TY - JOUR
AU - Mistry,, Jehangir
AB - Polygenic mouse models for obesity-induced type 2 diabetes (T2D) more accurately reflect the most common manifestations of the human disease. Two inbred mouse strains (NON/Lt and NZO/HlLt) separately contributed T2D susceptibility- conferring quantitative trait loci to F1 males. Although chronic administration of rosiglitazone (Rosi) in diet (50 mg/kg) effectively suppressed F1 diabetes, hepatosteatosis was an undesired side effect. Three recombinant congenic strains (designated RCS1, -2, and -10) developed on the NON/Lt background carry variable numbers of these quantitative trait loci that elicit differential weight gain and male glucose intolerance syndromes of variable severity. We previously showed that RCS1 and -2 mice responded to chronic Rosi therapy without severe steatosis, whereas RCS10 males were moderately sensitive. In contrast, another recombinant congenic strain, RCS8, responded to Rosi therapy with the extreme hepatosteatosis observed in the F1. Longitudinal changes in multiple plasma analytes, including insulin, the adipokines leptin, resistin, and adiponectin, and plasminogen activator inhibitor-1 (PAI-1) allowed profiling of the differential Rosi responses in steatosis-exacerbated F1 and RCS8 males vs. the resistant RCS1 and RCS2 or moderately sensitive RCS10. Of these biomarkers, PAI-1 most effectively predicted adverse drug responses. Unexpectedly, mean resistin concentrations were higher in Rosi-treated RCS8 and RCS10. In summary, longitudinal profiling of multiple plasma analytes identified PAI-1 as a useful biomarker to monitor for differential pharmacogenetic responses to Rosi in these new mouse models of T2D. OBESE RODENTS PRODUCED by monogenic defects in either leptin or its receptor genes are commonly used in pharmaceutical research. However, the extreme phenotypes reflected in such monogenic obesity mutants are not reflective of the “garden-variety” human obesity/diabetes syndromes from multiple standpoints. The common human obesity/diabetes syndromes are polygenic, not monogenic in origin (1, 2). Furthermore, the clinical phenotypes observed in humans are only rarely as extreme as those produced by the mouse monogenic obesity mutations, e.g. massive juvenile-onset obesity associated with hyperphagia, extremes of leptin in circulation (either very high or none), extreme insulin resistance and hyperinsulinism, thermoregulatory defects, hypercorticism, and reproductive failure (3). In an effort to develop new mouse models of obesity-induced diabetes (diabesity) more closely resembling the human metabolic syndrome underlying type 2 diabetes (T2D), we generated new polygenic models through the construction of recombinant congenic strains (RCS) (4). These RCS all showed less extreme behavioral and endocrinologic phenotypes than did the mice with monogenic obesity mutations; yet all RCS developed moderate obesity and certain ones among them showed a maturity-onset transition from impaired glucose tolerance to a stable nonfasting hyperglycemia (4, 5). The strategy for RCS construction was to pick two unrelated inbred strains, NON/Lt and NZO/HlLt, each with separate sets of quantitative trait loci (QTL) contributing to impaired glucose tolerance. Males of the NON/Lt strain exhibit elevated fasting blood glucose, impaired glucose tolerance, and low glucose-stimulated insulin secretion from isolated islets, phenotypes indicative of an intrinsic pancreatic β-cell defect (6). NZO/HlLt males develop a polygenic obesity with an early body weight gain threshold determining those males that transit from impaired glucose tolerance to T2D. NON/Lt strain males, unlike NZO/HlLt males, do not spontaneously transit from impaired glucose tolerance to overt diabetes. However, the NON/Lt genome contributes QTL that interact negatively with NZO-derived QTL to exacerbate diabesity in (NZO × NON)F1 males (7). Knowledge of the chromosomal locations of certain of these diabesity QTL from both parental strains permitted development of a panel of 10 RCS by two cycles of backcrossing of the NZO QTL into the NON/Lt genetic background followed by inbreeding with selection for different combinations of diabesity-promoting QTL (4). Each RCS derives approximately 88% of its genome from the NON/Lt background strain and approximately 12% from the obese NZO/HlLt strain. With this introduction of variable numbers of obesity QTL, weight gain in most of the RCS exceeds that of the NON/Lt parental mice but none shows the extremely rapid weight gain of NZO/HlLt parental males. These RCS display a spectrum of male diabesity frequencies. For example, NONcNZO10 contains the maximum number of identified diabesity QTL from both parental strains and nearly all develop maturity-onset diabesity, whereas another, NONcNZO5, selected for resistance alleles at these QTL, remains diabesity-free despite developing comparable obesity. NONcNZO1 males develop diabesity at approximately the same frequency as parental NZO/HlLt males, but at a considerably lower body weight. Comparably weighted NONcNZO2 males, differing from NONcNZO1 males by the loss (through recombination) of one quantitative trait locus, originally designated as Nidd3, develop less diabesity (4). Hence, selected RCS express either diabetogenic or nondiabetogenic obesity syndromes. In those diabesity-prone RCS, the syndromes conferred by the different admixtures of QTL nicely model for the more common garden variety human obesity/diabesity syndromes in terms of a more moderate obesity without hyperphagia and reproductive failure and only moderately elevated insulin and leptin concentrations in plasma (4, 5). To determine the usefulness of these RCS as pharmacogenetic screening tools, we initially analyzed the efficacy of diet-administered rosiglitazone (Rosi) on the severe diabesity syndrome developing in (NZO × NON)F1 males (8). Although the drug produced excellent remission from hyperglycemia and hyperinsulinemia, chronic treatment also produced a significant increase in body weight and severely exacerbated an underlying hepatic steatosis. This hepatosteatotic effect was associated with unusual changes in the lipid-metabolic profile of the treated mice (8). More recently, males from selected RCS, as well as males of the NON/Lt parental background strain and the hepatosteatosis-prone F1 males, were treated with Rosi from 8–22 wk of age (9). Development of hyperglycemia and hyperinsulinemia was completely suppressed in NONcNZO10, the RCS that, when left untreated, showed a comparable diabesity frequency to that of F1 males. However, unlike the extreme hepatosteatosis produced by Rosi in F1 males, hepatic steatosis was only moderately increased. In contrast, in comparably weighted NONcNZO8 males with lower diabesity frequency on chow diet, Rosi administration created the same severe hepatosteatotic effect as observed in the considerably more obese F1 males. Neither diabesity-prone NONcNZO1 nor diabesity-resistant NONcNZO2 males exhibited this adverse pharmacologic effect. These differential pharmacogenetic responses suggest that the RCS are variously fixed for combinations of diabesity QTL that may regulate hepatosteatotic responsiveness to Rosi. In the present study, we report the results of longitudinal profiling of plasma changes in endocrine factors using a multiplex assay system that permits differentiation of these various polygenic obesity/diabesity syndromes. Materials and Methods Mice NON/Lt, (NONxNZO)F1, and four previously reported RCS (4) [NONcNZO1 (designated RCS1), NONcNZO2 (RCS2), NONcNZO8 (RCS8), and NONcNZO10 (RCS10)] were bred in our research vivarium. They were housed 2–4 per pen in double pen plexiglass boxes on autoclaved shaved pine bedding and given free access to food and acidified water. All mice shared the same mouse room with controlled temperature and humidity and a 14-h light, 10-h dark cycle. Nine to 12 males of each strain at 8 wk of age were split into two groups. Control group males continued to receive the maintenance (control) diet (NIH-31 containing 6% fat, Purina Test Diets, Richmond, IN). The remaining half received the same diet supplemented (wt/wt) with 50 mg Rosi/kg (a kind gift from Dr. S. Smith, GlaxoSmithKline). Mice were maintained on the irradiation-sterilized diets for 14 wk and represent the same mice described in another study examining the hepatosteatotic effects of drug treatment at the 22-wk necropsy point (9). All procedures involving the use of animals were approved by the Animal Care and Use Committee of The Jackson Laboratory. Clinical markers For longitudinal profiling of plasma endocrine changes in response to chronic Rosi treatment, 100-μl volumes of blood were collected in two heparinized capillary tubes from the retroorbital sinus and plasma obtained over the biweekly time points shown in Figs. 2 and 3. Plasma glucose levels were measured by a Glucose 2 analyzer (Beckman Instruments, Fullerton, CA). Plasma insulin, leptin, resistin, and plasminogen activator inhibitor 1 (PAI-1) were measured with a commercially available LINCOplex Mouse Adipokine Immunoassay kit from Linco Research, Inc., capable of simultaneously measuring six different adipokines in mouse serum or plasma. Because one of the RCS (RCS8) exhibited an unanticipated resistin and PAI-1 response to Rosi treatment (see below), adiponectin was also measured in that group by the LINCOplex bead assay kit specific for mouse adiponectin (detecting all forms except the monomer). The two kits are based on the Luminex xMAP technology and use microsphere bead sets that are uniquely labeled with a mixture of two fluorescent dyes. The capture antibodies specific for each analyte are covalently coupled to individual bead sets. At the time of the assay, a mixture of beads is incubated with 10 μl of standards or mouse serum samples in a 96-well filter-bottom plate overnight at 4 C. On the next day, the beads are washed; biotinylated detection antibody cocktail is added and incubated for 30 min at room temperature, followed by the addition of streptavidin-phycoerythrin and incubation for another 30 min. After a final wash, the resuspended beads are read on a Luminex 100 reader and the concentration of each analyte in unknown samples is determined based on individual standard curves. For the longitudinal profile measurement of each endocrine factor comparing NON/Lt to the RCS generated on the NON/Lt genetic background, plasma samples were collected from each Rosi-untreated and Rosi-treated cohort (n = 5–6 mice per strain except for Rosi-treated RCS2, where n = 4) at 2-wk intervals from 8–20 wk, stored frozen until all the time points were collected, and then all the samples were run at the same time. At the 22-wk necropsy point, brown adipose tissue, inguinal fat, and gonadal fat were dissected from each mouse and weighed. Fig. 2. Open in new tabDownload slide Temporal changes in nonfasting matinal plasma glucose of Rosi-untreated (□) vs. Rosi-treated (•) strains. Significant differences for main effects over the course of treatment are noted where applicable. Fig. 2. Open in new tabDownload slide Temporal changes in nonfasting matinal plasma glucose of Rosi-untreated (□) vs. Rosi-treated (•) strains. Significant differences for main effects over the course of treatment are noted where applicable. Fig. 3. Open in new tabDownload slide Temporal changes in plasma insulin of Rosi-untreated (□) vs. Rosi-treated (•) strains. Significant differences for main effects over the course of treatment are noted where applicable. Fig. 3. Open in new tabDownload slide Temporal changes in plasma insulin of Rosi-untreated (□) vs. Rosi-treated (•) strains. Significant differences for main effects over the course of treatment are noted where applicable. Statistics For longitudinally sampled phenotypes, a mixed linear model analysis with repeated measures was performed using PROC MIXED SAS 9.1.3 (SAS Institute, Inc., Cary, NC). In this analysis, treatment effects, strain effects and treatment by strain-interaction effects were treated as between-subject effects, whereas time and time by between-subject effect interactions were treated as within-subject effects (10). All effects, including strain, treatment, time, and their interactions, were tested. Akaike’s Information Criterion was used for model selection. PROC MIXED was also used to test within-strain treatment effects by comparing least-square means. ANOVA (JMP, SAS Institute, Inc., Cary, N.C.) was used for within-strain comparisons for effect of treatment on fat pad weights. Significance was assumed at P ≤ 0.05. Results Longitudinal profiling of changes in body weight, plasma glucose, and plasma insulin Repeated-measure testing for the six phenotypes analyzed (Table 1) revealed significant between- and within-subjects effects. Significant time by treatment by strain interaction was observed for all variables except resistin. Within-strain responses to Rosi treatment across time are shown in Figs. 1–4. Data in Fig. 1 display the weight gains of the parental NON/Lt background males vs. the selected RCS beginning at 8 wk of age, the time at which mice were either left untreated or switched to the Rosi-supplemented diet. Between the 8–20-wk study period, Rosi-untreated NON/Lt male controls gained weight gradually, with no significant increase produced by Rosi supplementation of the diet. The untreated F1 males were markedly obese at 8 wk and continued to increase their body weight to a mean of 57 g by 20 wk. Body weights for Rosi-treated F1 were higher than untreated controls (P = 0.0564) from 12 wk onward, with a 66 g mean body weight recorded at 20 wk. Although the main effect of treatment was only of borderline significance, a test of treatment by time interaction showed a highly significant difference (P < 0.0001). None of the Rosi-untreated RCS males showed the NZO-like development of massive obesity characteristic of the F1 males. Indeed, Rosi-untreated RCS1 and RCS2 control males showed postpubertal patterns of weight gain quite similar to the NON/Lt parental background strain. RCS8 and RCS10 fed the control chow showed rates of weight gain intermediate between the NON parental background strain and the extremely obese F1 males (Fig. 1). Interestingly, Rosi did not produce increased weight gain in either NON or RCS1 males, the latter previously found to develop a late maturity-onset diabetes at approximately 50% frequency (4). In contrast, Rosi-supplemented diet produced significant increases in weight gain over time in RCS2, RCS8, and RCS10 males (Fig. 1). TABLE 1. P values for each phenotype as a function of treatment, strain, and time Trt Strain Trt × strain Time Time × Trt Time × strain Time × Trt × strain BW <0.0001 <0.0001 0.003 <0.0001 <0.0001 <0.0001 <0.0001 PG 0.0021 0.0552 NS 0.0063 <0.0001 0.0113 0.0032 Insulin <0.0001 0.0002 0.0032 <0.0001 <0.0001 <0.0001 0.0005 Leptin 0.0463 NS NS <0.0001 0.0063 0.0002 0.0397 PAI-1 NS <0.0001 0.0316 0.0111 0.0004 <0.0001 0.0008 Resistin 0.0048 <0.0001 0.0056 NS NS <0.0001 NS Trt Strain Trt × strain Time Time × Trt Time × strain Time × Trt × strain BW <0.0001 <0.0001 0.003 <0.0001 <0.0001 <0.0001 <0.0001 PG 0.0021 0.0552 NS 0.0063 <0.0001 0.0113 0.0032 Insulin <0.0001 0.0002 0.0032 <0.0001 <0.0001 <0.0001 0.0005 Leptin 0.0463 NS NS <0.0001 0.0063 0.0002 0.0397 PAI-1 NS <0.0001 0.0316 0.0111 0.0004 <0.0001 0.0008 Resistin 0.0048 <0.0001 0.0056 NS NS <0.0001 NS BW, Body weight; PG, plasma glucose; NS, not significant; Trt, treatment. Open in new tab TABLE 1. P values for each phenotype as a function of treatment, strain, and time Trt Strain Trt × strain Time Time × Trt Time × strain Time × Trt × strain BW <0.0001 <0.0001 0.003 <0.0001 <0.0001 <0.0001 <0.0001 PG 0.0021 0.0552 NS 0.0063 <0.0001 0.0113 0.0032 Insulin <0.0001 0.0002 0.0032 <0.0001 <0.0001 <0.0001 0.0005 Leptin 0.0463 NS NS <0.0001 0.0063 0.0002 0.0397 PAI-1 NS <0.0001 0.0316 0.0111 0.0004 <0.0001 0.0008 Resistin 0.0048 <0.0001 0.0056 NS NS <0.0001 NS Trt Strain Trt × strain Time Time × Trt Time × strain Time × Trt × strain BW <0.0001 <0.0001 0.003 <0.0001 <0.0001 <0.0001 <0.0001 PG 0.0021 0.0552 NS 0.0063 <0.0001 0.0113 0.0032 Insulin <0.0001 0.0002 0.0032 <0.0001 <0.0001 <0.0001 0.0005 Leptin 0.0463 NS NS <0.0001 0.0063 0.0002 0.0397 PAI-1 NS <0.0001 0.0316 0.0111 0.0004 <0.0001 0.0008 Resistin 0.0048 <0.0001 0.0056 NS NS <0.0001 NS BW, Body weight; PG, plasma glucose; NS, not significant; Trt, treatment. Open in new tab Fig. 1. Open in new tabDownload slide Temporal changes in body weight of Rosi-untreated (□) vs. Rosi-treated (•) strains. Significant differences for main effects over the course of treatment are noted where applicable. Fig. 1. Open in new tabDownload slide Temporal changes in body weight of Rosi-untreated (□) vs. Rosi-treated (•) strains. Significant differences for main effects over the course of treatment are noted where applicable. Fig. 4. Open in new tabDownload slide Temporal changes in plasma leptin, resistin, and PAI-1 of strains of Rosi-untreated (open diamonds with solid line) vs. Rosi-treated (open boxes with dotted lines) strains. Significant differences for main effects over the course of treatment are noted where applicable. Fig. 4. Open in new tabDownload slide Temporal changes in plasma leptin, resistin, and PAI-1 of strains of Rosi-untreated (open diamonds with solid line) vs. Rosi-treated (open boxes with dotted lines) strains. Significant differences for main effects over the course of treatment are noted where applicable. Data in Fig. 2 display the glycemic status of the RCS over time compared with the NON parental background on which selected NZO QTL had been introgressed. Both Rosi-untreated and treated NON males remained normoglycemic throughout the study period. Untreated F1, RCS8, and RCS10 males showed an aging-dependent increase in glycemia to mean levels of more than 250 mg/dl by 20 wk. RCS10 males contain only a selected subset of the many codominant or additive obesity QTL found in NZO males, all of which are present in heterozygous state in the F1 males. Nevertheless, untreated RCS10 males developed an obesity-induced hyperglycemia comparable to that developed in the F1 males (Fig. 2). Rosi treatment over the same time course significantly (P = 0.0033) suppressed hyperglycemia development in F1 as well as in diabesity-prone RCS10 males (P = 0.0001) and the more diabesity-resistant RCS1 males (P = 0.0425). RCS2 males, like the NON/Lt males, were completely diabesity resistant. RCS8 males did not show a significant change in glycemic state in response to Rosi over the time course followed (Fig. 2). Inability to show a significant main effect of treatment in this strain may, in part, be due to a more protracted (and variable) time course at which individual RCS8 males transit from normoglycemia to hyperglycemia compared with the more rapid diabetes development in either F1 or RCS10 males (4). Hence, it appears that untreated controls were just transiting into overt hyperglycemia at 20 wk, the last time point sampled (Fig. 2). Indeed, testing for interaction of treatment by time showed borderline significance (P = 0.06). Comparative plasma insulin data are shown in Fig. 3. NON and RCS2 males maintained a constant, low plasma insulin over the study period that was not significantly altered by dietary Rosi exposure. Consistent with their development of extreme obesity, F1 males developed the most extreme hyperinsulinemia over the study period. The Rosi-mediated suppression of hyperglycemia shown in Fig. 2 was accompanied by suppression of hyperinsulinemia development (Fig. 3). RCS1, -8, and -10 males developed a modest, age-dependent hyperinsulinemia. In RCS10, dietary administration of Rosi throughout the same period completely suppressed hyperinsulinemia development. It is noteworthy that RCS10 males, which exhibited higher mean hyperglycemic values than the heavier F1 males, did not exhibit the early development of hyperinsulinemia observed in the latter. Indeed, an unusual feature of diabesity development in RCS10 that distinguished it from the F1 model was the modest elevations in plasma insulin. This followed rather than preceded transition to hyperglycemia seen already by 10 wk of age (Fig. 2). Although chronic Rosi treatment of RCS8 males suppressed further progression of hyperinsulinemia, the treatment never restored plasma insulin levels below a relatively high 5 ng/ml level (Fig. 3), suggesting that RCS8 males were more resistant to the antidiabetogenic action of Rosi. No significant main or interaction effects distinguished plasma insulin changes in Rosi-treated vs. untreated RCS1 males, perhaps because hyperinsulinemia was so mild and variable in untreated controls. Longitudinal changes in plasma adipokines Data in Fig. 4 show the longitudinal changes in plasma leptin, resistin, and PAI-1 concentrations in response to Rosi treatment. Not surprisingly, the rapid increase in both subcutaneous and visceral adiposity in hyperphagic and Rosi-untreated (NONxNZO)F1 males produced higher plasma leptin concentrations than observed in Rosi-untreated NON or any of the Rosi-untreated RCS males. Mice fed the Rosi-supplemented diet showed strain-dependent changes in plasma leptin. Corroborating the observation that both NON or RCS1 males were resistant to the adiposity-stimulating effects of chronic Rosi feeding (Fig. 1), both strains also failed to increase mean leptin concentrations above control values. In contrast, those strains where Rosi treatment produced age-associated increases in body weight (e.g. F1, RCS8, and RCS10) showed significantly increased plasma leptin concentrations above Rosi-untreated control values. RCS2 leptin values trended higher in the Rosi-treated group but statistical testing did not show a significant difference. Chronic Rosi treatment has been reported to suppress resistin secretion into plasma in obese mouse models such as leptin-deficient (ob/ob) mice, leptin receptor-deficient (db/db) mice, and diet-induced obese B6 mice (11). However, in this panel of mouse strains, Rosi-mediated changes in temporal patterns of plasma resistin were more complex (Fig. 4). Interestingly, the glucose-intolerant, but nondiabetic NON/Lt males registered the highest initial resistin levels at 8 wk of age. However, these concentrations declined sharply with age (∼6-fold within 4 wk), with Rosi-treated NON/Lt males showing an identical pattern. Resistin was initially lower in all RCS at 8 wk compared with NON/Lt, with Rosi treatment producing significant main effects only in RCS8 (P = 0.0058) and RCS10 (P = 0.0012), the two most diabesity-prone RCS strains. In both these strains treated with Rosi, resistin concentrations were significantly increased over time in comparison with concentrations in Rosi-untreated controls (Fig. 4). Genetic disruption of the gene encoding PAI-1 in mice confers resistance to diet-induced insulin resistance and obesity (12). Hence, because chronic Rosi treatment of the strains in this study prevented development of hyperglycemia and reduced plasma insulin concentrations, we anticipated that the increased insulin sensitivity would correlate with decreases in plasma concentrations of this adipokine. As shown in Fig. 4, a significant increase in mean PAI-1 concentrations was demonstrable in Rosi-treated F1, RCS8, and RCS10 males. This analyte in RCS10 males showed a significant main effect of Rosi treatment, but the absolute values of PAI-1 were not showing temporal increases comparable with those seen in Rosi-treated F1 and RCS8 males. When these two cohorts were necropsied at 22 wk of age, and liver histology analyzed, they were distinguished from all the others in terms of a more extreme macrovesicular steatosis (9). Because significant main effects on elevated plasma resistin concentrations were also observed in RCS8, we examined adiponectin levels in our RCS8 samples to verify that Rosi was indeed effective as an insulin sensitizer. As shown in Fig. 5, Rosi treatment produced a significant treatment by time increase in plasma adiponectin. Thus, the failure of Rosi to suppress PAI-1 secretion over time, at least in RCS8, was dissociated from its potential insulin-sensitizing effects and very likely reflected the hepatic injury (see Discussion). Fig. 5. Open in new tabDownload slide Temporal changes in plasma adiponectin of Rosi-untreated RCS8 controls (⋄) vs. Rosi-treated (□) strains. Significant differences for main effects over the course of treatment are noted where applicable. Fig. 5. Open in new tabDownload slide Temporal changes in plasma adiponectin of Rosi-untreated RCS8 controls (⋄) vs. Rosi-treated (□) strains. Significant differences for main effects over the course of treatment are noted where applicable. Rosi-mediated changes in fat pad weights Chronic Rosi treatment significantly increased brown adipose depots (Fig. 6A) and inguinal fat depots (Fig. 6B) in all RCS and F1 males. The gonadal fat depot, in contrast, did not show comparable Rosi-mediated increases (Fig. 6C). Because body weights of all RCS except RCS1 taken at necropsy were significantly increased by Rosi, chronic Rosi treatment actually produced decreases in the ratios of gonadal fat pad weights to body weights (Fig. 6D). The basis for this differential drug effect on specific fat depots is unclear, but reflects the complexity of the drug by tissue interaction previously observed in multiple tissues of the (NZO × NON)F1 males on Rosi treatment (8). Fig. 6. Open in new tabDownload slide Differences in fat pad weights at necropsy of Rosi-untreated controls (open bars) vs. Rosi-treated (closed bars) males. Significant differences were determined by ANOVA with P < 0.05 (*) and P < 0.005 (**) noted. Fig. 6. Open in new tabDownload slide Differences in fat pad weights at necropsy of Rosi-untreated controls (open bars) vs. Rosi-treated (closed bars) males. Significant differences were determined by ANOVA with P < 0.05 (*) and P < 0.005 (**) noted. Discussion This study represents the first longitudinal profiling of changes in multiple plasma endocrine/adipokine analytes in new polygenic mouse models of T2D. The new models are important not only because their polygenic mode of inheritance is more reflective of the genetics of human obesity and T2D (1, 2), but also because the phenotypes (body weight, plasma insulin, plasma leptin) are less extreme than either the NZO/HlLt or F1 males from whence the new models were developed. Each RCS represents a selected subset of NZO diabesity QTL introgressed into the NON/Lt genetic background. In some respects, the NON strain resembles the GK (Goto-Kakazaki) rat (13) in that NON males in particular show a very impaired glucose tolerance in response to parenteral glucose administration, an impairment associated with a decreased first phase and blunted second phase insulin secretory response (6). As shown in the present study, NON males nevertheless are able to maintain nonfasting euglycemia on the 6% fat-containing chow diet used in the present study, despite the low plasma insulin concentrations. In contrast, diabesity does develop in approximately 50% of NZO/HlLt males fed a 4% fat-containing diet. Although all NZO/HlLt males become morbidly obese by 12 wk of age, the male subset developing diabesity is the one showing the greatest rate of prematurational weight gain (14). The degree of morbid obesity does not distinguish NZO males from the more intensively studied monogenic obesity models such as the Lepob (ob/ob) mouse or the Leprdb (db/db) mouse. What distinguishes both the NZO and the (NON × NZO)F1 males from the latter models is the maturity-onset development of hyperglycemia, hyperinsulinemia, and hyperleptinemia. As shown in Fig. 3 for plasma insulin and Fig. 4 for plasma leptin, the concentrations of these analytes are not markedly elevated at the 8-wk age point, even though the males are already quite obese (Fig. 1). RCS10 illustrates how a mouse model can be further refined to better reflect the so-called garden variety T2D patients. These patients are commonly overweight/obese but not morbidly so. RCS10 was generated by transfer of the maximum number of known NZO diabesity-promoting QTL onto the diabesity-sensitized NON genetic background (4, 5). The present data show that compared with the Rosi-untreated and morbidly obese F1 male controls, Rosi-untreated RCS10 male controls are developing diabesity at a higher frequency (reflected in Fig. 2 by a higher mean plasma glucose at the various sampling points) but at a body weight considerably lower than observed in F1 males (Fig. 1). Because it develops the highest frequency of diabesity, RCS10 serves as a useful starting point for comparing longitudinal changes in plasma insulin and leptin concentrations as the males gain weight. Plasma insulin concentrations in Rosi-untreated RCS10 male controls remained modestly elevated during the postmaturational period (at least 10–14 wk; see Fig. 3), a period during which loss of glucose homeostasis was evident. Attainment of a more prominent hyperinsulinemia at the 20-wk sampling point suggested that RCS10 has inherited the NON strain characteristic of pancreatic β-cell insensitivity to acute glucose challenge. Similarly, RCS8 males, although more slow to develop a stable hyperglycemic pattern, also exhibited this retarded insulin hypersecretory response. By contrast to both these hyperglycemia-developing RCS, the F1 males reflect the published NZO insulin hypersecretory behavior and manifest it at a much earlier time point and at a considerably lower mean hyperglycemia (Fig. 2). Rosi-untreated RCS8 male controls resembled RCS10 in terms of weight gain but were more F1-like in terms of developing a less extreme hyperglycemia. The high standard errors with plasma glucose in Rosi-untreated RCS8 male controls reflected a much higher interindividual variation than was observed in RCS10. Temporal changes in neither plasma leptin nor plasma resistin concentrations distinguished Rosi-untreated RCS8 from RCS10 male controls. However, chronic Rosi feeding of RCS8 elicited an unexpected phenotype: a significant increase in plasma PAI-1 over time, a phenotype shared only with the extremely obese F1 males (Fig. 4). As expected, Rosi feeding suppressed the temporal increases in hyperglycemia and hyperinsulinemia in F1 and RCS10 strains. Rosi also produced increases in mean body weight of F1 and RCS2, RCS8, and RCS10 males. We have previously reported necropsy data at 22 wk for the same cohorts of Rosi-treated and untreated males (15) for which longitudinal plasma profiling is reported in this study. The salient histopathologic finding was that, like the long-term-treated F1 males, but unlike males of the other RCS, RCS8 males developed severe Rosi-exacerbated hepatosteatosis (15). This adverse side effect also developed in diabetic F1 males treated with Rosi for only a month (8). Lipid metabolome profiling of livers from these Rosi-treated F1 males indicated that triglyceride accumulation and de novo fatty acid synthesis, rather than phospholipid production and lipoprotein export, were being favored. This F1 and RCS8 sensitivity to the hepatosteatotic effect of Rosi was correlated with suppression of enzyme components of both pathways for phosphatidylcholine biosynthesis in the liver (9, 15). This suppression of at least two of the major phosphatidylcholine biosynthetic enzymes was found to be peroxisome proliferator-activated receptor (PPAR) γ dependent (15). Thiazolidinedione-mediated increases in plasma adiponectin may be a biomarker for restoration of insulin sensitivity in human T2D patients (16). The finding of significant Rosi-mediated increases in plasma adiponectin in RCS8, in combination with the suppression of hyperglycemia and hyperinsulinemia development, confirmed a potent insulin-sensitizing action. Resistin is an important regulator of glucose metabolism in mouse models (17). Several studies conducted to evaluate thiazolidinedione effects on resistin measurement in various models demonstrate an inconsistent trend in levels of mRNA and of protein. For example, Steppan et al. (11) originally reported that circulating resistin concentrations were elevated in various obese models (ob/ob, db/db, and diet-induced obese B6 mice) and that Rosi treatment lowered serum resistin in these animals. In contrast, Way et al. (18) reported that resistin mRNA levels were severely decreased in various obese mouse models (ob/ob, db/db, diet-induced obesity) and that Rosi treatment resulted in 8.5-fold increase in mRNA in white adipose tissue of ob/ob mice. Fukui and Motojima (19) also found that an 8-d treatment with pioglitazone or troglitazone slightly induced resistin levels in lean C57BL/6, as well as in KK, KKAy, and db/db mice. We found that Rosi treatment increased brown adipose tissue and inguinal fat depot weights (but not gonadal fat pad weights) in all strains tested (Fig. 6), but Rosi-mediated resistin increases were significant only in RCS8 and RCS10. Overall, circulating levels of resistin could be influenced by body weight, degree of adiposity, levels of insulin and leptin, dose and treatment duration of thiazolidinedione, mouse strain used, as well as other factors such as stability or clearance rate of the protein. In the present study, NON and the RCS derived from it differ from F1 and NZO/HlLt males in that adiposity in the former is primarily due to increased visceral rather than a combination of visceral and subcutaneous white fat. Chronic Rosi feeding produced a significant treatment by time increase in body weight of RCS8 and RCS10 males (Fig. 1), the only two RCS to show a comparably significant treatment by time effect in circulating resistin (Fig. 4). Because Rosi suppressed development of hyperglycemia in the diabesity-prone F1 males without significantly changing plasma resistin concentrations, we have no evidence to suggest that diabesity-related secondary modifications in the resistin protein might account for the increased immunoassayable concentrations limited to RCS8 and RCS10 (Fig. 4). Rather, the relatively late-developing increases in Rosi-treated RCS8 and 10 males suggest that the change is more likely associated with Rosi-mediated increases in adiposity. The effect of PPARγ ligands on PAI-1 levels has been studied in human endothelial cells (20, 21) and adipose tissue (22). PPARγ ligands increase PAI-1 mRNA and protein expression in human endothelial cells. PAI-1 secretion from human white fat was reported to be suppressed after thiazolidinedione-mediated increase in insulin sensitivity. In the present study, we found that, among the strains tested, the F1 and RCS8 males showed a significant treatment by time increase in PAI-1 despite evidence for improved insulin sensitivity in both. In the case of RCS8, the only RCS matching the F1 stock in terms of the extent of Rosi-exacerbated hepatosteatosis (9, 15), increased insulin sensitivity was suggested not only by the blunting of age-dependent hyperinsulinemia development (Fig. 3) but also a significant treatment by time increase in circulating adiponectin (Fig. 5). If the unsuppressed PAI-1 is coming from the fatty liver, it may serve as a potential biomarker for development of severe hepatosteatosis in these models. Indeed, major sources of circulating PAI-1 are reported to be endothelial cells, hepatocytes, and adipocytes (20). In summary, we have characterized longitudinal changes in multiple plasma analytes that collectively better define newly developed mouse models of obesity, insulin resistance, and T2D. The panel of RCS, in combination with the multiplex assay system, provides useful tools to dissect pharmacogenetic differences in the new obesity/diabetes models with minimal volumes of mouse blood per sampling. Acknowledgments We gratefully acknowledge the skilled technical assistance of Ms. Pam Stanley. This work was supported by National Institutes of Health Grant DK56853 (to E.H.L.). H.J.P. was supported by a mentor-based fellowship from The American Diabetes Association. Institutional shared services were supported by National Cancer Institute Center Support Grant CA34196. Present address for H.-j.P.: Institute of Biomedical Sciences, Academica Sinica, Taipei, Taiwan. Abbreviations PAI-1, Plasminogen activator inhibitor-1; PPAR, peroxisome proliferator-activated receptor; QTL, quantitative trait loci; RCS, recombinant congenic strain(s); Rosi, rosiglitazone; T2D, type 2 diabetes. References 1 Chiu YF , Chuang LM , Hsiao CF , Hung YJ , Lin MW , Chen YT , Grove J , Jorgenson E , Quertermous T , Risch N , Hsiung CA 2005 An autosomal genome-wide scan for loci linked to pre-diabetic phenotypes in nondiabetic Chinese subjects from the Stanford Asia-Pacific Program of Hypertension and Insulin Resistance Family Study. Diabetes 54 : 1200 – 1206 Google Scholar Crossref Search ADS PubMed WorldCat 2 Pratley R , Thompson D , Prochazka M , Baier L , Pima Diabetes Genes Group 1997 A genome scan for linkage to quantitative traits predicting NIDDM in Pima Indians. Diabetes 46 : 170 A WorldCat 3 Herberg L , Leiter EH 2001 Obesity/diabetes in mice with mutations in the leptin or leptin receptor genes. In: Sima AAF , Shafrir E , eds. Animal models of diabetes: a primer. Amsterdam : Harwood Academic Publishers ; 63 – 107 COPAC 4 Reifsnyder PC , Leiter EH 2002 Deconstructing and reconstructing obesity-induced diabetes (diabesity) in mice. Diabetes 51 : 825 – 832 Google Scholar Crossref Search ADS PubMed WorldCat 5 Leiter EH , Reifsnyder PC 2004 Differential levels of diabetogenic stress in two new mouse models of obesity and type 2 diabetes . Diabetes 53 ( Suppl 1 ): S4 – S11 Google Scholar Crossref Search ADS PubMed WorldCat 6 Leiter EH , Herberg L 1997 The polygenetics of diabesity in mice. Diabetes Rev 5 : 131 – 148 WorldCat 7 Leiter EH , Reifsnyder PC , Flurkey K , Partke H-J , Junger E , Herberg L 1998 Non-insulin dependent diabetes genes in mice: deleterious synergism by both parental genomes contributes to diabetogenic thresholds. Diabetes 47 : 1287 – 1295 Google Scholar Crossref Search ADS PubMed WorldCat 8 Watkins SM , Reifsnyder PC , Pan H-J , German B , Leiter EH 2002 Lipid metabolome-wide effects of the peroxisome proliferator-activated receptor gamma agonist rosiglitazone on a new mouse model of type 2 diabetes. J Lipid Res 43 : 1809 – 1817 Google Scholar Crossref Search ADS PubMed WorldCat 9 Pan H-J , Reifsnyder PC , Vance DE , Xiao Q , Leiter EH 2005 Pharmacogenetic analysis of rosiglitazone-induced hepatosteatosis in new mouse models of type 2 diabetes. Diabetes 54 : 1854 – 1862 Google Scholar Crossref Search ADS PubMed WorldCat 10 Littell R , Milliken G , Stroup W , Wofinger R 1996 SAS® system for mixed models. Cary, NC : SAS Institute, Inc. ; 87 – 134 COPAC 11 Steppan CM , Bailey ST , Bhat S , Brown EJ , Banerjee RR , Wright CM , Patel HR , Ahima RS , Lazar MA 2001 The hormone resistin links obesity to diabetes. Nature 409 : 307 – 312 Google Scholar Crossref Search ADS PubMed WorldCat 12 Ma LJ , Mao SL , Taylor KL , Kanjanabuch T , Guan Y , Zhang Y , Brown NJ , Swift LL , McGuinness OP , Wasserman DH , Vaughan DE , Fogo AB 2004 Prevention of obesity and insulin resistance in mice lacking plasminogen activator inhibitor 1. Diabetes 53 : 336 – 346 Google Scholar Crossref Search ADS PubMed WorldCat 13 Portha B , Serradas P , Bailbe D , Suzuki K-I , Goto Y , Giroix M-H 1991 β-cell insensitivity to glucose in the GK rat, a spontaneous nonobese model for type II diabetes. Diabetes 40 : 486 – 491 Google Scholar Crossref Search ADS PubMed WorldCat 14 Koza RA , Flurkey K , Graunke DM , Braun C , Pan HJ , Reifsnyder PC , Kozak LP , Leiter EH 2004 Contributions of dysregulated energy metabolism to type 2 diabetes development in NZO/H1Lt mice with polygenic obesity. Metabolism 53 : 799 – 808 Google Scholar Crossref Search ADS PubMed WorldCat 15 Pan H-J, Watkins S , Lin Y , Chen YE , Vance DE , Leiter EH , Adverse hepatic and cardiac responses to rosiglitazone in new mouse model of type 2 diabetes: relation to dysregulated phosphatidylcholine metabolism. Vasc Pharm, in press COPAC 16 Miyazaki Y , Mahankali A , Wajcberg E , Bajaj M , Mandarino LJ , DeFronzo RA 2004 Effect of pioglitazone on circulating adipocytokine levels and insulin sensitivity in type 2 diabetic patients. J Clin Endocrinol Metab 89 : 4312 – 4319 Google Scholar Crossref Search ADS PubMed WorldCat 17 Banerjee RR , Rangwala SM , Shapiro JS , Rich AS , Rhoades B , Qi Y , Wang J , Rajala MW , Pocai A , Scherer PE , Steppan CM , Ahima RS , Obici S , Rossetti L , Lazar MA 2004 Regulation of fasted blood glucose by resistin. Science 303 : 1195 – 1198 Google Scholar Crossref Search ADS PubMed WorldCat 18 Way JM , Gorgun CZ , Tong Q , Uysal KT , Brown KK , Harrington WW , Oliver Jr WR , Willson TM , Kliewer SA , Hotamisligil GS 2001 Adipose tissue resistin expression is severely suppressed in obesity and stimulated by peroxisome proliferator-activated receptor gamma agonists. J Biol Chem 276 : 25651 – 25653 Google Scholar Crossref Search ADS PubMed WorldCat 19 Fukui Y , Motojima K 2002 Expression of resistin in the adipose tissue is modulated by various factors including peroxisome proliferator-activated receptor alpha. Diabetes Obes Metab 4 : 342 – 345 Google Scholar Crossref Search ADS PubMed WorldCat 20 Marx N , Bourcier T , Sukhova GK , Libby P , Plutzky J 1999 PPARγ activation in human endothelial cells increases plasminogen activator inhibitor type-1 expression: PPARγ as a potential mediator in vascular disease. Arterioscler Thromb Vasc Biol 19 : 546 – 551 Google Scholar Crossref Search ADS PubMed WorldCat 21 Xin X , Yang S , Kowalski J , Gerritsen ME 1999 Peroxisome proliferator-activated receptor gamma ligands are potent inhibitors of angiogenesis in vitro and in vivo. J Biol Chem 274 : 9116 – 9121 Google Scholar Crossref Search ADS PubMed WorldCat 22 Zirlik A , Leugers A , Lohrmann J , Ernst S , Sobel BE , Bode C , Nordt TK 2004 Direct attenuation of plasminogen activator inhibitor type-1 expression in human adipose tissue by thiazolidinediones. Thromb Haemost 91 : 674 – 682 Google Scholar PubMed WorldCat Copyright © 2006 by The Endocrine Society
TI - Differential Endocrine Responses to Rosiglitazone Therapy in New Mouse Models of Type 2 Diabetes
JO - Endocrinology
DO - 10.1210/en.2005-0839
DA - 2006-02-01
UR - https://www.deepdyve.com/lp/oxford-university-press/differential-endocrine-responses-to-rosiglitazone-therapy-in-new-mouse-NOXNO0E0xt
SP - 919
VL - 147
IS - 2
DP - DeepDyve
ER -