections (separated by 200 μm) per mouse. Islet sizes were then binned into ranges and the histogram analysis function in Microsoft Excel (2016) was used to determine the frequency of islets per size range. Frequency was represented as a percentage of total number of islets assessed per animal.

Statistical analysis

Statistical analysis was performed using Graphpad Prism 7.0 software (Graphpad Software, La Jolla, California) with a p value <0.05 considered statistically significant. All values are expressed either as individual data points with median or as mean ± SEM where indicated. Unless otherwise specified, a one-way ANOVA (or non-parametric Kruskal–Wallis test), two-way ANOVA, or Student's t-test were performed, with a Tukey post-hoc test.

Results

Validation of Lepr restoration in β cells

To assess recombination of the LeprloxTB/loxTB site in the presence of Ins1Cre, islets were isolated from 11 to 13 week old LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice, and gDNA was extracted for analysis by qPCR. A ∼75% rate of recombination of Lepr was observed in islets of LeprloxTB/loxTB Ins1Cre mice, while the recombination of Lepr in islets of LeprloxTB/loxTB mice was less than 10% (Figure 1B). Recombination of Lepr in islets was compared to recombination of Lepr in the hypothalamus to assess the tissue specificity of Ins1Cre. The percent of recombination in the hypothalamus was minimal and was comparable between LeprloxTB/loxTB Ins1Cre and LeprloxTB/loxTB mice (Figure 1C).

Next, we crossed Ins1Cre mice with mT/mG reporter mice to determine whether Ins1Cre produced recombination in β cells. These double fluorescent reporter mice express membrane-targeted dimer Tomato prior to Cre-mediated excision and membrane-targeted enhanced green fluorescent protein (EGFP) after excision [33]. Immunostaining of EGFP in fixed pancreas from mT/mG Ins1Cre and mT/mG control mice revealed EGFP expression co-localized in β cells of mT/mG Ins1Cre mice (Figure 1D), indicating that Cre-induced recombination occurred in β cells. To determine whether replacing 1 insulin allele with Cre influenced glucose homeostasis and metabolism, we also assessed body weight, fasting glucose, glucose tolerance, and insulin sensitivity in male mT/mG Ins1Cre mice between the ages of 4–12 weeks. Fasting insulin, blood glucose, and body weight levels were similar between mT/mG Ins1Cre mice and mT/mG mice (Figure 1E–G). Similarly, no differences in insulin sensitivity, glucose tolerance, or glucose stimulated insulin secretion were observed (Figure 1H–J). Therefore, our characterization of the Ins1Cre line confirms and extends previously published data [34] and demonstrates that knock-in of Cre into one allele of Ins1 does not appear to significantly affect glucose homeostasis.

Lepr in β cells is not sufficient to regulate islet morphology or β cell area

It has been previously reported that whole body mutation of Lepr [21] or knockout of Lepr in the pancreas of chow-fed mice [25] results in increased β cell area. To determine if restoration of Lepr in β cells is sufficient to prevent the increase in β cell area, we immunostained pancreas sections collected from 12 to 14 week old Lepr Ins1Cre, LeprloxTB/loxTB and Leprwt/wt Ins1Cre mice with insulin, glucagon, and DAPI (Figure 2A). Quantification revealed that Lepr insulin positive area was ∼10 fold higher compared to Leprwt/wt Ins1Cre mice (p < 0.01; Figure 2C). Restoration of Lepr in LeprloxTB/loxTB Ins1Cre mice resulted in a trend for reduced insulin positive area compared to LeprloxTB/loxTB mice (p = 0.06). In contrast, no differences in α cell area, as determined by glucagon positive immunostaining, were observed between experimental groups (Figure 2D). Analysis of islet size revealed a significantly lower proportion of small islets (<500 μm2) in LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice compared to Leprwt/wt mice, while LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice had a greater number of large (>10,000 μm2) islets than Leprwt/wt littermates. To determine if restoration of Lepr in β cells is able to prevent the loss of Glut-2 expression in LeprloxTB/loxTB mice, pancreas sections were co-immunostained for Glut-2, insulin, and DAPI (Figure 2B). Glut-2 immunoreactivity was clearly present in Leprwt/wt Ins1Cre mice, but generally lacking in LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice. Together, these findings suggest that direct action of leptin in β cells may not be sufficient to modify islet size or rescue Glut-2 expression in mice lacking leptin receptors elsewhere.

Figure 2

Representative images were taken from pancreas sections of Leprwt/wt, LeprloxTB/loxTB and LeprloxTB/loxTBIns1Cre mice immunostained for insulin (INS; red), glucagon (GCG; green), and DAPI (gray) (A). Presence of Glut-2 in β cells was determined by co-staining pancreas sections for insulin (red) Glut-2 (green) and DAPI (white) in Leprwt/wt, LeprloxTB/loxTB, and LeprloxTB/loxTBIns1Cre mice (B). Scale bar represents 100 μm. Relative β and α cell area were calculated by measuring total insulin and glucagon positive areas, respectively, relative to total pancreas area (C, D). Islet size was approximated by measuring insulin positive area per islet. Islet size was divided into 6 categories and frequency of islets in each category was determined relative to total number of islets measured per animal (E). Values are presented as individual data points with a line representing the median and the shaded area indicating the range of values. *P < 0.05 compared to Leprwt/wtIns1Cre. A non-parametric one-way ANOVA with Kruskal–Wallis post hoc analysis was used to determine statistical significance (n = 4–5 per group).

Restoration of Lepr in β cells has no significant impact on body weight, insulin levels, or fasting glucose levels

Along with increased insulin levels, obesity has been reported in Leprflox/flox RIPCre mice [24]. To examine the effects of restoration of Lepr in β cells, we measured body weight weekly between 6 and 11 weeks of age and collected plasma at 13 weeks of age to measure leptin levels. Body weight of male and female LeprloxTB/loxTB Ins1Cre mice was comparable to LeprloxTB/loxTB mice. Male LeprloxTB/loxTB Ins1Cre mice weighed 50% more than Leprwt/wt littermate mice, while female LeprloxTB/loxTB Ins1Cre mice weighed 80% more than their respective control littermates (Figure 3A, B). Similar to prior studies with LeprloxTB/loxTB mice [29], we observed hyperleptinemia in LeprloxTB/loxTB male and female mice, with leptin levels that were ∼150 times higher compared to Leprwt/wt Ins1Cre littermates (LeprloxTB/loxTB = 252.9 ± 26.7 ng/ml, Leprwt/wt Ins1Cre = 1.65 ± 0.20 ng/ml). Despite restoring Lepr in β cells, fasting leptin levels of LeprloxTB/loxTB Ins1Cre mice (279.8 ± 11.67 ng/ml) were comparable to those of LeprloxTB/loxTB mice. To assess the effects of β cell-specific signaling of leptin on insulin secretion and glucose homeostasis, we monitored blood glucose weekly between 6 and 11 weeks of age and collected plasma from male and female mice at 7 and 11 weeks of age to measure insulin levels. Hyperinsulinemia was evident in LeprloxTB/loxTB Ins1Cre mice and plasma insulin levels were comparable to that of LeprloxTB/loxTB mice (Figure 3C, D). In contrast, plasma insulin levels of Leprwt/wt and Leprwt/wt Ins1Cre mice were ∼20 fold lower than LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice. Similar to fasting insulin levels, fasting glucose levels were elevated in LeprloxTB/loxTB mice and LeprloxTB/loxTB Ins1Cre mice compared to Leprwt/wt and Leprwt/wt Ins1Cre mice (Figure 3E, F). However, by 10–11 weeks of age, female LeprloxTB/loxTB Ins1Cre mice exhibited significantly higher glucose levels than their LeprloxTB/loxTB littermates. Together these findings suggest that selective restoration of Lepr in β cells does not prevent obesity and hyperinsulinemia in mice lacking leptin receptors elsewhere. Moreover, restoration of Lepr in β cells does not improve fasting glucose levels of male or female LeprloxTB/loxTB Ins1Cre mice relative to their respective littermate LeprloxTB/loxTB mice.

Figure 3

Body weight was measured weekly between 6 and 11 weeks of age in male (A) and female (B) Leprwt/wt, Leprwt/wtIns1Cre, LeprloxTB/loxTB, and LeprloxTB/loxTBIns1Cre, mice. Fasting plasma insulin levels at 7 and 11 weeks of age in male (C) and female mice (D). Values are presented as individual data points with a line indicating the median and the shaded region spanning the range of values. Fasting blood glucose levels were monitored in male (E) and female (F) mice. Values in A, B, E, & F are presented as mean ± SEM. *P < 0.05 comparing LeprloxTB/loxTBIns1Cre and LeprloxTB/loxTB using a two-way repeated measures ANOVA with Tukey post hoc analysis (n = 4–13 females, n = 3–5 males).

LeprloxTB/loxTBIns1Cre mice are glucose intolerant and have an exaggerated insulin response to glucose

We next assessed the effects of restoring Lepr in β cells on insulin secretion in both male and female LeprloxTB/loxTB Ins1Cre mice. Following a 6 h fast, Leprwt/wt mice were euglycemic, while LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice were hyperglycemic (blood glucose >11 mM; Figure 4A, D). Fifteen minutes after glucose gavage, blood glucose levels of male and female Leprwt/wt and Leprwt/wt Ins1Cre mice peaked between 15 and 18 mM. In contrast, glucose levels of LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice peaked above the limit of detection (33.3 mM, represented as a dashed line on the graph), indicating they were severely glucose intolerant. Analysis of area under the curve (AUC) revealed that LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice were significantly glucose intolerant compared to control littermate mice (Figure 4B, E). In response to glucose gavage, male and female Leprwt/wt and Leprwt/wt Ins1Cre mice had fasting plasma insulin levels between 0.1 and 0.7 ng/ml, and insulin levels peaked at 0.3–1.1 ng/ml at the 15-minute time point (Figure 4C, F). In contrast, LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice had fasting insulin levels that were ∼10-fold higher than Leprwt/wt mice and peak glucose stimulated insulin secretion levels that were between 15 and 30-fold higher than Leprwt/wt mice. Fasting as well as glucose stimulated plasma insulin levels were comparable between LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice. Despite similar glucose levels in response to oral glucose gavage, basal plasma insulin levels and glucose stimulated insulin secretion was different between male and female mice. Female LeprloxTB/loxTB and LeprloxTB/loxTB Ins1Cre mice had lower basal insulin levels compared to males (Figure 3, Figure 4C, F), and insulin secretion in response to glucose was much higher in female LeprloxTB/loxTB as compared to male mice of the same genotype (Figure 4C, F).

Figure 4

Male (A) and female (D) Leprwt/wt, Leprwt/wtIns1Cre, LeprloxTB/loxTB, and LeprloxTB/loxTBIns1Cre mice between 7 and 12 weeks of age were fasted for 6 h to assess glucose tolerance by gavage of 40% glucose at a dose of 1.5 g glucose/kg body weight. Blood glucose was measured every 15 min and area under the curve was analyzed (B, E). Blood collected at 0, 15 and 60 min of the glucose tolerance test was used to measure plasma insulin levels (C, F). Values are presented as mean ± SEM in panels A, C, D, and F while values in B and E are presented individually with a line indicating the median. A one-way ANOVA with Tukey post hoc analysis was used to assess statistical significance (n = 3–7 per group).

Discussion

To examine the effects of restoring Lepr in β cells on insulin secretion and morphology, we crossed mice expressing Cre knocked into the Ins1 gene with LeprloxTB/loxTB mice, which have a transcriptional blockade inserted between exon 16 and exon 17 of Lepr, resulting in impaired leptin signaling. Disruption of the leptin receptor signaling domain in LeprloxTB/loxTB mice results in a phenotype similar to db/db mice, characterized by obesity, hyperinsulinemia, hyperglycemia, insulin resistance, and islet hyperplasia [35], [36]. Cre-mediated removal of the transcriptional blockade and restoration of Lepr can be achieved in LeprloxTB/loxTB mice in a cell-type selective manner, enabling the assessment of the physiological impact of Lepr in specific tissues. Restoration of Lepr in β cells of LeprloxTB/loxTB Ins1Cre mice did not alter the development of hyperinsulinemia, hyperglycemia, glucose intolerance, and obesity, suggesting that leptin signaling in β cells does not suppress insulin secretion. Moreover, islet hyperplasia was still present in LeprloxTB/loxTB Ins1Cre mice, similarly to what is reported in db/db mice [35]. However, as we did not follow the mice for an extended period of time, our studies cannot rule out the possibility that the restoration of Lepr in β cells may have prevented age related β cell failure that eventually results in overt severe diabetes, like in db/db mice [37], [38].

A distinct difference between the present study and previous studies in which Lepr was selectively deleted [24], [25], [28] is the inability of compensation to occur from non-β cells in LeprloxTB/loxTB Ins1Cre mice. In prior studies that examined the role of leptin signaling in β cells by deletion of Lepr, it is unclear whether Lepr acting in other tissues influenced pancreas morphology and β cell function. The ability of non-β cell leptin receptor action to influence β cell function is evident in mice with restored Lepr in brains of db/db mice. Overexpression of Leprb driven by synapsin 1 and neuron specific enolase promoters in db/db mice resulted in normalized islet morphology and partial correction of body weight [39]. These authors concluded that Lepr in the CNS is able to rescue the diabetic and obese phenotype of db/db mice. To prevent Lepr in non-β cells from influencing β cell function, we selectively re-expressed Lepr in β cells of mice lacking Lepr elsewhere. Our results suggest that Lepr in β cells is not sufficient to prevent hyperinsulinemia and glucose intolerance from occurring. It is likely that Lepr in the CNS and other islet cell populations are involved in inhibiting insulin secretion and preventing hyperplasia of β cells.

Previous studies have reported Lepr expression in β cells of human and rodent islets and cell lines [9], [40], [41], [42], [43]; however, in another study, non-detectable levels of Lepr were reported in β cells from dispersed islets using RT-qPCR [28]. More recently, studies using single-cell RNA-seq to determine the transcriptome profiles of human and murine islets revealed that the expression of Lepr was higher in δ cells than β cells [44], [45]. Though insulin secretion and expression was reduced in isolated murine and human islets treated with leptin [8], [10], [12], this effect may have been a result of leptin acting on δ cells. Collectively, this suggests that despite restoration of Lepr in LeprloxTB/loxTB Ins1Cre mice, the relatively low levels of Lepr in β cells may not be sufficient to provoke improvements in β cell function in the absence of leptin receptor signaling in other pancreatic islet populations.

The presence of hyperinsulinemia and islet hyperplasia despite restoration of Lepr in β cells of LeprloxTB/loxTB mice may be due to hyperleptinemia that was present in LeprloxTB/loxTB Ins1Cre mice. Plasma leptin levels were ∼150-fold higher in LeprloxTB/loxTB Ins1Cre mice compared to control mice. Hyperleptinemia has been previously observed in LeprloxTB/loxTB [29], and in mice with disruption of Lepr in peripheral tissues [46]. This increase of plasma leptin levels was attributed to the disruption of Lepr in white adipose tissue (WAT), which led to inactivation of a negative feedback loop regulating leptin synthesis in WAT [46]. Indeed the presence of hyperleptinemia can cause leptin resistance [47]. Thus, despite restoration of Lepr in β cells, the chronically elevated leptin levels in LeprloxTB/loxTB Ins1Cre mice may have resulted in a state of leptin resistance in β cells of LeprloxTB/loxTB Ins1Cre mice.

Conclusion

In summary, selective restoration of Lepr in pancreatic β cells of Lepr knockout mice does not prevent the development of hyperinsulinemia and β cell expansion. These findings suggest that leptin receptor signaling directly in β cells does not regulate β cell function. We propose that actions of leptin on β cells are secondary to action of leptin signaling in other tissues. Whether this involves a combination of leptin signaling in other islet cell populations and in the CNS warrants further study.