Short-term high-fat consumption stimulates mouse islet β-cell replication through unknown mechanisms. Resident macrophages (MΦs) are capable of secreting various factors involved in islet development and tissue remodeling. We hypothesized that a short-term high-fat diet (HFD) promotes MΦ infiltration in pancreatic islets and that MΦs serve as a regulator of β-cell replication. To test these hypotheses and dissect mechanisms involved in HFD-induced β-cell replication, adult C57BL/6J mice were fed a HFD for 7 days with or without administration of clodronate-containing liposomes, an MΦ-depleting agent. Mouse body and epididymal fat pad weights, and nonfasting blood glucose and fasting serum insulin levels were measured, and pancreatic islet β-cell replication, oxidative stress, and MΦ infiltration were examined. Short-term HFD promoted an increase in body and epididymal fat pad weight and blood glucose levels, along with an increased fasting serum insulin concentration. β-Cell replication, islet MΦ infiltration, and the percentage of inducible NO synthase positive MΦs in the islets increased significantly in mice fed the HFD. Immunofluorescence staining for 8-oxo-2′-deoxyguanosine or activated caspase-3 revealed no significant induction of DNA damage or apoptosis, respectively. In addition, no change in stromal-derived factor 1-expressing cells was found induced by HFD. Despite continuous elevation of nonfasting blood glucose and fasting serum insulin levels, depletion of MΦs through treatments of clodronate abrogated HFD-induced β-cell replication. These findings demonstrated that HFD-induced MΦ infiltration is responsible for β-cell replication. This study suggests the existence of MΦ-mediated mechanisms in β-cell replication that are independent of insulin resistance.
- β-cell replication
islet β-cell mass adaptation to obesity has been observed in both rodents (36) and humans (29), with human autopsies showing increases of up to 50% in obese patients (29). The growth of β-cells under high-fat diet (HFD) conditions represents a natural adaptation process of the cells and provides an excellent in vivo model for investigating the potential of and the mechanisms involved in β-cell regeneration. The compensatory β-cell proliferation following long-term HFD is believed to be secondary to the development of insulin resistance (30) in response to liver-derived systemic factors (12). It has been revealed recently, however, that increases in β-cell proliferation are initiated within days after the start of high-fat feeding in mice, with (32) or without (23) the deveolopment of insulin resistance, suggesting potential alternate mechanisms involved. The goal of this investigation was to identify early changes in the islet microenvironment in response to HFD consumption that might contribute to early β-cell proliferation.
Although it has long been known that long-term HFD induces peripheral tissue inflammation, especially in adipose tissue, driven by innate immune responses to lipotoxicity, the potentially pathogenic changes occurring in the islets during early stages of high-fat feeding are unclear (24, 28, 35, 39). There have been various reports describing increased inflammation in pancreas and islets of humans with type 2 diabetes (T2D) (11) and in a nonobese rodent model of T2D (17). Nondiabetic juvenile nonhuman primates (NHPs) in the early stages of insulin resistance after chronic (>12 mo) HFD also showed the presence of increased numbers of islet-associated macrophages (MΦs) along with increased cytokine expression in the pancreatic tissue (25). All these findings suggest a role for local inflammation in the impairment of islet function in metabolic diseases. On the other hand, MΦs are normal constituents of fetal, neonatal, and adult pancreata in humans and rodents, capable of secreting various factors involved in islet development and tissue remodeling (1, 9). Moreover, they are found to be important in promoting adult β-cell regeneration in mouse models of pancreatic tissue injury (2, 6, 38). As an initial step in testing the hypothesis that islet-infiltrating MΦs serve as a potential signaling source for rapid, HFD-induced β-cell replication, this work paid particular attention to MΦ infiltration in pancreatic islets and β-cell growth after a 7-day HFD treatment. In this paper, we characterize oxidative stress level and MΦ infiltration, as well as depletion of MΦ in relation to β-cell replication in pancreatic islets of short-term HFD-treated C57BL/6 mice.
MATERIALS AND METHODS
Animals and treatments.
Ten-week-old male C57BL/6J mice (Jackson Laboratory) were housed in a temperature-controlled room with a 12:12-h light-dark cycle under specific pathogen-free conditions in the animal barrier facility at the Columbia University Medical Center. For the experiment, mice were fed either a normal chow (Teklad Global Rodent T.2018.R12, Harlan Laboratories) containing 5.5% fat or a HFD (Teklad Adjusted Calories Diet TD88137, Harlan Laboratories) with 21% total fat. Clodronate-liposomes that induce apoptosis of MΦs were injected (100 mg/kg in 100 μl iv) into a subgroup of mice 1 day before starting HFD followed by a second injection 3 days later. Another subgroup of HFD-fed mice received daily oral gavage containing an insulin sensitizer, pioglitazone (25 μg/g body wt po, Selleck Chem). There were at least five animals in each experimental group. Nonfasting blood glucose levels were measured daily using a glucometer (ACCU-CHEK). All mice had free access to water containing a DNA precursor analog 5-bromo-2′-deoxyuridine (BrdU, 0.8 mg/ml; BD Pharmingen) during the 7-day experiment. Mice were euthanized at the end of the experiment, at which time the epididymal fat pads were harvested for weight determination, pancreatic tissues were harvested for immunofluorescence staining, and fasting serum samples were obtained for insulin measurement. All animal procedures were approved by the Institutional Animal Care and Use Committee of Columbia University.
Immunofluorescence staining of pancreatic tissue sections.
Upon harvest, the whole mouse pancreas tissue was immediately fixed in 4% paraformaldehyde for 30 min, cryopreserved in 30% sucrose overnight, embedded in OCT, and stored at −80°C. The entire pancreatic tissue was sectioned (5-μm slices) and used for the various immunofluorescence stainings described below. In order to avoid repeated staining of the same islets for the same types of antibody combinations for statistical consideration, pancreas sections selected for analyses were at least 100 μm apart. They were stained with primary antibodies specific for insulin (Dako), the MΦ marker F4/80 (Invitrogen), BrdU (BD Pharmingin), inducible nitric oxide synthase (iNOS, Abcam), stromal cell-derived factor 1 (SDF1, Abcam), and/or 8-oxo-2′-deoxyguanosine (8-OHdG, Santa Cruz Biotechnology,). In addition, pancreatic tissue sections were also stained for glucagon (Abcam), monocyte chemoattactin protein-1 (MCP1, Abcam), or cleaved caspase-3 (Cell Signaling). Secondary antibodies conjugated with various fluorochromes (ImmunoLab) were used to label cells bound with the respective primary antibodies. Images were recorded using LAS AF 6.2 software on a motorized Leica DMI 600B fluorescence microscope (Leica Microsystems, Wetzlar, Germany).
Insulin-, BrdU-, F4/80-, iNOS-, SDF1-, or 8-OHdG-positive cells were counted manually with the assistance of the ImageJ program. BrdU-positive β-cells, iNOS, and F4/80 staining were analyzed from four distinct pancreas samples with more than 150 islets and 4,000–8,000 total insulin-positive cells from each experimental group. The percentages were determined as the total BrdU- or 8-OHdG-positive nuclei of β-cells divided by the total nuclei of β-cells and multiplied by 100. Islet-associated MΦs were identified as a F4/80-positive cell within a cluster of insulin-positive cells and presented as a ratio of MΦs to β-cells with more than 1,000 insulin-positive cells counted for each sample. The percentages of iNOS-positive MΦs were calculated as F4/80 and iNOS double positive cells divided by total MΦs in the islets and multiplied by 100. The percentages of SDF1-positive cells were calculated as SDF1-positive cells divided by total insulin-positive cells in over 100 islets (totaling more than 1,500 β-cells) for each the control and HFD group multiplied by 100.
Fasting mouse serum insulin detection by ELISA.
At the end of the 7-day experiment, all mice were deprived of food for a 6-h period while water was continuously provided. After the fasting period, blood samples were collected and serum was prepared at the time animals were euthanized. Serum insulin levels were measured using a mouse insulin ELISA kit (Mercodia) according to the manufacturer's instructions. At least four animals per experimental condition were examined.
For multiple-group comparisons, statistical differences were determined by one-way ANOVA with post hoc Bonferroni test (Prism-GraphPad). For two-group comparisons, statistical differences were determined by unpaired two-tailed Student's t-test. P < 0.05 indicates significance.
MΦ infiltration and β-cell replication occur early after HFD feeding.
Despite the extensive studies of the proinflammatory effects of long-term HFD on peripheral tissues driven by innate immune responses to lipotoxicity (24, 28, 35, 39), the potentially pathogenic changes occurring in the pancreatic tissue during early stages of high-fat feeding are unclear. By use of F4/80, a well-characterized and extensively referenced membrane protein on mature mouse macrophage as a marker, the number of islet-targeted MΦs was significantly increased (P < 0.05) by day 7 of HFD, with the percentage of macrophages in islet β-cells reaching 10.74 ± 0.95% compared with the control baseline level of 5.32 ± 0.94% (Fig. 1, A, B, E). At the same time, β-cell replication, as defined by BrdU incorporation in an insulin-positive cell, increased from 1.25 ± 0.35% to 7.10 ± 0.80% 7 days after the short-term HFD treatment (Fig. 2, A, B, E). The increased MΦ infiltration and cellular proliferation were islet localized and specifically β-cell targeted, respectively, as no significant increase in cellular infiltration or replication was observed in the pancreatic exocrine tissue or other islet endocrine cells. Glucagon- and BrdU-double-positive cells were not detected in pancreata of all groups (data not shown). No significant difference in the size of the β-cells was observed between different treatment groups.
MΦ depletion and insulin sensitivity manipulation.
In order to examine whether MΦ infiltration is linked to β-cell replication, we administered clodronate, a commonly used agent for MΦ depletion (3, 20), to a subgroup of HFD-fed mice. Clodronate is encapsulated in liposomes that are phagocytized by MΦs, eventually initiating programmed cell death. Administration of clodronate-liposomes significantly reduced the number of MΦs infiltrating the islets close to baseline level 5.67 ± 0.82% (Fig. 1, C and E) and at the same time abrogated the HFD-induced β-cell replication to 3.1 ± 0.4% (Fig. 2, C and E), both were similar to that of the untreated control group (p > 0.05). Fasting serum insulin levels, indicative of insulin sensitivity (19), in mice on normal chow, HFD, HFD with clodronate treatment, or HFD with pioglitazone treatment, were significantly increased from 0.15 ± 0.07 to 2.05 ± 0.75, 2.40 ± 0.85, or 0.84 ± 0.76 ng/ml, respectively (Fig. 3). These data suggest that insulin sensitivity, which is negatively correlated to the fasting state serum insulin level, was decreased after the 7-day HFD treatment, consistent with what others have reported (21, 27, 32). Depletion of MΦs by clodronate-liposomes did not prevent the short-term HFD-induced insulin resistance as the fasting serum insulin levels in both treatment groups were significantly higher than those in the control mice. This is consistent with the reports showing that short-term HFD-induced insulin resistance is independent of inflammation (21) and develops as a result of increased mitochondrial emission of reactive oxygen species (ROS) in the adipose tissue in the absence of MΦ infiltration (27). Treatment with pioglitazone reduced the fasting serum insulin level (Fig. 3), although no significant difference was detected compared with the levels of other groups. Pioglitazone treatment, however, reduced HFD-induced β-cell proliferation to 2.03 ± 0.18% (Fig. 2), which was not significantly different (P > 0.05) from that of the control animals or animals on HFD and clodronate but was significantly different than that of mice treated with HFD alone (P < 0.05). It is interesting to note that pioglitazone also inhibited HFD-induced islet MΦ infiltration to 3.60 ± 2.32% (Fig. 1, D and E), probably due to its direct inhibitory effect on macrophages (31, 41).
Short-term HFD promotes an increase in body and epididymal fat pad weight and blood glucose level.
To confirm that the metabolic status of mice in all the treatment groups were comparably affected by the HFD treatment, mouse body weight, fat pad weight, and blood glucose level were measured at the end of the treatment. The body weight of mice from all treatment groups ranged from 23 to 27 g at the start of the experiment. Mice on normal chow feeding showed an insignificant increase in body weight after the 7-day experiment. Feeding mice HFD for 7 days resulted in significant increases in their body and fat pad weights (Fig. 4, A and B) and in their nonfasting blood glucose levels (Fig. 4C), which is similar to what was reported by others (32). Mice on HFD that received clodronate-liposome or pioglitazone treatment did not show significant increase in body weight (Fig. 4A), presumably due to the increased stress level experienced by these mice as a result of the treatment procedures. The fat pad weight (Fig. 4B), however, was significantly increased in all mice fed HFD, serving as an internal control that all mice on HFD consumed a similar amount of the diet and had a similar amount of excess calories disproportionally stored in the adipose tissue. Depletion of MΦs by clodronate-liposome treatment did not prevent HFD-induced elevation of blood glucose level, whereas the insulin-sensitizer pioglitazone, as expected, reduced the increase to an insignificant level, as shown in Fig. 4C. Last, neither clodronate nor pioglitazone had an impact on HFD-induced fat pad weight gain.
The oxidative stress level and apoptotic cells in islets.
High-fat feeding may lead to lipotoxicity of tissues, including metabolic interference and stress responses such as oxidative stress and autophagy (15). To better understand the phenomenon of MΦ infiltration in islets and to examine whether MΦs might have migrated to islets in response to cellular injury, we evaluated the presence of 8-OHdG, a biomarker of oxidative stress, in islets of mice fed with normal chow or HFD. As shown in Fig. 5, approximately 2% of the islet cells or total pancreatic cells showed positive 8-OHdG staining in their nuclei, with no significant difference detected between the two experimental groups. In addition, examination of pancreatic tissue samples from all treatment groups for the presence of activated caspase-3, a marker of apoptosis, did not reveal any positive result (data not shown).
Pancreatic expression of SDF1.
SDF1 is a peptide chemokine expressed ubiquitously in stromal tissues including that of mouse pancreata (40). Its role in attracting MΦs to the site of islet injury and promoting regeneration has recently been suggested (5). To view whether the presence of SDF1 was altered in the pancreas of HFD-treated mice, we examined and compared the expression pattern of SDF1 in islets of mice fed normal chow or HFD. As shown in Fig. 6, SDF1 staining was detected in the periphery of islets in insulin- and glucagon- (not shown) negative cells. The percentages of SDF1-positive cells were calculated to be 22.24 ± 5.20 and 20.76 ± 4.78%, in the control and the HFD groups, respectively. No statistical differences in the expression level or the number of positive cells were detected between the two groups.
Islet MΦ subtype characterization.
MΦs demonstrate incredible plasticity in response to various environmental stimuli, which allows them to specialize and display polarized functional properties such as inflammatory or anti-inflammatory actions. As an initial step in characterizing the phenotype of MΦs in the pancreatic islets at the time of harvest, we checked for the expression of iNOS, the marker of M1 proinflammatory MΦ, in F4/80+ cells. As shown in Fig. 7, islets in normal control mice contained both iNOS− and iNOS+ MΦs. After the 7-day HFD, the number of infiltrating MΦs, especially the iNOS+ MΦs, increased, leading to a significantly higher percentage (89.89 ± 6.06%) of iNOS+ MΦs compared with 76.19 ± 8.61% of the controls (P < 0.05). The fluorescence of iNOS immunostaining was also detected in most of the β-cells, in the periphery of islets and in the inner lining of blood vessels. The intensity of the staining and the number of β-cells positive for iNOS were not significantly different between the control and the HFD groups (Fig. 8).
The aim of this study was to identify factors that contribute to early β-cell proliferation and to investigate the immediate changes occurring in the islet microenvironment in response to a HFD. This report shows, for the first time, that MΦs accumulate in islets in a matter of days after HFD feeding, accompanied by increased β-cell replication. Such a rapid accumulation of MΦs was not observed in adipose tissue of C57BL/6 mice after a 7-day HFD treatment (27), even though insulin resistance was developed in these mice as a result of increased mitochondrial emission of ROS in the adipose tissue.
This finding suggests that pancreatic islets are early targets of MΦ, although the specific subtypes of the cells and the signals attracting them to islets remain to be identified. Cellular damage and death oftentimes signal MΦ migration. We examined the presence of 8-OHdG, a marker of oxidative DNA damage induced by reactive oxidative stress in cells undergoing metabolic disturbance, as an indicator of cellular damage. No significant differences in the number of 8-OHdG-positive cells between mice on chow-fed diet vs. those on HFD were found. Consistently, no signs of apoptosis were detected by us (data not shown), or by others in the short-term HFD experiments (6), nor in the pancreata of mice or NHP juveniles exposed to HFD for 14 wk or >12 mo, respectively (25, 37). These results argue against a traditionally projected scavenger image of MΦs in islets of early-stage HFD consumption. We also found no change in the expression of SDF1, which was suggested to play a critical role in the recruitment of MΦs to the injured islets (5). Further characterization of the changes occurring in islets upon HFD treatment could potentially lead to better understanding of the signals that attract the MΦs into islets.
The increase in mouse body weight, adipose tissue mass, and nonfasting blood glucose level, as well as insulin resistance found at day 7 of HFD consumption are consistent with what others have reported (21, 26, 27, 34). Our results also show a correlation between the presence of infiltrating MΦs and the induced replication frequency of the islet β-cells under HFD regardless of the state of insulin resistance. However, depletion of MΦs by clodronate-liposomes did not attenuate the early-stage insulin resistance in our mice. Despite the presence of insulin resistance, HFD-induced β-cell replication was abolished in clodronate-treated mice, excluding the possibility that the increased replication frequency of β-cells in this study was a direct response to increased insulin resistance, although we cannot rule out the possibility that the clodronate-liposomes also have a direct inhibitory effect on β-cell replication in addition to their ability to eliminate MΦs. β-Cell replication independent of insulin resistance was also reported previously in mice that consumed a HFD for 3 days (23), although no mechanism was revealed. Our studies suggest a supporting or mediator role of infiltrating MΦs in islet β-cell replication. High-fat-fed mice that also received pioglitazone treatment showed improvement in their insulin sensitivity and did not respond to HFD with β-cell replication. The latter seemed to indicate that the early increase in β-cell replication was associated with the development of insulin resistance but could also suggest that the replication was associated with MΦ infiltration, since pioglitazone also has an inhibitory effect on MΦ (31, 41), consistent with the observed absence of islet-infiltrating MΦs in mice treated with pioglitazone. It will be of interest to confirm this finding of the effect of pioglitazone on β-cell replication in other islet regeneration models, such as injury, when MΦs are involved (2, 6, 33, 38).
Activation of the innate immune system is a hallmark feature of obesity and diabetes. In many strains of rodents, HFDs induce an inflammatory response in the liver, adipose tissue, and hypothalamus (4, 7, 18). The first indication that islet MΦs might contribute to islet inflammation in diabetes was identified by histological studies documenting the accumulation of CD68-positive MΦs in the islets of patients with T2D and in islets of mice fed HFD for 4–16 wk (11). Since then, more studies have reported islet MΦs as regulators of islet inflammation that have pathogenic impacts contributing to the development of T2D (8, 10, 22). On the other hand, MΦs can be protective to the islet, as they support β-cell replication by producing trophic factors that are necessary in some experimental rodent models of pancreas regeneration (2, 6, 38). Transplantation of wild-type, but not MΦ-deficient, Csf1r−/− bone marrow in immune-deficient mice leads to MΦ accumulation and angiogenesis within islets and subsequently increased islet mass (33). There are also emerging data demonstrating that activated islet MΦs, driven by hyperglycemia, stimulate islet α-cell proliferation (13, 14). Clearly, activated MΦs may reveal many distinct phenotypes within the islets. This is consistent with the fact that MΦs are incredibly plastic in response to various environmental stimuli that allows them to specialize and display polarized functional properties, such as inflammatory or anti-inflammatory actions in response to cytokine products (16).
Islet-infiltrating MΦs may secrete trophic factors that aid in β-cell replication under short-term HFD conditions similar to that occurring in pancreatic tissue injury models where MΦ-derived TGFβ was found to upregulate islet SMAD7 production leading to an increase in replication (2, 38). However, our initial characterization of the increased mouse islet-infiltrating MΦ population as a result of the short-term HFD treatment indicates that these are likely M1, the classic MΦs that engulf and digest cellular debris and foreign substances. M1 cells are capable of secreting proinflammatory cytokines such as tumor necrosis factor-α, interleukin-1β, etc., as well as NO, a product of iNOS. These factors have been suggested to play a detrimental role and result in islet dysfunction and death. However, in our study, no signs of islet damage were observed with short-term HFD treatment. The alternative type, named M2, is known to secrete a wide range of chemokines and growth factors to promote neovascularization and tissue repair. It is becoming clearer that more and more MΦ phenotypes in between these two polar types exist. A more detailed characterization of the islet-infiltrating MΦs, including phenotypes such as cytokines released, and time course studies are needed in future studies.
Our findings demonstrate that islet β-cell replication in response to high-fat feeding are dynamically regulated and challenge the perception that MΦ accumulation in islets is always a pathological feature as it occurs in T2D (9). It opens the possibility that the actions of MΦs may differ depending on the timing and pathological context within islets. This study lays the groundwork for defining the proliferative signals that are present in response to HFD. We argue that carefully unraveling the processes underlying HFD-induced changes in islet micro- and macroenvironments will illuminate the potential role of inflammation in β-cell replication. Furthermore, characterizing islet-infiltrating MΦs at different stages of HFD treatment will allow definition of MΦ ontogeny, heterogeneity, and functionality in normal, diseased, and regenerating islets, allowing proper therapeutic interventions at various stages of disease development and advancing our ability to preserve functional β-cell mass in persons with obesity and T2D.
This work was supported by funding from JDRF (17-2012-429), National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases (UC4 DK-104207), and a Startup fund from Columbia Center for Translational Immunology to X. Chen, a National Heart, Lung, and Blood Institute award (2-T32 HL-007854-19) to D. C. Woodland, and a NIDDK summer research fellowship award (5 U24 DK-076169-09) to J. Leong. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
No conflicts of interest, financial or otherwise, are declared by the author(s).
D.C.W., W.L., J.L., and M.L.S. performed experiments; D.C.W., W.L., J.L., M.L.S., and X.C. analyzed data; D.C.W., W.L., and X.C. interpreted results of experiments; D.C.W., W.L., and X.C. prepared figures; D.C.W., W.L., J.L., P.L., and X.C. edited and revised manuscript; P.L. and X.C. approved final version of manuscript; X.C. conception and design of research; X.C. drafted manuscript.
We are grateful for the excellent technical assistance provided by Sidikha Ashraf and Victoria Li.
- Copyright © 2016 the American Physiological Society