Loss of growth hormone signaling in the mouse germline or in adulthood reduces islet mass and alters islet function with notable sex differences

Abstract

In the endocrine pancreas, growth hormone (GH) is known to promote pancreatic islet growth and insulin secretion. In this study, we show that GH receptor (GHR) loss in the germline and in adulthood impacts islet mass in general but more profoundly in male mice. GHR knockout (GHRKO) mice have enhanced insulin sensitivity and low circulating insulin. We show that the total cross-sectional area of isolated islets (estimated islet mass) was reduced by 72% in male but by only 29% in female GHRKO mice compared with wild-type controls. Also, islets from GHRKO mice secreted ∼50% less glucose-stimulated insulin compared with size-matched islets from wild-type mice. We next used mice with a floxed Ghr gene to knock down the GHR in adult mice at 6 mo of age (6mGHRKO) and examined the impact on glucose and islet metabolism. By 12 mo of age, female 6mGHRKO mice had increased body fat and reduced islet mass but had no change in glucose tolerance or insulin sensitivity. However, male 6mGHRKO mice had nearly twice as much body fat, substantially reduced islet mass, and enhanced insulin sensitivity, but no change in glucose tolerance. Despite large losses in islet mass, glucose-stimulated insulin secretion from isolated islets was not significantly different between male 6mGHRKO and controls, whereas isolated islets from female 6mGHRKO mice showed increased glucose-stimulated insulin release. Our findings demonstrate the importance of GH to islet mass throughout life and that unique sex-specific adaptations to the loss of GH signaling allow mice to maintain normal glucose metabolism.

NEW & NOTEWORTHY Growth hormone (GH) is important for more than just growth. GH helps to maintain pancreatic islet mass and insulin secretion throughout life. Sex-specific adaptations to the loss of GH signaling allow mice to maintain normal glucose regulation despite losing islet mass.

INTRODUCTION

Growth hormone (GH) is a peptide hormone secreted by the somatotroph cells of the anterior pituitary (1, 2). It binds to the GH receptor (GHR), which is found on almost all cells/tissues with its main targets being liver, adipose tissue (AT), bone, kidney, and muscle. Although GH is well known for its ability to induce postnatal growth, it is involved in numerous other processes including lipid, carbohydrate, nitrogen, and mineral metabolism (3, 4). GH was identified as early as 1930s as a “diabetogenic hormone” (5, 6) primarily due to the stimulation of lipolysis to induce insulin resistance (7–10). Therefore, to obtain the required energy in times of stress, GH promotes the use of lipids instead of consumption of carbohydrates and protein (11, 12). Because of this, GH excess is associated with poor glucose homeostasis, increased circulating glucose, hyperinsulinemia, and a reduction of insulin receptor levels in the membrane of peripheral tissues (13).

Due to GH’s physiological actions, mice with excess GH in circulation, termed bovine GH (bGH) transgenic mice, are giant, insulin insensitive, and prone to glucose intolerance and diabetes (14). In contrast, mice insensitive to GH due to a germline inactivating mutation in the extracellular domain of the GHR (GHR^−/− mice, also called GHRKO or Laron mice) (15) are characterized by smaller size, decreased circulating insulin, improved insulin sensitivity, resistance to diet-induced diabetes and cancer, and increased longevity (16–19). Importantly, these small mice have clinical significance as a model of Laron Syndrome (LS), that is, patients who also carry a germline inactivating mutation in the GHR. Similar to GHRKO mice, these patients have decreased body size and increased adiposity, with low levels of insulin-like growth factor-1 (IGF-1) and high levels of GH; they also have a very low incidence of cancer and diabetes (20). Furthermore, mice with disruption of the GHR just after sexual maturation (∼6 wk of age) are also highly insulin sensitive, have low circulating IGF-1 and insulin levels, and females have extended longevity (21).

Although GH has well-established metabolic effects in bone, muscle, fat, and liver (12, 22), the fact that GH also has important effects on many other tissues is often overlooked. One organ that is sometime forgotten in discussions of GH and glucose metabolism is the endocrine pancreas (23). Islets of Langerhans are micro-organs within the pancreas that regulate blood glucose by secreting hormones such as insulin from pancreatic β cells that signal to peripheral tissues such as liver, fat, and muscle to take up glucose (23, 24). Therefore, these cells are key players in glucose homeostasis, and their loss or failure to secrete insulin appropriately leads to the development of diabetes (25). Studies of GH action in primary pancreatic islets or β cells have demonstrated two key effects: 1) increased proliferation of pancreatic β cells (26–28) and 2) increased glucose-stimulated insulin secretion (29–31). In addition, GH can impact β-cell growth (32, 33). For example, the germline GHRKO mice displayed reduced pancreatic size and β-cell mass (34). These findings suggest that GHR-induced signaling in the pancreas plays a key role in regulating glucose metabolism. However, isolated islets from GHRKO mice have never been examined in vitro to directly assess islets in the absence of neural, humoral, and other factors.

Much of the activity of GH as a driver of development, growth, and metabolism is heavily influenced by sex hormones. Circulating GH levels have been reported to be as much as 125 times higher in women compared with men (35), which results in substantial differences in downstream signaling. For example, GH administration (adjusted for body mass) results in substantially lower insulin-like growth factor-1 (IGF-1) levels in women compared with men. This increase of GH levels in premenopausal women is thought to be a result of the inhibitory effect of estrogen on GH-induced intracellular signaling (36, 37). These sex differences in GH action have effects throughout the body. GH-related action is arguably the single most important cause of sex differences in bone mass (38), and sensitivity in bone mass changes in response to long-term GH treatment is also sex dependent (39). Sex differences have also been observed in gene expression patterns in the liver related to GH signaling (40, 41). Even the neuroendocrine control of GH pulsatility shows strong sex differences, with men having much stronger pulsatility compared with women (42).

In the present study, we aimed to evaluate whether sex differences and developmental stage has an impact on the size and function of islets. Therefore, islets harvested from both male and female mice with disrupted GHR at two different developmental stages, germline and adulthood, were used for this study. Experiments were performed in both male and female germline GHRKO and mice with disrupted GHR after 6 mo of age (6mGHRKO). We showed for the first time that the age and sex at which GH action is disrupted impacts pancreatic function, with germline male GHRKO mice showing a larger reduction in islet size relative to controls than observed with female mice. Furthermore, we found that disruption of GHR at an adult age results in reduced islet size in both male and female 6mGHRKO mice when compared with controls, although, again, this reduction is greater in male mice. In addition, islets of GHRKO and 6mGHRKO mice showed sex differences in glucose-stimulated insulin secretion. Collectively, these data suggest that GH action is needed throughout life to maintain islet cell mass and function, and that sexual dimorphism needs to be considered when evaluating the effects of GH action in glucose metabolism.

MATERIALS AND METHODS

Mouse Housing and Breeding

The GHRKO mouse line used in this study has been described previously (15). For all studies, adult male and female GHRKO mice ages 13.5 ± 4.5 mo were used, with wild-type (WT) littermates serving as controls. Male and female GHRKO mice (18-mo-old) were used for evaluation of glucose and insulin tolerances.

The 6mGHRKO mice were generated by crossing C57BL/6J mice carrying a floxed Ghr allele (43, 44) with mice that express an inducible ubiquitous Cre recombinase gene, transcription of which is driven by the ROSA26 gene promoter/enhancer (ROSA26-Cre-ERT2) [B6.129Gt(ROSA)26Sortm1(cre/ERT2)Tyj/J mice] from the Jackson Laboratory. Homozygous mice for both the floxed Ghr gene and the Cre alleles were bred as described previously (21). To induce Ghr gene disruption and produce the 6mGHRKO mice, 6-mo-old mice received 100-μL intraperitoneal (ip) injections of tamoxifen dissolved in peanut oil once per day over 5 consecutive days, for a total dose of 0.32 mg of tamoxifen/g of body weight. Control 6mGHRKO mice received injections of only peanut oil.

Note that we did not specifically control for tamoxifen effects because our assessment points were 6 mo after the last tamoxifen injection and unlikely to have effects. Although histological changes in tissues have been observed following tamoxifen injections at similar concentrations, these effects were reversed by 28 days (45). Furthermore, tamoxifen has been shown to enhance glucose tolerance (46), which was not observed in our study. Our findings are thus consistent with changes in GHR signaling and not likely due to tamoxifen. Also, although expression of GHR in no-Cre mice was not assessed, several indirect measurements such as body composition and femur length show that the GHR was only ablated after tamoxifen treatment (data not shown).

Mice were housed at 22°C under a 14-h light, 10-h dark cycle, 3–4 mice per cage, with ad libitum access to water and standard laboratory chow (ProLab RMH 3000). To collect the pancreas and pancreatic islets, mice were euthanized by exposure to CO₂ before islet isolation. Mice were dissected at ∼12 mo of age after a 12-h overnight fast. All experiments were approved by the Ohio University Animal Care and Use Committee.

Glucose and Insulin Tolerance Tests

Protocol for 6mGHRKO mice.

Glucose tolerance tests (GTTs) and insulin tolerance tests (ITTs) were performed on male and female 6mGHRKO and controls of 11 mo of age. For GTTs, filtered PBS was used to prepare the 10% glucose solution. Mice were fasted overnight before receiving an intraperitoneal injection of 0.01 mL glucose solution/g body weight (1.0 mg/kg body wt). For ITTs, recombinant human insulin (Novolin-R; Novo Nordisk) was diluted in filtered PBS for a final concentration of 0.075 U/mL. Mice were fasted for 6 h before receiving an intraperitoneal injection of 0.01 mL insulin solution/g body weight (0.75 U/kg). For both tests, blood glucose was measured before injections and 15, 30, 45, 60, and 90 min after injections. All glucose measurements were taken with tail-snip blood using OneTouch Ultra glucose strips and glucometers (Lifescan, Inc., Milpitas, CA).

Protocol for GHRKO mice.

GTTs and ITTs were performed as described for 6mGHRKO mice with the following exceptions. For the GTT, each mouse was given an intraperitoneal injection of 15% glucose at 0.01 mL/g body weight after fasting for 6 h (1.5 mg/kg body wt). Blood collections were made at 0, 20, 60, and 120 min postinjection. For the ITT, mice were fasted for 5 h, and each mouse was given 0.075 U/mL at 0.01 mL/g body weight (0.75 U/kg) of recombinant human insulin (Humulin-R, Eli Lilly & Co, Indianapolis, IN). Blood was collected at 0, 20, 40, 60, and 90 min postinjection. These exceptions were made to avoid stress to the GHRKO mice.

Staining and Quantification of Pancreatic Sections

The pancreas was harvested and placed in dry tubes for freezing in liquid nitrogen for a subset of 6mGHRKO (n = 4 male, n = 4 female) and control (n = 4 male, n = 4 female) mice. Pancreata were subsequently fixed in 10% formalin, blocked, and sectioned at 5-μm intervals. Three contiguous 5-μm sections were chosen for hematoxylin and eosin (H&E) staining for each mouse to identify the largest area among the sections for each islet; each islet was counted once. This procedure was similarly repeated in another set of pancreas sections located more than 200 μm away from the first. Individual islets were identified by differential H&E staining and photographed with a digital camera mounted to a light microscope. For each digital image, each islet was traced using computer software ImageJ to quantify the surface area for each islet. Each digital photo was magnified so that the nuclei of cells contained in the islets could be counted manually. Data collected for each islet per mouse were normalized to the area of each islet to determine the number of nuclei per square millimeter for male and female control and 6mGHRKO mice.

Insulin and glucagon staining were performed on 5-μm-thick pancreatic sections using standard deparaffin and rehydration procedures and using an anti-insulin antibody [Abcam, #ab7842 (1:200 dilution), guinea pig polyclonal to insulin, IgG isotype] and antiglucagon antibody [Abcam #ab10988 (1:200 dilution), mouse monoclonal to glucagon, IgG1 isotype]. To visualize insulin and glucagon in the pancreatic slices, secondary antibodies conjugated to Alexa488 (insulin) and Alexa594 (glucagon) were used. The images were acquired on a Nikon Microphot-SA upright fluorescent microscope. ImageJ was used to determine total cross-sectional area of hand-drawn regions of interest around each islet circumference. The ImageJ thresholding function was used to determine the area of glucagon-positive cells (red) and insulin-positive cells (green) within each islet to estimate α-cell and β-cell mass, respectively. The remaining area was considered as other cells types. Note that this technique does not count the number of each cell type; it determines the area of the fluorescence signal to estimate the “mass” of each cell type in each islet. Thus, α cells, which are generally smaller than other pancreatic endocrine cells, may be undercounted with this technique.

Islet Isolation and Islet Area Assessment

Pancreatic islets were isolated by collagenase-P digestion (Roche Diagnostics, Indianapolis, IN) followed by centrifugation with Histopaque 1100 (Sigma-Aldrich, St. Louis, MO), as previously described (47). Islets were incubated overnight in RPMI 1640 medium (Invitrogen) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin to allow recovery from collagenase digestion before further treatment.

The day after isolation, islets were imaged using an EVOS XL Core from Life Technologies using a ×4 magnification objective lens. The perimeter of each islet was traced using National Institutes of Health (NIH) ImageJ to calculate the cross-sectional area. It should be noted that islet isolation procedures typically yield ∼15%–25% of the total islet content within a pancreas, thus providing a large sample size of the islet population. The typical murine pancreas contains ∼1,500 islets; we retrieved 200–400 islets per pancreas for 6mGHRKO mice (less for GHRKO). Effective canulation, collagenase digestion, and other factors can impact islet yield (47). The estimated diameter of islets from our control group was ∼100–200 µm, which is typical for an average size mouse islet (48, 49). A rigorous histological study of entire pancreata section-by-section in young rats estimated the maximum diameter of islets from area measurements at ∼100–150 µm (50). These studies suggest that measurements of islet area from histological sections and isolated islets are within the same range. It must be stressed that rigorous side-by-side testing of both techniques has not been conducted to fully validate the use of isolated islets to estimate average size.

Insulin Secretion

Sets of 20 islets/well in 12-well or 24-well plates were used for all studies of insulin secretion, as previously described (51). Islets were carefully selected to be size-matched, so that relative insulin secretion rates can be compared accurately since differences in islet size can affect insulin secretion. We have previously shown that size-matching is as effective as normalizing to insulin, protein, or RNA content (52). Because changes in islet composition could affect these results, we stained islets for insulin in parallel studies to show that the percentage of β cells was ∼70% regardless of sex or genotype. Media was collected at ∼24-h intervals or during glucose-stimulated insulin secretion (GSIS) tests using 3 mM and 11 mM glucose for GHRKO and 3 mM and 28 mM glucose stimulation for 6mGHRKO genotypes. Note that the glucose stimulation index was reported for GHRKOs because the same islets were tested sequentially, whereas separate islets were stimulated in 3 or 28 mM glucose for 6mGHRKOs. Insulin secretion was measured by standard insulin enzyme-linked immunosorbent assay (ELISA) following manufacturer’s directions (ALPCO, Salem, NH). Intra-assay variability was kept <10% for all studies.

Statistical Analysis

Differences in mouse body weight, body fat percentage, fasting blood glucose, GTT, and ITT were analyzed by one-way analysis of variance (ANOVA) with Tukey honestly significant difference (HSD) post hoc analysis focused on GHR mutant mice (GHRKO or 6mGHRKO) versus controls for each sex. Area under the curve (AUC) was performed to analyze GTTs and IITs. All data collected from islets were analyzed for GHR mutant mice versus control comparisons for each sex separately by two-tailed Student t test unless otherwise stated. Significance was defined as P < 0.05, near significance as P < 0.10 (but P > 0.05), and not significant as P > 0.10.

RESULTS

Male and Female GHRKO (Laron) Mice Have Altered Physiology and Metabolism

We first examined basic metabolic parameters of body weight, fasting blood glucose, glucose tolerance, and insulin tolerance. We found that body weight was substantially lower for both male and female GHRKO mice compared with WT controls, with no sex differences observed (Fig. 1A). As shown in Fig. 1B, fasting blood glucose was significantly lower for female (P < 0.001) GHRKO mice, but not male (P = 0.26).

Figure 1.

Physiological characteristics of GHRKO mice. Body weight (A) and fasted blood glucose levels (B) before glucose tolerance test for the following test groups: wild-type male (WT M, n = 9), growth hormone (GH) receptor knockout male (GHRKO M, n = 9), WT female (WT F, n = 14), and GHRKO female (GHRKO F, n = 12). C: serum insulin levels from n = 4 WT vs. n = 4 GHRKO mice (***P < 0.005). D: glucose tolerance test (GTT): mean blood glucose values at 0, 20, 60, and 120 min following a glucose bolus of 0.01 mL/g body wt. E: area under the curve (AUC) of the GTT. F: insulin tolerance test (ITT): mean blood glucose values at 0, 20, 40, 60, and 90 min following an insulin bolus of 0.075 U/mL at 0.01 mL/g body wt. G: AUC of the ITT. N.S., not significant, *P < 0.05, ***P < 0.001. Mice were 18 mo old for all studies shown in this figure. Note that GHRKO was shortened to “KO” in D and F. GHRKO, growth hormone receptor knockout.

GHRKO mice have low levels of circulating insulin, smaller numbers of islets per pancreas, and smaller islet areas as measured from pancreatic sections (34). In Fig. 1C, we confirm that GHRKO mice have reduced levels of circulating insulin compared with WT controls (n = 4 GHRKO mice, n = 4 WT, P < 0.001). Differences in serum insulin could not be properly evaluated for each sex separately due to the small sample size (n = 2). However, prior studies noted no sex differences in serum insulin levels in this mouse model (14, 15, 34), which is consistent with these limited observations.

We next examined glucose tolerance and insulin sensitivity. Blood glucose during the glucose tolerance test was lower in female GHRKO mice compared with controls at 60 min (P < 0.05) and 120 min (P < 0.001), but not at 20 min (Fig. 1D). The area under the curve (AUC) was significantly lower for female GHRKO mice compared with female WT; however, this appears to be due to reduced glucose tolerance in the WT females. No differences in glucose tolerance were observed among the males (Fig. 1, D and E). To determine whether the glucose tolerance differences were due to enhanced insulin sensitivity or due to factors associated with insulin secretion/levels, we conducted an insulin tolerance test. Both male and female GHRKO mice showed exceptionally high insulin sensitivity; however, there were no sex-dependent differences (Fig. 1, F and G). Note that WT mice did not show robust responses to the ITT because the insulin dose (0.075 U/mL at 0.01 mL/g body wt) was chosen to avoid severe hypoglycemic events that occur with higher insulin doses in GHRKO mice due to their higher insulin sensitivity. We conclude that the decrease in glucose tolerance in WT females relative to GHRKO mice does not appear to be due to changes in insulin sensitivity.

Glucose-Stimulated Insulin Secretion Is Reduced in Islets from GHRKO Mice

Isolation of pancreatic islets from GHRKO mice is a substantial challenge because of the small size of the mice and their pancreata. No study of isolated pancreatic islets from GHRKO mice has been reported to date. As shown in Fig. 2, we report for the first time that islets isolated from male GHRKO mice secrete substantially less insulin than WT controls using size-matched islets from each genotype. GHRKO islets from male mice secreted significantly less insulin when treated with low (3 mM) glucose compared with WT mice (Fig. 2A). When stimulated with11 mM glucose, islets from male GHRKO mice also had reduced insulin secretion near significance compared with controls (Fig. 2A, P = 0.08). Islets from female GHRKO mice did not show a statistically significant difference from WT controls for 3 mM or 11 mM glucose (Fig. 2B), although it should be noted that both sexes of the GHRKO mice showed an approximate 50% reduction in stimulated insulin secretion (11 mM) compared with their respective controls. The glucose stimulation index (insulin secretion in 11 mM glucose divided by secretion in 3 mM glucose) was significantly greater for male GHRKO mice relative to WT controls (Fig. 2C), largely due to very low insulin secretion in 3 mM glucose. The glucose stimulation index for female mice, however, did not significantly differ between genotypes (P = 0.16), but was in the opposite direction as males.

Figure 2.

Glucose-stimulated insulin secretion is reduced in islets isolated from GHRKO mice. A and B: insulin released by islets from GHRKO and wild type (WT) mice during 1-h incubation in 3 mM glucose (3 G) followed by 1 h in 11 mM glucose (11 G). n = 8 sets of 20 islets collected from n = 4 mice per sex and per genotype for a total of 16 mice. A: wild-type male (WT M) and GHRKO male (KO M). B: WT and GHRKO female (GHRKO F). All sexes and genotypes showed significant glucose stimulation (difference in insulin secretion in 3 vs. 11 mM glucose) except for islets from female GHRKO mice. C: glucose stimulation index (secretion in 11 mM divided by 3 mM glucose) for each sex compared by genotype. N.S., not significant. *P < 0.05, #P < 0.10. Isolated islets were collected from mice at approximately 1 yr of age. GHRKO, growth hormone receptor knockout.

Islet Size Is Significantly but Differentially Reduced in Male and Female GHRKO Mice

Previous work using measurements of islet areas within pancreatic sections indicated that islets were substantially reduced in size in GHRKO mice (34). However, this study did not note any sex differences among the male and female mice studied (34). We measured the cross-sectional area for nearly 2,000 islets isolated from male and female GHRKO and WT mice. As shown in Fig. 3, male GHRKO mice showed smaller and decreased number of isolated islets compared with WT mice (Fig. 3, A and B). Furthermore, smaller and fewer isolated islets were also seen in female GHRKO mice compared with controls. This difference was not as pronounced as in male mice (Fig. 3, D and E).

Figure 3.

GHRKO mice have significantly reduced islet size. A and B images of isolated islets collected from a male wild-type (WT) mouse (A) and a male GHRKO mouse (B). C: cross-sectional areas of islets from male GHRKO mice and WT controls. Images of isolated islets collected from a female WT mouse (D) and a female GHRKO mouse (E). F: cross-sectional areas of islets from female GHRKO mice and WT controls. Percent of total islets from male (G) and female (H) mice organized by size into three categories as indicated on the x-axis. n = 509 islets from six male GHRKO; n = 827 islets from six male WT; n = 271 from five female GHRKO; n = 353 islets from five female WT mice. ***P < 0.001 by islet (P < 0.002 by mouse for each sex). Isolated islets were collected from mice at approximately 1 yr of age. GHRKO, growth hormone receptor knockout.

The measured cross-sectional area of male GHRKO islets was only 28% that of WT controls (Fig. 3C, P < 0.001, t test). Islets isolated from female mice showed a much smaller difference, with a mean area of 71% of WT controls (Fig. 3F, P < 0.001, t test). Similar statistical differences were found comparing islet size by mouse for males (5 WT and 6 GHRKO mice, P < 0.01) and females (5 WT and 5 GHRKO mice, P < 0.01). When the average islet area for each mouse was compared using two-way ANOVA, significant differences were observed between WT and GHRKO mice (P < 0.001) and between males and females (P = 0.05), suggesting a sex difference in islet mass due to GHR deficiency. Using a one-way ANOVA, male WT islets were significantly larger than both male and female GHRKO islets, but male GHRKO islets were not statistically different in size from female GHRKO islets. Thus, this sex difference can be viewed as islets from male GHRKO mice being “female-like” in size, as opposed to “male-like.” Note that in this statistical test, islet size from female WT and female GHRKO mice did not differ significantly.

We also observed that the distribution of islets from GHRKO mice is shifted toward smaller islets, especially among islets from male mice. As shown in Fig. 3G for males, 33.8% of WT control islets had areas >32,000 µm² (diameter > ∼200 µm), whereas only 1.2% of GHRKO islets had areas >32,000 µm² (P < 0.0001). In contrast, nearly half (44.2%) of all GHRKO islets had cross-sectional areas <8,000 µm², (diameter < ∼100 µm), whereas only 5.8% of the islets from WT controls were in this range (P < 0.0001 by Fisher’s exact test). Islets from female GHRKO mice also showed smaller islet size than female controls, but to much less of a degree. As shown in Fig. 3H, 36.9% of islets from female GHRKO mice were <8,000 µm² in cross-sectional area versus 11.4% for female WT mice (P < 0.0001). Our observed range of islet sizes agrees with established islet sizes detected by whole pancreas sectioning in rats (50). Furthermore, a prior study of GHRKO mice similarly showed fewer large islets, more smaller islets, and similar numbers of medium-sized islets in GHRKO pancreatic sections compared with WT (34).

Effects of Disrupted GH Signaling on Physiology and Metabolism in Adulthood

Because GH action impacts body composition and glucose metabolism, we next examined basic metabolic parameters in mice with normal GH signaling throughout life into adulthood. That is, mice that had the Ghr gene ablated after they reached 6 mo of age (6mGHRKO mice) were used for these experiments. Ghr gene expression was shown to be reduced ∼50% or more in all tissues tested at 1 yr and 2 yr of age (53). Mice were examined 6 mo after Ghr gene disruption to compare 6mGHRKO and peanut oil-injected control mice for changes in physiology and metabolism. As shown in Fig. 4A, both male and female 6mGHRKO mice showed an average 1–2 g decrease in body mass (not significant) compared with their respective controls. As expected, ablation of the GHR significantly increased fat mass percentage in both male and female 6mGHRKO mice compared with controls (Fig. 4B). This result is consistent with previous reports of other mouse lines with reduced GH action (14). The increased fat mass was more evident in males; percent body fat nearly doubled due to GHR deficiency. These metabolic changes did not appear to significantly increase elevated fasting blood glucose levels for either male or female mice (Fig. 4C). Despite significant changes in percent body fat, glucose tolerance tests did not differ in 6mGHRKO mice compared with controls for either male (Fig. 4D) or female mice (Fig. 4E). Insulin sensitivity at 12 mo of age was enhanced in male 6mGHRKO mice compared with controls (Fig. 4F), however, female mice did not show any differences in insulin tolerance (Fig. 4G).

Figure 4.

Differences in metabolism between male and female 6mGHRKO mice. Means ± SE for body weight (A), percent body fat measured by magnetic resonance imaging (B), and 6-h fasted blood glucose (C) for male and female 6mGHRKO at 1 yr of age. n = 26 control male; n = 26 6mGHRKO (labeled 6mKO) male; n = 27 control female; n = 27 6mGHRKO (labeled 6mKO) female. *P < 0.05, ***P < 0.001, not significant (N.S., P > 0.10). Glucose tolerance tests (GTTs) were performed at 11 mo of age, following 6 h of fasting for both male (D) and female (E) mice. Glucose response data were analyzed by repeated-measures ANOVA from calculated area under the curve (AUC). Insulin tolerance tests (ITTs) were performed at 11 mo of age, following a 6 h fast for both male (F) and female (G) mice. Glucose response data were analyzed by repeated-measures ANOVA from calculated AUC. All values are means ± SD for D–G. n = 10 mice per group, N.S., not significant; *P < 0.05 between control and 6mGHRKO mice of same sex. GHRKO, growth hormone receptor knockout.

Pancreatic Sections Show That Islets Are Much Smaller in Male 6mGHRKO Mice Compared with Controls

We fixed and sectioned the pancreas from 6mGHRKO and control mice to determine whether there were differences in islet size as reported in global GHRKO mice. Figure 5A illustrates a representative islet from a control male pancreas section, whereas a typical male 6mGHRKO islet is shown in Fig. 5B. The male control islets had an average diameter of 108.5 ± 25.8 µm whereas islets from the male 6mGHRKO mice had an average diameter of 63.6 ± 10.7 µm. The overall average size of islets was significantly smaller in 6mGHRKO males, with the islets being ∼64% smaller than their counterparts (Fig. 5C). As shown in Fig. 5D, the size of individual cells within each islet appeared to be smaller in islets from 6mGHRKO mice as evidenced by a significant increase in the density of nuclei per unit area of each islet (P = 0.0015).

Figure 5.

Pancreatic sectioning shows substantially smaller islets from male 6mGHRKO mice due in part to smaller cell bodies. Images of pancreatic sections from male control (A) and male 6mGHRKO (B) mice. C: comparison of islet surface area ± SE by pixel tracing and subsequent micrometer conversions between control (n = 6) and 6mGHRKO male mice (n = 6) *P < 0.05. D: mean number of nuclei counted per islet area in millimeter square ± SE for all islets from male control (n = 41 islets) and male 6mGHRKO mice (n = 40 islets) ***P = 0.0015 by t test. Images of individual islets within pancreatic sections from female control (E) and female 6mGHRKO (F) mice. G: comparison of islet surface area ± SE by pixel tracing and subsequent micrometer conversions between n = 6 control and n = 8 6mGHRKO female mice. H: mean number of nuclei counted per islet area in millimeter square ± SE for all islets from female control (n = 34 islets) and female 6mGHRKO mice (n = 34 islets). Not significant (N.S., P = 0.91). Pancreas sections were collected from mice at approximately 1 yr of age.

For female mice, the typical islet from female controls and 6mGHRKO are shown in Fig. 5, E and F, respectively. The size of islets found in sections from female control mice had an average diameter of 80.7 ± 9.6 µm, whereas the female 6mGHRKO mice had a diameter of 74.6 ± 8.3 µm. Although large and small islets were observed from both 6mGHRKO and control mice, the overall average cross-sectional area of islets from female 6mGHRKO mice were ∼17% smaller than female controls (Fig. 5G), although this difference was not significant. The number of islets per pancreatic slide did not differ between female 6mGHRKO and control mice nor did the density of nuclei per unit area of each islet (Fig. 5H).

Insulin and Glucagon Staining Suggests GHR Deficiency at an Adult Age Does Not Impact Islet Architecture or Composition

To determine whether GHR deficiency in adulthood causes changes in the composition of islets, we stained pancreatic sections for insulin (green) and glucagon (red). As shown in Fig. 6A, islets were smaller among male 6mGHRKO mice compared with controls as expected. However, islets appeared to maintain similar architecture for each genotype with respect to insulin-staining β cells in the core (center) and glucagon-staining α cells near the periphery. Figure 6B shows that the islet surface area consisting of insulin-staining β cells is lower for male 6mGHRKO mice. The same is true for α cells to an even greater extent, causing the average male 6mGHRKO islet area to be 43% smaller than their control. The surface area of each cell type as a percent of the total islet composition was not significantly different for β cells between control and 6mGHRKO for males. However, the percentage of α cells in the male 6mGHRKO islets was lower at near significance (#P < 0.10) and “other” islets cell types were higher at near significance (#P < 0.10) compared with controls (Fig. 6C). In contrast, there were no significant changes in any aspect of islet composition or architecture in females as shown in Fig. 6, D–F. The average female 6mGHRKO islet area was only 8% smaller than the respective control.

Figure 6.

Insulin and glucagon staining of pancreatic sections shows that islets from 6mGHRKO mice have a similar composition of cell types when compared with controls, but male 6mGHRKO islets are smaller. A: example of male control and 6mGHRKO islets stained with insulin (green) and glucagon (red). B: surface area determined from fluorescence coverage of green (β cells), red (α cells), or no staining (other) for n = 19 islets (4 mice) control and n = 32 islets (5 mice) 6mGHRKO. C: green, red, and no staining normalized to total islet area to produce an estimated percent of islet mass for each cell type n = 19 islets (4 mice) control and n = 32 islets (5 mice) 6mGHRKO. *P < 0.05, **P < 0.01, ***P < 0.001, #P < 0.10. D–F: islet composition in female mice using the same description as for (A–C); n = 22 islets (5 mice) control and n = 30 islets (5 mice) 6mGHRKO. Pancreatic sections were collected from mice at approximately 1 yr of age. GHRKO, growth hormone receptor knockout.

Islets Isolated from Mice Support Histological Findings of Smaller Islet Size in 6mGHRKO Males

Isolated islets were used to further document the islet size measurements and to assess islet function outside of the in vivo environment. By taking images and tracing the circumference of each islet, we were able to determine the cross-sectional area and approximate diameter of each islet. We found that although islet sizes showed a large variability (Fig. 7, A and B), the overall size of islets isolated from male 6mGHRKO mice was significantly smaller than controls by 34% (Fig. 7C). For females, although there were large and small islets isolated from both 6mGHRKO and control mice (Fig. 7, D and E), the overall size of islets isolated from 6mGHRKO mice was significantly reduced by 22% compared with controls (Fig. 7F). Additional analysis by two-way ANOVA showed differences by sex and by genotype (P < 0.05 for both). One-way ANOVA with Tukey HSD analysis showed that the only significant difference was between WT male and 6mGHRKO female islet areas. As shown in Fig. 7, G and H, the distribution of islets areas for 6mGHRKO mice showed a similar pattern of fewer large islets and more small islets as observed with GHRKO mice. These shifts in distribution were significant for both male and female by Fisher’s exact test (P < 0.001) for large (>32,000 µm²) and small islets (<8,000 µm²), but not significant for medium-sized islets (between 8,000 and 32,000 µm²).

Figure 7.

6mGHRKO mice have significantly reduced islet size. A and B: images of isolated islets collected from a male control mouse (A) and a male 6mGHRKO mouse (B). C: cross-sectional areas of islets from male 6mGHRKO mice and controls. Images of isolated islets collected from a female control mouse (D) and a female 6mGHRKO mouse (E). F: cross-sectional areas of islets from female 6mGHRKO mice and controls. Percent of total islets from male (G) and female (H) mice organized by size into three categories as indicated on the x-axis. n = 1,159 male control islets and n = 836 male 6mGHRKO islets. n = 480 female control islets and n = 351 female 6mGHRKO islets (***P < 0.001 by t test). Isolated islets were collected from mice at approximately 1 yr of age. GHRKO, growth hormone receptor knockout.

Insulin Secretion Is Elevated in Islets from Female, but Not Male, 6mGHRKO Mice

We next examined insulin secretion as a key measurement of islet function. As shown in Fig. 8A, tests of glucose-stimulated insulin secretion between 3 mM and 28 mM glucose among size-matched islets showed no differences between genotypes in islets from males. These data suggest that GHR deficiency induced in adulthood does not reduce insulin secretion in response to glucose in males. Among islets isolated from female 6mGHRKO mice, tests of glucose-stimulated insulin secretion in 28 mM glucose showed significantly increased insulin secretion compared with controls (Fig. 8B).

Figure 8.

Insulin release from islets is generally enhanced in female 6mGHRKO mice. A: insulin secretion for islets from male mice during 60-min incubation in 3 mM glucose (3 G) followed by 60-min incubation in 28 mM glucose (28 G). n = 8 sets of 20 islets/group for each test. No significant differences were observed. B: insulin secretion for islets from female mice during 60-min incubation in 3 mM glucose (3 G) followed by 60-min incubation in 28 mM glucose (28 G). n = 8 sets of 20 islets/group for each test. *P < 0.05. Islets were isolated from GHRKO mice at approximately 1 yr of age. GHRKO, growth hormone receptor knockout.

DISCUSSION

GH has long been known to play important functions in growth and development. With this study, we showed for the first time that the sex and developmental stage of the mice impacts the GH-mediated effects in β-cell mass and function. Our results suggest that GH action affects 1) islet growth as evidenced in germline GHRKO mice and also islet maintenance as indicated by the 6mGHRKO mice results and 2) the degree of islet growth and maintenance differently between male and female mice. Previous work has shown that germline GHRKO mice have a 68% reduction in islet size (34) as determined by histology, however, no sex differences were noted. We showed that this decreased islet size is more pronounced in male GHRKO and 6mGHRKO mice. Furthermore, effects of changes in GH signaling on insulin secretion in both mouse lines were also markedly sex dependent. The overall findings of the present study for GHRKO and 6mGHRKO mice are summarized in Fig. 9.

Figure 9.

Summary of the effects of growth hormone (GH) signaling disruptions on metabolism and islet function in male and female mice. Arrows indicate an increase (top) or decrease (bottom) relative to control mice. Dashed boxes emphasize limited to no changes in metabolic physiology (glucose tolerance and insulin sensitivity) due to opposing influences on islets (islet mass and insulin secretion). Thin-lined boxes indicate decreases in islet mass and/or fiunction, and associated thick-lined boxes emphasize an increase in insulin sensitivity to maintain glucose tolerance. # indicates reliance on published findings (54) that were not repeated in the present study. + indicates that although insulin secretion was reduced in male GHRKO mice, the glucose stimulation index was increased, suggesting increased islet sensitivity to hyperglycemia to help maintain normal glucose tolerance; islets from female GHRKO mice did not show this adaptation.

Reduced Islet Mass in Germline GHRKO and 6mGHRKO Mice

It is known that pancreatic β cells express GH and IGF-1 receptors (55). GHRKO mice have increased circulating GH and decreased IGF-1 and insulin levels (14, 15). Prior studies of pancreatic sections have shown that islets from GHRKO mice were ∼68% smaller than WT controls (34). Reduced islet mass has also been observed in β cell-specific GHR knockout (βGHR^−/−) mice (56) and adult-onset of GH-deficient (AOiGHD) mice (57). In both studies, diminished β-cell proliferation and mass expansion were found in response to nutrient challenge (higher caloric diet) when compared with WT mice; however, no change was seen in β-cell mass when βGHR^−/− mice were fed with normal chow diet (56) or when AOiGHD mice were fed with low-fat diet (57). Extending these findings, we observed a significant loss of islet mass from 6mGHRKO mice on regular chow diet and stark differences between sexes that were not reported previously. Cross-sectional area measurements of nearly 2,000 isolated islets showed that males had a greater reduction in islet size than females, with male and female GHRKO mice displaying ∼70% and ∼30% decreases in islet area, respectively, compared with controls.

Similar to GHRKO mice, the 6mGHRKO mice showed a substantial decrease in islet area accompanied by a significant increase in cell density within each islet. We hypothesized that the reason for this reduction in islet area could be one or more of the following: 1) decreased β-cell proliferation, 2) increased β-cell death, and/or 3) β-cell atrophy (reduced mass of individual islet cells). GH has both β-cell proliferation and antiapoptotic effects in insulin-producing cell lines (58), thus the loss of GH signaling could produce the opposite effect. Increased apoptosis or decreased proliferation should reduce the number of islet cells. Therefore, islets would be smaller because of fewer islet cells, which would not change the density of nuclei in the islet. Yet, we observed an increase in nuclei per unit area. The third possibility, β-cell atrophy, is consistent with our data. It is known that β cells also can undergo hypertrophy based on metabolic demands. In fact, Cho et al. (59) observed a 30% increase in β-cell size in patients with type 2 diabetes based on an assessment of the total β-cell area in islets and an observed reduction in β cell number. We observed this effect in reverse; reduced GH signaling leading to islet cell atrophy, rather than nutrient overload leading to hypertrophy. An alternative explanation for the results could be that islets from 6mGHRKO mice may have stopped growing at the point of tamoxifen injection while peanut oil-injected controls continued to grow. This change in somatic mass is not large enough to account entirely for the observed changes in islet mass, so changes in apoptosis and/or proliferation also likely contribute.

As mentioned in the materials and methods, it is important to note that tamoxifen control mice were not used. There is a remote possibility that tamoxifen as used in these studies could have an effect of islet cell mass and physiology. Because our assessments were performed 6 mo after the last tamoxifen injection, by this time acute tamoxifen effects are expected to be eliminated from the system. In terms of long-terms effects, it has been shown that histological changes in tissues are reversed by 28 days, even when using tamoxifen concentration that are higher than the dose used in this experiment (45). Furthermore, islet cell mass findings follow a similar trend in GHRKO (no-tamoxifen treated) and 6mGHRKO mice. Thus, our findings are not likely due to a tamoxifen effect.

GH and the Glucose-Stimulated Insulin Secretion Pathway

Our findings suggest that decreased GH action has different impact on glucose-stimulated insulin secretion at particular developmental stages. That is, germline GHRKO mice show a decreased insulin output for both sexes. In contrast, reduced GHR signaling in adulthood does not reduce insulin secretion and even increases secretion from islets as seen in female 6mGHRKO mice.

We found that islets from both male and female GHRKO mice secrete approximately half the insulin in response to glucose stimulation as size-matched controls (although this was only significant in males). These findings are consistent with defects in insulin secretion observed by Wu et al. (56) in islets from βGHRKO mice. Thus, it seems that germline disruption of GHR in the β cells leads to impaired insulin secretion and/or insulin production per β cell. Furthermore, male and female GHRKO mice are highly insulin sensitive. Therefore, it is possible that the reduced insulin secretion is due to a reduced demand since less insulin is needed to respond to glucose stimulation. This idea is supported by the glucose stimulation index data for male GHRKO mice, which indicates greater sensitivity to response to hyperglycemic conditions due to low basal insulin release. It is particularly intriguing that the failure of β cells to secrete adequate insulin is a major cause of both autoimmune type 1 diabetes and obesity-associated type 2 diabetes, yet very low levels of insulin in patients with Laron Syndrome have been suggested to contribute to reduced rates of diabetes and cancer (60). Unraveling this paradox may provide a better understanding of metabolic disorders.

In the case of 6mGHRKO mice, disruptions in GH signaling in adulthood had a different effect on insulin secretion. Male 6mGHRKO mice showed no differences in glucose-stimulated insulin secretion compared with controls, and female 6mGHRKO mice showed enhanced insulin secretion. These data indicate that islets from 6mGHRKO do not lose insulin-secreting capacity as islets from GHRKO mice do. In males, increased insulin sensitivity compensates for substantially reduced islet mass as insulin secretion remains unchanged. In females, insulin sensitivity is unchanged, islets appear to augment their insulin release to compensate for a slight loss of islet mass. Thus, different mechanisms appear to compensate for loss of GH signaling in adulthood between the sexes to maintain glucose homeostasis.

Comparing and Contrasting Mouse Models of GH Action

It is interesting that our results in 6mGHRKO mice differ from results found in the other mouse lines with postnatal reduction in GH action. Male adult-onset of GH-deficient (AOiGHD) mice fed with both low- and high-fat diet have decreased circulating insulin levels and reduced insulin mRNA levels in the pancreas when compared with controls (61). Furthermore, glucose-stimulated insulin secretion was impaired in older (46 wk old), but not in younger (22 wk old) male AOiGHD mice (61). The contradictory results obtained in the GSIS of male AOiGHD and 6mGHRKO mice could be derived from differences in the generation of the mice. The AOiGHD were generated by killing the somatotroph cells of the anterior pituitary at 10–12 wk of age with diphtheria toxin (57, 61), whereas the 6mGHRKO mice were produced by knocking out the Ghr gene globally in mice of 6 mo of age (13).

The 6mGHRKO mice also have a significant reduction in Ghr gene expression in all tissues; therefore, GH action is reduced throughout the body after 6 mo of age (53). In contrast, GHR is fully active in all the tissues in the AOiGHD mice, and circulatory levels of GH are decreased only ∼50% (57). Therefore, endocrine GH can still have an effect in all tissues of the mice. Also, the age at which the GH action was impaired could also affect islet mass and function; recall that the GH gene was disrupted at 10–12 wk of age (during or just after puberty in mice), whereas the Ghr gene was ablated during adulthood at 6 mo of age (53, 57).

A key factor in these different phenotypes is IGF-1, a hormone that is secreted primarily by the liver in response to GH-induced signaling. As the name implies, this peptide can bind to insulin receptors and have insulin-like effects, therefore, differences in circulatory IGF-1 could affect glucose homeostasis and insulin secretion. Although AOiGHD mice have normal to a modest reduction in IGF-1, the GHRKO and 6mGHRKO have more than a 90% decrease in circulating IGF-1 when compared with controls (53, 57). Hence, it is possible that circulatory IGF-1 in AOiGHD mice assist to supply the demand for insulin signaling to maintain glucose homeostasis. Also, mice with a β cell-specific deletion in the Igf1r gene (βIgf1r^−/− mice) have reduced glucose tolerance, reduced insulin secretion, and reduced expression of genes related to glucose handling such as the glucose transporter 2 and glucokinase (62). Hence, because of the diminished serum IGF-1 levels found in GHRKO and 6mGHRKO mice, it is likely that either enhanced insulin sensitivity and/or increased insulin secretion is needed to maintain glucose homeostasis. Although male 6mGHRKO mice showed increased insulin sensitivity, female 6mGHRKO mice demonstrated increased insulin secretion compared with controls. This provides a sex dimorphic means to the same end.

Furthermore, prolactin, a hormone intimately related with GH, also impacts glucose homeostasis. Prolactin and GH are closely related and belong to the Class I cytokine family. Both are secreted by the anterior pituitary (63) but by different cell types. Although it has been shown that prolactin increases β-cell proliferation and insulin release, experiments performed in rats and humans revealed that prolactin excess leads to decreased glucose tolerance and induces insulin resistance (64–67). Although, the prolactin effect on β cells and glucose metabolism is of importance because human GH has been shown to bind to both growth hormone and prolactin receptors, thus inducing prolactin intracellular signaling, mouse GH does not (68). Therefore, the high GH circulating levels seen in GHRKO and 6mGHRKO mice should not affect the prolactin activity in these mice.

The Potential Role of Estrogens and Androgens

Pancreatic β cells express both estrogen receptors (69) and androgen receptors (70), and GH has well-established interactions with these sex steroids (37, 71, 72). It is thus reasonable to speculate that differences in estradiol and/or testosterone signaling might play an important role in the sex differences in islet mass and insulin secretion observed in GHRKO mice. Activation of estrogen receptors, for example, has been shown to protect β-cell mass in rodent models of type 1 and type 2 diabetes (73). Also, the development of human embryonic stem cells (hESCs) directed to become pancreatic β cells is more successful depending on the sex of the host, with females showing an accelerated maturation of the hESCs implanted compared with males (74).

This is consistent with our observation that reduction of the islet mass was not as pronounced in females compared with males of both mouse lines. Estrogen receptor activation can also enhance glucose-stimulated insulin secretion, as reviewed in (73), and several human studies have shown that females secrete more insulin and/or C-peptide in response to a given glucose load compared with males (75). These results are congruent with the increased GSIS seen in female 6mGHRKO mice.

Testosterone may also play an important role in pancreatic islet GH signaling. Reports have noted a relationship between hyperglycemia and β-cell dysfunction with androgen deprivation therapy, a standard treatment for prostate cancer (76–78). Both in vivo and at the cellular level, androgen receptor activation enhances glucose-stimulated insulin secretion (79). Interestingly, male GHRKO and 6mGHRKO mice showed a reduction or no change in glucose-stimulated insulin secretion. In terms of islet cell mass, β cell-specific androgen receptor knockout (BARKO) mice and castration-induced testosterone deficiency in rats result in decreased β-cell mass compared with controls (80). These results are congruent with those obtained in both of the GHR-deficient mouse lines (79). Taken together, it appears that some of the observed effects in GHRKO and 6mGHRKO mice could be attributed to changes in sex steroid signaling. The extent to which estrogen and/or androgen signaling is involved with GH in pancreatic islets will require further study.

Conclusions

As summarized in Fig. 9, we have shown that germline GHRKO mice have reduced islet mass and reduced ability to secrete insulin, however, these mice are able to maintain normal glucose metabolism through enhanced insulin sensitivity. We have also shown for the first time that sustained GHR signaling throughout the body at an adult age is important for the maintenance of islet mass. Furthermore, results obtained for glucose-stimulated insulin secretion from islets demonstrate a sexual dimorphism in response to GHR ablation at 6 mo of age. Perturbations in GH signaling appear to have much greater impact on male mice with regard to islet mass and insulin sensitivity compared with female mice. Studies to clarify the molecular mechanisms that mediate the sexual dimorphism of β cells in response to glucose stimulation will be required to understand the consequences of reduced GH action in β-cell maintenance and glucose regulation.

GRANTS

Financial support for this work was provided by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Mouse Metabolic Phenotyping Centers (National MMPC, RRID:SCR_008997, www.mmpc.org) under the MICROMouse Program, Grant DK076169, by the NIH R01 AG059779 (to J.J.K.), and by funding from the Osteopathic Heritage Foundation (OHF) and Ohio University Heritage College of Osteopathic Medicine (OU-HCOM) (to C.S.N.). J.J.K. is supported, in part, by the State of Ohio’s Eminent Scholar’s Program, which includes a gift by Milton and Lawrence Goll.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

S.D.-O., E.O.L., J.J.K., and C.S.N. conceived and designed research; S.D.-O., K.L.C., I.J., N.B.W., S.E.M., K.G.S., I.M.M.-H., E.O.L., and C.S.N. performed experiments; S.D.-O., K.L.C., I.J., N.B.W., S.E.M., A.N.B., K.G.S., H.L.W., I.M.M.-H., E.O.L., J.J.K., and C.S.N. analyzed data; K.L.C., A.N.B., K.G.S., H.L.W., I.M.M.-H., E.O.L., J.J.K., and C.S.N. interpreted results of experiments; K.G.S., H.L.W., I.M.M.-H., J.J.K., and C.S.N. prepared figures; S.D.-O., K.L.C., K.G.S., J.J.K., and C.S.N. drafted manuscript; S.D.-O., K.G.S., I.M.M.-H., E.O.L., J.J.K., and C.S.N. edited and revised manuscript; S.D.-O., K.L.C., I.J., A.N.B., K.G.S., I.M.M.-H., E.O.L., J.J.K., and C.S.N. approved final version of manuscript.