High-fat diet-induced β-cell proliferation occurs prior to insulin resistance in C57Bl/6J male mice

Experimental animals.

With the exception of hyperinsulinemic euglycemic clamp studies, all experiments were performed in the Vanderbilt University facility. Male C57Bl/6J mice (Jackson Laboratory, Bar Harbor, ME) were delivered to the Association for Assessment and Accreditation of Laboratory Animal Care International-accredited Division of Animal Care at Vanderbilt University at 7 wk of age. After 1 wk of acclimatization, mice were weighed and randomly assigned to one of two groups: 1) CD (11% kcal from fat, Lab Diet 5LJ5; Purina, St. Louis, MO) and 2) HFD (60% kcal from fat; BioServ F3282, Frenchtown, NJ). Mice were housed on a 12:12-h light-dark cycle, receiving water and food ad libitum. Body weight was obtained prior to starting the diets and prior to each metabolic measurement. Euthanasia was performed at time of pancreatic dissection, using isoflurane until the mice were unconscious, followed by cervical dislocation. These methods are consistent with the Panel on Euthanasia of the American Veterinary Medical Association. All procedures were conducted in accordance with Vanderbilt University Division of Animal Care and Use Committee-approved protocols and under the supervision of the Division of Animal Care. For hyperinsulinemic euglycemic clamps, male C57Bl/6J mice (Jackson Laboratory, Bar Harbor, ME) were delivered to the Animal Facility of the Research Center of the University of Montreal Hospital Cente (CRCHUM) at 7 wk of age and the clamps performed at 8 wk of age. All procedures were approved by the Institutional Committee for the Protection of Animals at the CRCHUM.

Metabolic measurements.

Intraperitoneal glucose tolerance tests (IPGTT) and IPITT were performed as described previously (18). For IPGTT, animals were fasted for 16 h. Fasting blood glucose was measured from 2 μl of tail vein blood with an Accuchek glucometer and glucose test strips (Abbott Diabetes Care). Animals received an intraperitoneal (ip) injection of filter-sterilized glucose (2 mg dextrose/g body wt), and blood glucose was measured at 15-, 30-, 60-, 90-, and 120-min intervals following injection; n = 3 (weeks 3, 7, and 9) or 6 (weeks 1, 5, and 11). For IPITT, animals were fasted for 6 h. Fasting blood glucose was measured as described above, and animals received an ip injection of 0.075 U/ml insulin (recombinant human insulin, no. I9278; Sigma-Aldrich) in filter-sterilized 1× PBS at 0.1 ml/10 g body wt. Subsequent changes in blood glucose were measured at 15-, 30-, 60-, 90-, and 120-min intervals following injection; n = 3 (weeks 3, 7, and 9) or 6 (weeks 1, 5, and 11). Plasma insulin assays were performed as described in (31), with the modification that blood was harvested from the saphenous veins of 16-h-fasted animals using heparinized Natelson blood-collecting tubes (Kimble Chase, Rockwood, TN).

Hyperinsulinemic euglycemic clamps.

Two-hour hyperinsulinemic-euglycemic clamps were performed in 4-h-fasted mice (n = 6 for CD and 7 for HFD), as described previously (4). Briefly, following a 1-min bolus insulin infusion (85 mU/kg; Humulin R), insulin was infused at 8 mU·kg⁻¹·min⁻¹. Twenty percent dextrose was infused beginning 5 min after the insulin infusion to clamp glycemia at ∼120 mg/dl. Insulin levels during the steady state were measured at 90 and 120 min using the AlphaLISA kit. The insulin sensitivity index (M/I) was calculated as the glucose infusion rate (GIR) divided by the average insulinemia during the last 30 min of the clamp (I); n = 6 CD and 7 HFD.

Tissue preparation and histology.

At euthanization, pancreata were processed as described previously (14). Antibodies were guinea pig anti-insulin (1:500; Dako, Carpinteria, CA), rabbit anti-Ki67 (1:500; AbCam, Cambridge, MA), Cy2-conjugated anti-guinea pig IgG (1:300; Jackson Laboratories, Bar Harbor, ME), Cy3-conjugated anti-rabbit IgG (1:300, Jackson Laboratories), and horseradish peroxidase-conjugated anti-guinea pig IgG (1:300, Jackson Laboratories).

β-Cell mass, β-cell proliferation, and β-cell death.

Analysis and quantification of β-cell mass was performed as described in (18). For β-cell mass assessment, ∼2% of each pancreas was immunolabeled and analyzed (5–10 sections/animal, each separated by 250 μm). Slides were scanned at ×20 magnification (Scan Bright field Scope System; Aperio, Vista, CA), and an algorithm developed from a Genie macro within Spectrum (Aperio) was used to identify β-cells and other tissue (15); n = 3 (weeks 1, 3, 7, and 9 for CD and HFD), 5 (week 11 for CD), or 6 (week 5 for CD and HFD and week 11 for HFD). β-Cell proliferation was determined by immunolabeling sections ∼400 μm apart (∼5 slides/animal) for insulin and Ki-67 (Abcam; 1:500) or insulin and phosphorylated histone H3 (pHH3, 1:200; Cell Signaling Technology). For Ki-67 labeling, antigen retrieval consisted of microwaving slides for 14 min in 10 mM sodium citrate buffer; n = 3 (weeks 3, 7, and 9 for CD and HFD), 4 (3 days for CD), 5 (3 days for HFD), or 6 (weeks 1, 5, and 11 for CD and HFD). For pHH3 labeling, antigen retrieval was placed in TEG buffer at pH 9.0 and microwaved on high power for 1 min and then 10% power for 7.5 min. Nuclei were labeled with 1 μg/ml 4′,6′-diamidino-2-phenylindole (DAPI, Molecular Probes, Grand Island, NY) and mounted with Aqua-Mount (Thermo Scientific, Kalamazoo, MI); n = 3 for all time points and diets. Slides were scanned as above. At least 5,000 insulin-positive cells/mouse were counted using MetaMorph software. Calculations were made by dividing the number of insulin/Ki-67 or insulin/pHH3 colabeled cells by the total number of insulin-positive cells. To assess β-cell death, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) was performed using the ApoAlert Kit (Clontech) according to the manufacturer's instructions (16). For TUNEL assay, pancreata (3 sections/animal) from three CD- and three HFD-fed animals were analyzed at each of three time points (3 days, 1 wk, and 11 wk).

Quantitative RT-PCR.

Islet RNA from 1-wk-treated mice was isolated, and quantitative RT-PCR (qRT-PCR) was performed as described previously (1). Primer sequences are listed in Table 1. Data are shown as 2^−ΔΔC_T (24); n = 6.

Metabolomic analysis.

Whole liver, epididymal fat, skeletal muscle (gastrocnemius, soleus, and plantaris muscles), bone (fibula and tibia), and plasma were collected from 9-wk-old C57Bl/6J mice either fed a HFD (n = 8) or maintained on a CD (n = 8) for 1 wk. Tissues were dissected, flash-frozen, and kept at −80°C before being shipped to Metabolon (Durham, NC) for metabolite analysis. Sample preparation, instrument analysis, and data processing analysis were performed by Metabolon, as detailed in previous publications (9, 27).

Statistical analysis and calculations.

Data are shown as means ± SE (12). P values were calculated with either the two-tailed unpaired Student's t-test or the two-way ANOVA with the Sidek correction for multiple comparisons as indicated. P values ≤0.05 were considered significant, and P values >0.05 were not reported. For metabolomic data shown in Fig. 5B, any missing values were assumed to be below the detection limit and were imputed with the minimum observed value. Following imputation and log transformation, P values were calculated using the Welsh two-sample t-test, and estimates of the false discovery rate (q value) were calculated to account for multiple comparisons, as described previously (36).

Weight gain and impaired glucose tolerance in HFD-fed mice.

To correlate the consumption of HFD with changes in body weight, glucose homeostasis, and β-cell growth, 8-wk-old male mice were either maintained on CD or switched to HFD. The average initial body weight for the CD group was identical to the HFD group (20.4 ± 2.1 vs. 20.5 ± 1.5 g, respectively). Mice from each diet were examined at 1, 3, 5, 7, 9, and 11 wk after the start of the study for body weight, glucose homeostasis, β-cell mass, and β-cell proliferation. HFD-fed mice weighed significantly more than CD-fed mice throughout the study, starting with an initial weight gain of 1.1 ± 0.9 g (P = 0.006) in the 1st wk compared with the static weight of the CD group (−0.7 ± 1.5 g, P = not significant) (Fig. 1A). The largest weight increase for both diets occurred between weeks 3 and 5, where CD-fed mice gained 3.9 ± 1.0 g (P = 0.001) and HFD-fed mice gained a similar 3.7 ± 1.1 g (P = 0.008). After week 5, the HFD group continued to gain weight, showing an 8.4 ± 1.5-g (P = 0.0001) increase over the next 6 wk, whereas body weights of CD-fed mice plateaued during this time frame with an insignificant increase in body weight of 1.6 ± 0.8 g.

Changes in glucose tolerance in response to both acute and long-term HFD consumption were tracked by performing an IPGTT every other week. As expected from previously published studies, glucose tolerance was already impaired after 1 wk of HFD and was consistently impaired throughout the study despite unchanged fasting blood glucose (Fig. 1, B–F). To depict subtle but significant changes in glucose tolerance as mice were maintained on HFD, we calculated the area under the curve for glucose (AUC_glc) both between diets and between intradiet time points (Fig. 2, A–C). The AUC_glc for HFD-fed mice increased gradually between weeks 1 and 9 and culminated with a highly significant increment in AUC_glc from weeks 9 to 11 (P = 0.001); the AUC_glc for CD-fed mice was relatively unchanged (Fig. 2A). The gradual increase in HFD-fed AUC_glc between weeks 1 and 9 was due to an elevation in glucose during the first 30 min of the IPGTTs (Fig. 2B). In fact, the last 90 min of the IPGTTs remained relatively consistent between intradiet time points, with the exception of week 11 in the HFD-fed group, when a large increase in AUC_glc occurs (Fig. 2C).

Insulin resistance is not apparent with short term HFD feeding, but impaired insulin tolerance develops with prolonged HFD feeding.

The progression of impaired glucose tolerance in response to both acute and long-term HFD suggests impairments in insulin secretion and/or insulin sensitivity. First, we determined the effects of short- and long-term HFD feeding on insulin tolerance. Since injection of insulin can dramatically lower blood glucose levels, mice were fasted for only 6 h prior to insulin injection. Fasting blood glucose concentrations were not altered in response to short-term HFD feeding but were elevated with long-term HFD feeding (Fig. 3C). Interestingly, mice fed HFD for 1 wk exhibited a significantly enhanced decline in blood glucose during the latter time points of the IPITT (Fig. 3A), and this was still apparent after normalization to fasting blood glucose (Fig. 3D). In week 5, the blood glucose concentrations of the HFD-fed group were indistinguishable from controls (Fig. 3, B and E). After 11 wk of HFD feeding, we observed elevated blood glucose throughout the test relative to CD-fed mice (Fig. 3C). These data show that insulin tolerance in HFD-fed mice gradually diminished with duration of HFD, whereas the CD-fed mice maintained a consistent response throughout the study. However, after normalization to basal glucose concentrations, insulin tolerance in the long-term HFD-fed mice was only slightly compromised at the later time points (Fig. 3F). Similar to week 5, there was no significant difference between the diets at weeks 3, 7, and 9 (data not shown).

Plasma was assayed every other week for circulating insulin in response to glucose. Although there was a significant increase in plasma insulin in response to the glucose challenge in animals fed HFD for 1 wk, this response was reduced by 50% compared with the response of the CD-fed group (Fig. 4A). The CD-fed plasma insulin response remained consistent through week 11 (Fig. 4, B and C). Plasma insulin values for HFD-fed mice were virtually unchanged for weeks 1–7 (Fig. 4, A and B, and data not shown). As predicted, hyperinsulinemia was apparent in long-term HFD-fed mice (Fig. 4C). Fasting plasma insulin was elevated in HFD-fed mice by week 9 (data not shown) and exacerbated further in week 11 (Fig. 4C), at which point it was nearly fourfold higher in the HFD-fed group compared with the CD-fed group (P = 0.02). Despite elevated fasting plasma insulin in HFD-fed mice at week 11, they still presented with a moderate but significant response to the initial glucose challenge at 15 min (P = 0.03).

Our data at later time points agree with a wealth of previously published studies demonstrating insulin resistance after prolonged HFD feeding. The more provocative finding of the present study is a lack of apparent insulin resistance 1 wk after initiation of HFD. To more rigorously investigate whether insulin resistance is present after 1 wk of HFD feeding, we performed hyperinsulinemic euglycemic clamps (Fig. 5). The GIR required to maintain blood glucose levels (Fig. 5A) at the target level was not different between the two groups (Fig. 5B). Although circulating insulin levels at the end of the clamp were slightly higher in the HFD-fed group (Fig. 5C), this difference was not statistically significant, and the M/I index was not different between both groups (Fig. 5D). These results suggest that insulin sensitivity was unaltered after 1 wk of HFD.

In humans, elevations of the metabolite α-hydroxybutyrate (α-HB) in plasma are indicative of insulin resistance (11, 13). Given the results from the IPITT and hyperinsulinemic euglycemic clamps suggesting lack of insulin resistance at 1 wk of HFD feeding, we queried results from metabolomic analyses of metabolically relevant tissues (including plasma) to determine whether levels of α-HB were increased in response to HFD at this time point. Metabolomic analysis of liver, skeletal muscle, epididymal fat, bone, and plasma collected from 1-wk-HFD- and -CD-fed mice revealed decreased α-HB in the liver (0.62-fold lower, P = 0.002, q = 0.012) and adipose tissue (0.73-fold lower, P = 0.011, q = 0.002) in HFD-fed mice compared with CD-fed mice; α-HB was unchanged in other tissues, including plasma, at this time point (Fig. 6).

β-Cell mass expansion begins after 3 wk of HFD consumption.

β-Cell mass expansion is well established in long-term HFD feeding studies; however, the timing of this expansion is less clear. To address this, β-cell mass was determined every other week between 1 and 11 wk (Fig. 7). A significant increase in β-cell mass was observed after 3 wk of HFD feeding; however, this increase was not sustained throughout the treatment period. The β-cell mass within each diet group remained relatively constant until week 9, when a significant increase was observed in both groups. Compared with the chow-fed animals, a further increase in β-cell mass was observed between weeks 9 and 11 in the HFD-fed group. Comparison between the two diets at the 11-wk time point revealed a 1.3-fold increase in β-cell mass for HFD-fed mice over CD-fed mice (P = 0.001).

β-Cell proliferation was induced with short-term HFD feeding.

To ascertain the kinetics of the β-cell-proliferative response to HFD feeding, pancreata were assayed at different durations of HFD for β-cell proliferation using Ki-67 and pHH3 immunolabeling. Whereas CD-fed mice exhibited a constant low percentage of Ki-67/insulin double-positive cells throughout the study, β-cell proliferation in the HFD-fed group was dynamic and enhanced after just 3 days (Fig. 8A). Compared with CD, β-cell proliferation was enhanced 2.6-, 2.2-, 2.9-, and 2.3-fold for the HFD-fed mice at 3 days (P = 0.05), 1 wk (P = 0.00005), 5 wk (P = 0.0001), and 7 wk (P = 0.003), respectively. Similar results were obtained using pHH3 as a proliferative marker; the greatest enhancement in β-cell proliferation was observed after only 3 days of HFD, and an increase in β-cell proliferation was maintained at 1 wk of HFD but was no longer apparent at 11 wk of HFD (Fig. 8B).

Since a significant increase in β-cell proliferation was detected within 1 wk of HFD treatment, qRT-PCR was conducted on RNA from isolated islets at 1 wk to analyze changes in gene expression for key proliferative genes. Expression of Ki67, Foxm1, cyclin A2, and cyclin B1 was upregulated 1.6-, 1.6-, 1.7-, and twofold (P = 0.003, 0.00004, 0.03, and 0.01), respectively, in 1-wk-HFD-fed mouse islets compared with CD islets. Expression of cyclins D1 and D2 was unchanged. Expression of the cell cycle inhibitors cdkn2A (p16), cdkn1A (p21), and cdkn1B (p27) was not changed significantly in HFD-fed mice (Fig. 8C).