Abstract
Obesity and diabetes are closely associated with hyperactivation of the hypothalamic–pituitary–adrenal (HPA) axis. In this study, the diet-induced obese C57BL/6 mouse was used to test the hypothesis that chronically elevated metabolic parameters associated with the development of obesity such as cholesterol and glucose can aggravate basal HPA axis activity. Because the lipocalin-type prostaglandin D2 synthase (L-PGDS) knockout (KO) mouse is a model of accelerated insulin resistance, glucose intolerance, and obesity, it was further hypothesized that HPA activity would be greater in this model. Starting at 8 weeks of age, the L-PGDS KO and C57BL/6 mice were maintained on a low-fat or high-fat diet. After 20 or 37 weeks, fasting metabolic parameters and basal HPA axis hormones were measured and compared between genotypes. Correlation analyses were performed to identify associations between obesity-related chronic metabolic changes and changes in the basal activity of the HPA axis. Our results have identified strong positive correlations between total cholesterol, LDL-cholesterol, glucose, and HPA axis hormones that increase with age in the C57BL/6 mice. These data confirm that obesity-related elevations in cholesterol and glucose can heighten basal HPA activity. Additionally, the L-PGDS KO mice show early elevations in HPA activity with no age-related changes relative to the C57BL/6 mice.
Introduction
Obesity and diabetes are closely associated with a hyperactive hypothalamic–pituitary–adrenal (HPA) axis. There is evidence that chronic stress can trigger obesity through the increased activity of the HPA axis. Chronic stress-induced elevations of glucocorticoids (GCs) increase expression of corticotrophin-releasing factor in the amygdale, increase the salience of compulsive activities such as the ingestion of high-fat (HF) foods, as well as increase abdominal fat depots (Dallman et al. 2003, Shin et al. 2010). Although many studies have focused on hyperactivity of the HPA axis as a cause of obesity, relatively a few have focused on the opposite relationship, i.e. obesity as a cause of HPA hyperactivity (Foss & Dyrstad 2011). The metabolic changes associated with obesity have the potential to aggravate the HPA axis. Elevations in glucose and cholesterol in obesity are alterations in homeostasis monitored by the hypothalamus and potentially lead to activation of the HPA axis.
The lipocalin-type prostaglandin D2 synthase (L-PGDS) knockout (KO) mouse has been used to examine the role of L-PGDS in obesity and type 2 diabetes. This mouse model demonstrates glucose intolerance, accelerated insulin resistance, aggravated obesity in the context of increased cholesterol, and atherosclerosis (Ragolia et al. 2005, Tanaka et al. 2009). In addition, the L-PGDS KO mouse presents with hypertrophic adipocytes in visceral as well as subcutaneous stores (Ragolia et al. 2005, Tanaka et al. 2009). In human obesity, L-PGDS in cerebral spinal fluid correlates with the orexigenic peptide NPY as well as visceral adiposity and CNS markers of the HPA axis (Elias et al. 2011).
L-PGDS is the principle PGH2 isomerase expressed in the CNS and was originally discovered as beta-trace. L-PGDS is produced in the choroid plexus and leptomeninges (Urade et al. 1993) and has been well studied in association with the regulation of sleep–wake cycle (Qu et al. 2006) and sensitivity to tactile pain (Eguchi et al. 1999). It also plays a protective, anti-inflammatory role in some neurological disorders such as Alzheimer's disease (Kanekiyo et al. 2007) and cerebral ischemia (Saleem et al. 2009). In the periphery, it is found in the vascular endothelium (Taba et al. 2000) as well as adipose tissue (Quinkler et al. 2006, Fujimori et al. 2007). Its role in the vasculature is thought to be a protective one, where its downstream prostaglandin (PGD2) and its derivative 15d-PGJ2 have anti-inflammatory properties (Sasaguri & Miwa 2004). In adipose tissue, L-PGDS and PGD2 are linked to adipogenic differentiation (Fujimori et al. 2007, Fujitani et al. 2010).
In this study, the diet-induced obese C57BL/6 mouse was used to test the hypothesis that chronically elevated metabolic parameters associated with obesity such as cholesterol and glucose can intensify basal HPA axis activity. Because the L-PGDS KO mouse is a model of insulin resistance and obesity, it was further hypothesized that basal HPA axis activity would be greater in this model. The L-PGDS KO and C57BL/6 mice were fed a low-fat (LF) or HF diet initiated at 8 weeks of age. After 20 and 37 weeks of feeding, metabolic and HPA axis hormone measurements were compared between genotypes. Correlation analyses between metabolic parameters and ACTH and corticosterone levels were used to identify associations between obesity and HPA activity.
Materials and methods
Animals
All animal protocols were approved by the Institutions Animal Care and Use Committee and adhere to the regulations outlined by the National Institutes of Health. C57BL/6 mice were purchased from Hilltop Animal Labs, Inc. (Scottsdale, PA, USA). The L-PGDS transgenic KO C57BL/6 mice were originally obtained from the Osaka Bioscience Institute (Osaka, Japan; Eguchi et al. 1999) and are bred in-house. The colony is periodically tested to confirm the mutation using standard genotyping protocols. The animals were housed under local vivarium conditions and fed a standard chow diet until 8 weeks of age when the experimental diets were introduced. Animals were allowed to adapt to local conditions for at least 1 week before experimental procedures. Only male mice of both genotypes were used in the experiments.
At 8 weeks of age, animals were grouped according to genotype and diet and singly housed: for the C57BL/6 mice n=20 per group and for the L-PGDS KO mice n=10 per group. Animals were fed ad libitum with either a HF diet (5.24 kcal/g) providing 60% of the calories as fat (D12492) or a LF diet (3.85 kcal/g) with 10% of the calories as fat (D12450B) (Research Diets, Inc., New Brunswick, NJ, USA). Animals were killed by decapitation between 0900 and 1100 h after 20 or 37 weeks of feeding. Trunk blood was collected in EDTA tubes and plasma was stored at −80 °C until assay. Variations in the final n represent animals excluded due to mortality and/or no plasma collected, not enough plasma for every assay, or significant outlier.
Glucose, triglyceride, and cholesterol measurements
Whole-blood glucose measurements were made at the time of decapitation using an Accu-Chek glucose monitor (Roche) and the corresponding glucose test strips. Total cholesterol was also measured at the time of decapitation using the Cardio-Chek P.A. (Polymer Technology Systems, Inc., Indianapolis, IN, USA) and the corresponding Lipid Panel test strips, which measure cholesterol, HDL-cholesterol, and triglycerides in whole blood. Because some of the whole-blood triglyceride and HDL and measurements were out of the test-strip range, the Max Discovery kits by BIOo Scientific (Austin, TX, USA) were used to measure triglyceride, LDL, and HDL using plasma.
Plasma ACTH and corticosterone measurements
Corticosterone was measured using the DetectX corticosterone enzyme immunoassay kit (Arbor Assays) with a detection limit at 16.9 pg/ml. Plasma ACTH levels were determined using an enzyme immunoassay kit from Phoenix Pharmaceuticals (Burlingame, CA, USA) with a detection range of 0–25 ng/ml.
Statistical analysis
Data were analyzed using three-way fixed effects ANOVA model. In exploring the raw data, several outcome variables violated the assumptions of ANOVA, namely normality. So, we transformed the variables glucose, ACTH, and CORT using Box-Cox transformation. SAS PROC TRANSREG was used to transform the variables and made sure it satisfied the usual assumption of ANOVA. The transformed variables are then used in performing ANOVA. Genotypes, diet, and time were used as the main effects, and all possible two-way and three-way interaction terms between them were considered (Montgomery 2008). SAS PROC MIXED was used to fit the above model that was followed by Tukey–Kramer's multiple comparisons test.
Pearson correlation co-efficient was calculated to assess correlations among the outcome variables that were normally distributed. The Spearman method was used in cases where the data were not normally distributed. Multiplicity adjustments were not made as the correlation analyses were performed as an exploratory nature to examine the trend of the data. All calculations were performed using SAS 9.2 (SAS Institute, Cary, NC, USA); results were considered statistically significant when P<0.05. One significant outlier among the L-PGDS KO mice fed the HF diet for 37 weeks was identified using Grubb's test with a significance level of P<0.05. Data from this animal were excluded from all analyses.
Results
The L-PGDS KO mouse develops features of metabolic syndrome with age independent of diet
As expected, the HF diet increased body weight among both the C57BL/6 mice and the L-PGDS KO mice. Weight gain by the HF-fed L-PGDS KO mice was greater and more rapid when compared with the C57BL/6 mice. The LF-fed L-PGDS KO mice also experienced greater weight gain after 37 weeks of ad libitum feeding when compared with the C57BL/6 mice (Fig. 1A). ANOVA results indicate that genotype, time, and diet had significant individual effects on body weight, P<0.0001. The LF-fed L-PGDS KO group also developed elevated fasting total cholesterol (Fig. 1B) that can be attributed to an elevation in LDL levels (Fig. 1C). After 37 weeks of feeding, the elevations in cholesterol were not significantly different than those observed in the HF-fed L-PGDS KO mice giving rise to a significant interaction effect of genotype×diet×time (F(3,105)=3.30, P=0.0233). Consistent with a metabolic syndrome, levels of HDL were lower in the LF-fed L-PGDS KO mice after 37 weeks of feeding reflected by a significant interaction effect of genotype×diet×time (F(4,97)=3.89, P=0.0057) (Fig. 1D).
Metabolic parameters in C57BL/6 mice and L-PGDS KO mice maintained on LF and HF diets for 20 and 37 weeks. (A) Body weight, (B) Total cholesterol, (C) Plasma LDL, (D) Plasma HDL, (E) Glucose and (F) Triglyceride were measured. Data are presented as mean±s.e.m. and n is indicated for each group within or above the bar. Comparisons between individual groups were made using the Tukey–Kramer adjustment after significant single or interaction effects found through three-way ANOVA. aSignificantly different from C57BL/6 counterpart, bsignificantly different from LF fed of same genotype, csignificantly different from 20 weeks of same genotype and diet.
Citation: Journal of Endocrinology 216, 2; 10.1530/JOE-12-0275
Fasting glucose levels were greater in all L-PGDS KO groups with the exception of the group fed the HF diet for 37 weeks (Fig. 1E). ANOVA results indicated a significant genotype effect on glucose (F(1,110)=7.68, P=0.0066) as well as a significant effect of diet (F(1,110)=31.64, P<0.0001).
In contrast to what was observed with cholesterol levels, fasting triglyceride levels in LF-fed L-PGDS KO mice did not increase after 37 weeks of feeding. In fact, triglyceride levels remained constant among all L-PGDS KO mice irrespective of time and diet where the C57BL/6 mice experienced elevated triglyceride with time on the LF diet as well increases in triglyceride with the HF diet. Three-way ANOVA confirmed a significant interaction effect between genotype and time (F(1,108)=7.44, P=0.0074; Fig. 1F).
L-PGDS KO mice experience early elevations in HPA axis activity relative to C57BL/6 mice but do not experience age-associated changes in ACTH levels
The C57BL/6 mice experience the expected age-associated change in HPA axis activity, i.e. with both the LF and the HF diet, their fasting ACTH levels increased significantly from 20 to 37 weeks of feeding (Fig. 2A). With the LF diet, corticosterone levels do not increase with age as would be expected, reflecting the increase in ACTH. When fed a HF diet, the C57BL/6 mice experience elevations in corticosterone compared with their LF-fed counterparts (Fig. 2B). In the L-PGDS KO mice, ACTH levels were elevated at 20 weeks relative to the C57BL/6 mice but did not increase with age leading to a significant interaction effect of genotype and time (F(1,108)=32.00, P<0.0001; Fig. 2A). Among the L-PGDS KO mice, corticosterone levels were elevated compared with the C57BL/6 mice as well as with HF feeding; however, there was no change with time on either diet. Three-way ANOVA revealed a significant effect of genotype (F(1,106)=9.78, P=0.0023) as well as a significant interaction effect of diet and time (F(1,106)=5.67, P=0.0191) (Fig. 2B).
HPA axis hormone and leptin levels in C57BL/6 and L-PGDS KO mice maintained on LF and HF diets for 20 and 37 weeks. (A) ACTH, (B) Corticosterone, and (C) Leptin were measured. Data are presented as mean±s.e.m. and n is indicated for each group within or above the bar. Comparisons between individual groups were made using the Tukey–Kramer adjustment after significant single or interaction effects found through three-way ANOVA. aSignificantly different from C57BL/6 counterpart, bsignificantly different from LF fed of same genotype, csignificantly different from 20 weeks of same genotype and diet.
Citation: Journal of Endocrinology 216, 2; 10.1530/JOE-12-0275
Because leptin plays a known regulatory role in the HPA axis (Malendowicz et al. 2007), leptin levels were measured and compared between genotypes. As expected and consistent with increases in body weight and metabolic indices, leptin levels increased significantly with HF feeding among both genotypes (Fig. 2C). Also consistent with metabolic indices, leptin levels among LF-fed L-PGDS KO mice were significantly elevated after 37 weeks when compared with their C57BL/6 counterparts. No significant differences in leptin between genotypes were found at 20 weeks under a LF diet or with HF feeding consistent with previous reports (Fig. 2C; Ragolia et al. 2005).
Correlation analyses of metabolic parameters and HPA axis hormones
In order to reveal relationships between elevated metabolic parameters associated with obesity and HPA axis activity, we performed correlation analyses. These analyses were performed by combining LF- and HF-fed groups of the same genotype and comparisons were made with regard to age, i.e. 20 weeks of feeding (28 weeks of age) and 37 weeks of feeding (45 weeks of age). First, to identify relationships between metabolic parameters and the peripheral branch of the HPA axis, correlations with corticosterone were made. Due to the strong in vivo relationship between corticosterone and glucose, the strong positive correlations between these two parameters observed in the C57BL/6 mice were expected. These correlations trended toward positive in the L-PGDS KO mice but were not significant (Fig. 3A and B). In addition, we identified significant correlations between corticosterone and total cholesterol among the C57BL/6 and L-PGD KO mice and this positive correlation strengthened with age (Fig. 3C and D). In line with the correlation results of corticosterone and total cholesterol, significant positive correlations between corticosterone and LDL and corticosterone and HDL were found among the C57BL/6 mice. Among L-PGDS KO mice, positive correlations between corticosterone and LDL and HDL also strengthened with age (Fig. 3E, F, G and H).
Correlation analyses of glucose (A & B), total cholesterol (C & D), and LDL (E & F) and HDL (G & H) with ACTH in C57BL/6 mice and L-PGDS KO mice. Analyses were performed combining LF- and HF-fed measurements after 20 weeks of feeding (28 weeks of age) and after 37 weeks of feeding (45 weeks of age). Gray squares – C57BL/6, black circles – L-PGDS KO.
Citation: Journal of Endocrinology 216, 2; 10.1530/JOE-12-0275
To identify relationships between metabolic parameters and the central branch of the HPA axis, correlations were made with ACTH. If elevated metabolic parameters associated with obesity aggravate the HPA axis then it was expected that positive correlations with ACTH would exist. As with corticosterone in the C57BL/6 mice, ACTH positively correlated with glucose and with total cholesterol, with the strength of the correlations increasing with time (Fig. 4A, B, C and D). Correlations of ACTH and LDL followed a similar trend, although at 45 weeks they did not reach significance (Fig. 4E and F). This may have been due to a few data points that approached the point of significant outlier. By contrast, no significant correlations were found between ACTH and HDL levels in these mice (Fig. 4G and H). Among L-PGDS KO mice, significant positive correlations between ACTH and all metabolic parameters were found at 20 weeks but not after 37 weeks of feeding (Fig. 4A, B, C, D, E, F, G and H).
Correlation analyses of glucose (A & B), total cholesterol (C & D), and LDL (E & F) and HDL (G & H) with corticosterone in C57BL/6 mice and L-PGDS KO mice. Analyses were performed combining LF- and HF-fed measurements after 20 weeks of feeding (28 weeks of age) and after 37 weeks of feeding (45 weeks of age). Gray squares – C57BL/6, black circles – L-PGDS KO.
Citation: Journal of Endocrinology 216, 2; 10.1530/JOE-12-0275
Correlation analysis of ACTH, corticosterone, and leptin
It is generally accepted that ACTH and corticosterone correlate. Our analysis showed a significant correlation only after 37 weeks of feeding among the C57BL/6 mice and no correlation of HPA axis hormones among L-PGDS KO mice (Fig. 5A and B).
Correlation analyses of leptin and the HPA axis hormones in C57BL/6 mice and L-PGDS KO mice. (A & B) ACTH vs corticosterone, (C & D) ACTH vs leptin, (E & F) corticosterone vs leptin. Analyses were performed combining LF- and HF-fed measurements after 20 weeks of feeding (28 weeks of age) and after 37 weeks of feeding (45 weeks of age). Gray squares – C57BL/6, black circles – L-PGDS KO.
Citation: Journal of Endocrinology 216, 2; 10.1530/JOE-12-0275
Correlation analyses between leptin and HPA axis hormones were also performed due to leptin's known regulatory role in this axis. Strong positive correlations were found between leptin and both ACTH and corticosterone among both genotypes with the correlations increasing in significance with length of feeding time (Fig. 5C, D, E and F).
Discussion
The L-PGDS KO mouse maintained on a HF diet is a model of increased cholesterol in the context of obesity (Tanaka et al. 2009). In our study, the L-PGDS KO mice also developed some characteristic features of metabolic syndrome independent of diet. Metabolic syndrome is defined as having three or more disorders relating to metabolism, i.e. obesity, high blood pressure, elevated fasting glucose, and elevated lipids with a decreased HDL. KO mice maintained on the LF diet from 8 weeks until 45 weeks of age experienced increased body weight as well as an elevated total cholesterol that could be attributed to increased LDL and decreased HDL cholesterol when compared with C57BL/6 mice on the same diet. The elevations in total cholesterol and LDL cholesterol are not significantly different from those observed in the C57BL/6 mice or L-PDGDS KO mice fed a HF diet for the same length of time (Fig. 1B and C). Consistent with the increase in body weight gain and metabolic parameters, the LF-fed L-PGDS KO mice also experienced significantly increased leptin levels after 37 weeks.
C57BL/6 mice fed a HF diet are a well-established model of metabolic syndrome and, consistent with this, they experienced an elevation in fasting triglyceride levels when fed a HF diet. Of note is the fact that this increase did not persist after 37 weeks of feeding due to an age-associated increase in fasting triglyceride among mice maintained on a LF diet. These data are consistent with the reports of a shift in fatty acid metabolism toward lipogenesis with aging in mice (Kuhla et al. 2011). It was expected that in L-PGDS KO mice, triglyceride levels in response to HF diet would mirror trends observed with cholesterol measurements as others found a trend toward higher fasting triglyceride levels in L-PGDS KO mice after 12 weeks of HF feeding (Tanaka et al. 2009). However, no significant difference in fasting triglyceride was found in L-PGDS KO mice with 20 weeks of HF feeding. Moreover, unlike the control mice, fasting triglyceride levels did not increase with age and after 37 weeks of feeding were lower than those observed in the control animals under both LF and HF diets giving rise to a significant interaction effect of genotype×time (Fig. 1F). Glucose levels in HF-fed L-PGDS KO mice were also not elevated above LF-fed mice after 37 weeks of feeding. Lower triglyceride levels as well as lower glucose levels in the L-PGDS KO mice could be indicative of increased conversion of glucose to triglyceride along with increased triglyceride storage in adipocytes. Adipocytes found in fat depots of L-PGDS KO mice are significantly larger than those of wild-type mice (Ragolia et al. 2005, Tanaka et al. 2009), and L-PGDS and PGD2 expression positively correlate with adipogenic differentiation (Fujimori et al. 2007, Fujitani et al. 2010). Adipogenesis in developing L-PGDS KO mice is impaired, and therefore they have significantly lower numbers of adipocytes. The increased lipid storage in individual cells may be to compensate for decreased numbers of adipocytes.
As expected, the C57BL/6 mice experienced an increase in basal ACTH with age as well as a concomitant relative hypocorticism; corticosterone levels did not increase to mirror the increase in ACTH levels. These results are consistent with previous experiments examining age-related changes in HPA activity of male C57BL/6 mice and have been explained as a consequence of a relative hyposensitivity of the adrenal with aging in mice (Dalm et al. 2005). By contrast, the L-PGDS KO mice did not experience age-associated changes in basal HPA axis hormones; ACTH did not increase with age under either diet. These data suggest an association of L-PGDS and/or its PGD2 product and stress axis activity with aging. CNS levels of L-PGDS increase with aging in humans (Miwa et al. 2008) and in animal models (Chen et al. 2009). L-PGDS levels also correlate with basal CRH (Elias et al. 2011). Thus, there is the possibility that lack of increased L-PGDS levels with age results in decreased CRH production and therefore ACTH production and release. Further study is needed to test this hypothesis.
Although the L-PGDS KO mice did not experience HPA axis hyperactivity with age, their ACTH and corticosterone levels were increased relative to the C57BL/6 mice at 28 weeks of age after 20 weeks of feeding on either diet (Fig. 2). This is not surprising and many factors could be contributing to the early changes in the stress axis of these animals. Both increased production of IL1β and sleep disturbances are potential contributors to increased HPA axis activity in L-PGDS KO mice. Lack of L-PGDS is associated with increased production of IL1β in L-PGDS KO mice (Tanaka et al. 2009). Increased production of this inflammatory cytokine is in turn associated with increased stimulation of the HPA axis (Hsieh et al. 2010, Sasayama et al. 2011). PGD2 produced from L-PGDS also plays a significant somnogenic role in physiological sleep in rodents and humans (Hayaishi 2002, Jordan et al. 2004, Qu et al. 2006). L-PGDS deficiency in the KO animals would result in lower PGD2 levels and disturbances in sleep. Sleep disturbances can have a significant impact on HPA axis activity resulting in significant elevations in basal HPA hormone levels (Balbo et al. 2010).
Chronic stress and subsequent elevated GC result in the release of glucose from hepatic stores and the redistribution of fat to abdominal depots. Therefore, stress and the associated basal HPA hyperactivity have been implicated as a trigger in the development of obesity and diabetes (Dallman et al. 2004). Our data confirm the strong in vivo relationship between glucose and corticosterone and also support the hypothesis that elevated metabolic indices can intensify HPA axis activity. ACTH levels correlated with glucose and total cholesterol levels, and these correlations also strengthened with time, confirming HPA axis hyperactivity. Therefore, the HPA axis can be aggravated by a HF diet as well as participate in the pathological progression of type 2 diabetes.
The mechanism behind HF diet activation of the HPA axis needs to be explored. It has been established that uncontrolled or poorly controlled diabetes is associated with an upregulated HPA axis (Chan et al. 2001). This increased activity is reportedly related to insulin deficiency as opposed to hyperglycemia; basal ACTH and corticosterone in streptozotocin (STZ)-treated rats are normalized with insulin but not phloridzin (Chan et al. 2005). Hypothalamic insulin resistance occurs with HF feeding and is attributed to the accumulation of lipid in this region (Posey et al. 2009). Therefore, our data are in line with these data suggesting that diet-induced obesity creates an insulin-deficient environment in the brain resulting in a hyperactive HPA axis.
The positive correlation between ACTH and total cholesterol levels in C57BL/6 mice after 20 and 37 weeks of feeding, as well as in the L-PGDS KO mice after 20 weeks of feeding, could possibly be explained by stimulation of the liver-X receptor-α in the pituitary by oxysterol. A high cholesterol diet or the administration of the synthetic liver X receptor agonist increases POMC mRNA in the pituitary and also increases pituitary production of ACTH resulting in elevated plasma ACTH (Matsumoto et al. 2009).
In addition, the impact of HF feeding on ACTH levels must be considered. Glucagon-like peptide-1 (GLP1), released in response to food intake, can increase the activity of the HPA axis in rodents and humans (Gil-Lozano et al. 2010). Therefore, bouts of hyperphagia due to the more palatable food among HF-fed animals may contribute to elevated circulating ACTH levels. However, GLP1-induced increases are transient and ACTH levels return to baseline within 20 min of injection (Gil-Lozano et al. 2010).
The correlation analyses also indicate a positive relationship between elevated total cholesterol and corticosterone. Others have concluded that in addition to de novo synthesis of cholesterol and catabolism of intracellular stored cholesterol, cholesterol derived from plasma lipoproteins is a significant source of the cholesterol needed in adrenal steroidogenesis (Morris & Chaikoff 1959, Hoekstra et al. 2010). Mice fed probucol to deplete plasma lipoproteins experience impaired adrenal steroidogenesis (Hoekstra et al. 2010). These data along with the strong correlation between corticosterone and total cholesterol in our study point to the possibility that an increase in lipoprotein would increase adrenal basal steroidogenesis and secretion.
It is generally expected that ACTH and corticosterone correlate, especially in the presence of an environmental stressor due to the fact that ACTH stimulates adrenal corticosterone release. Here, however, we found a significant correlation only after 37 weeks of feeding in the C57BL/6 mice and no correlation among the L-PGDS KO mice. Additionally, leptin, which plays a regulatory role in the HPA axis in mice, strongly correlated with both ACTH and corticosterone among both genotypes. Centrally, leptin increases ACTH secretion while peripherally it suppresses ACTH-stimulated steroid production (Malendowicz et al. 2007). However, leptin can also raise hepatic 11β-hydroxysteroid dehydrogenase (11β-HSD) type I activity in mice (Liu et al. 2003). Therefore, our data suggest that in mice, obesity-related increases in leptin disrupts the feedback between central and peripheral branches of the HPA axis by increasing ACTH secretion and inhibiting ACTH-stimulated adrenal corticosterone release. However, it can increase corticosterone through sources other than ACTH stimulation. This would explain the absence of the correlation between ACTH and corticosterone, the increases in corticosterone with HF feeding, and the significant correlations of leptin with both ACTH and corticosterone.
Overall, our results demonstrate that changes in metabolic indices associated with obesity promote increased activity in the HPA axis. The L-PGDS KO mouse displays features of the metabolic syndrome in the absence of a HF diet as well as with HF feeding and this correlates with HPA axis hyperactivity at 28 weeks of age. However, unlike the C57BL/6 model that displays age-related increases in HPA activity, the L-PGDS KO mice appear resistant to changes in HPA activity with age and long-term HF feeding indicating that these events are dependent on L-PGDS expression.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by the Medical Education Fund (Winthrop U Hospital) and the National Institutes of Health (K99HL091116).
References
Balbo M, Leproult R & Van Cauter E 2010 Impact of sleep and its disturbances on hypothalamo–pituitary–adrenal axis activity. International Journal of Endocrinology 2010 759234. (doi:10.1155/2010/759234)
Chan O, Chan S, Inouye K, Vranic M & Matthews SG 2001 Molecular regulation of the hypothalamo–pituitary–adrenal axis in streptozotocin-induced diabetes: effects of insulin treatment. Endocrinology 142 4872–4879. (doi:10.1210/en.142.11.4872)
Chan O, Inouye K, Akirav EM, Park E, Riddell MC, Matthews SG & Vranic M 2005 Hyperglycemia does not increase basal hypothalamo–pituitary–adrenal activity in diabetes but it does impair the HPA response to insulin-induced hypoglycemia. American Journal of Physiology. Regulatory, Integrative and Comparative Physiology 289 R235–R246. (doi:10.1152/ajpregu.00674.2004)
Chen CP, Chen RL & Preston JE 2009 Age-related increase of prostaglandin D(2) synthase concentration and glycation in ovine cerebrospinal fluid. Experimental Gerontology 44 639–645. (doi:10.1016/j.exger.2009.07.001)
Dallman MF, Pecoraro N, Akana SF, La Fleur SE, Gomez F, Houshyar H, Bell ME, Bhatnagar S, Laugero KD & Manalo S 2003 Chronic stress and obesity: a new view of "comfort food". PNAS 100 11696–11701. (doi:10.1073/pnas.1934666100)
Dallman MF, la Fleur SE, Pecoraro NC, Gomez F, Houshyar H & Akana SF 2004 Minireview: glucocorticoids – food intake, abdominal obesity, and wealthy nations in 2004. Endocrinology 145 2633–2638. (doi:10.1210/en.2004-0037)
Dalm S, Enthoven L, Meijer OC, van der Mark MH, Karssen AM, de Kloet ER & Oitzl MS 2005 Age-related changes in hypothalamic–pituitary–adrenal axis activity of male C57BL/6J mice. Neuroendocrinology 81 372–380. (doi:10.1159/000089555)
Eguchi N, Minami T, Shirafuji N, Kanaoka Y, Tanaka T, Nagata A, Yoshida N, Urade Y, Ito S & Hayaishi O 1999 Lack of tactile pain (allodynia) in lipocalin-type prostaglandin D synthase-deficient mice. PNAS 96 726–730. (doi:10.1073/pnas.96.2.726)
Elias E, Benrick A, Behre CJ, Ekman R, Zetterberg H, Stenlof K & Wallenius V 2011 Central nervous system lipocalin-type prostaglandin D2-synthase is correlated with orexigenic neuropeptides, visceral adiposity and markers of the hypothalamic–pituitary–adrenal axis in obese humans. Journal of Neuroendocrinology 23 501–507. (doi:10.1111/j.1365-2826.2011.02128.x)
Foss B & Dyrstad SM 2011 Stress in obesity: cause or consequence? Medical Hypotheses 77 7–10. (doi:10.1016/j.mehy.2011.03.011)
Fujimori K, Aritake K & Urade Y 2007 A novel pathway to enhance adipocyte differentiation of 3T3-L1 cells by up-regulation of lipocalin-type prostaglandin D synthase mediated by liver X receptor-activated sterol regulatory element-binding protein-1c. Journal of Biological Chemistry 282 18458–18466. (doi:10.1074/jbc.M701141200)
Fujitani Y, Aritake K, Kanaoka Y, Goto T, Takahashi N, Fujimori K & Kawada T 2010 Pronounced adipogenesis and increased insulin sensitivity caused by overproduction of prostaglandin D2 in vivo. FEBS Journal 277 1410–1419. (doi:10.1111/j.1742-4658.2010.07565.x)
Gil-Lozano M, Perez-Tilve D, Alvarez-Crespo M, Martis A, Fernandez AM, Catalina PA, Gonzalez-Matias LC & Mallo F 2010 GLP-1(7–36)-amide and exendin-4 stimulate the HPA axis in rodents and humans. Endocrinology 151 2629–2640. (doi:10.1210/en.2009-0915)
Hayaishi O 2002 Molecular genetic studies on sleep–wake regulation, with special emphasis on the prostaglandin D(2) system. Journal of Applied Physiology 92 863–868. (doi:10.1152/japplphysiol.00766.2001)
Hoekstra M, Korporaal SJ, Li Z, Zhao Y, Van Eck M & Van Berkel TJ 2010 Plasma lipoproteins are required for both basal and stress-induced adrenal glucocorticoid synthesis and protection against endotoxemia in mice. American Journal of Physiology. Endocrinology and Metabolism 299 E1038–E1043. (doi:10.1152/ajpendo.00431.2010)
Hsieh CH, Li HY & Chen JC 2010 Nitric oxide and interleukin-1β mediate noradrenergic induced corticotrophin-releasing hormone release in organotypic cultures of rat paraventricular nucleus. Neuroscience 165 1191–1202. (doi:10.1016/j.neuroscience.2009.12.003)
Jordan W, Tumani H, Cohrs S, Eggert S, Rodenbeck A, Brunner E, Ruther E & Hajak G 2004 Prostaglandin D synthase (beta-trace) in healthy human sleep. Sleep 27 867–874.
Kanekiyo T, Ban T, Aritake K, Huang ZL, Qu WM, Okazaki I, Mohri I, Murayama S, Ozono K & Taniike M et al. 2007 Lipocalin-type prostaglandin D synthase/beta-trace is a major amyloid β-chaperone in human cerebrospinal fluid. PNAS 104 6412–6417. (doi:10.1073/pnas.0701585104)
Kuhla A, Blei T, Jaster R & Vollmar B 2011 Aging is associated with a shift of fatty metabolism toward lipogenesis. Journals of Gerontology. Series A, Biological Sciences and Medical Sciences 66 1192–1200. (doi:10.1093/gerona/glr124)
Liu Y, Nakagawa Y, Wang Y, Li R, Li X, Ohzeki T & Friedman TC 2003 Leptin activation of corticosterone production in hepatocytes may contribute to the reversal of obesity and hyperglycemia in leptin-deficient ob/ob mice. Diabetes 52 1409–1416. (doi:10.2337/diabetes.52.6.1409)
Malendowicz LK, Rucinski M, Belloni AS, Ziolkowska A & Nussdorfer GG 2007 Leptin and the regulation of the hypothalamic–pituitary–adrenal axis. International Review of Cytology 263 63–102. (doi:10.1016/S0074-7696(07)63002-2)
Matsumoto S, Hashimoto K, Yamada M, Satoh T, Hirato J & Mori M 2009 Liver X receptor-α regulates proopiomelanocortin (POMC) gene transcription in the pituitary. Molecular Endocrinology 23 47–60. (doi:10.1210/me.2007-0533)
Miwa Y, Oda H, Shiina Y, Shikata K, Tsushima M, Nakano S, Maruyama T, Kyotani S, Eguchi N & Urade Y et al. 2008 Association of serum lipocalin-type prostaglandin D synthase levels with subclinical atherosclerosis in untreated asymptomatic subjects. Hypertension Research 31 1931–1939. (doi:10.1291/hypres.31.1931)
Montgomery DC 2008 Design and Analysis of Experiments. 7th Edition, New York: John Wiley and Sons.
Morris MD & Chaikoff IL 1959 The origin of cholesterol in liver, small intestine, adrenal gland, and testis of the rat: dietary versus endogenous contributions. Journal of Biological Chemistry 234 1095–1097.
Posey KA, Clegg DJ, Printz RL, Byun J, Morton GJ, Vivekanandan-Giri A, Pennathur S, Baskin DG, Heinecke JW & Woods SC et al. 2009 Hypothalamic proinflammatory lipid accumulation, inflammation, and insulin resistance in rats fed a high-fat diet. American Journal of Physiology. Endocrinology and Metabolism 296 E1003–E1012. (doi:10.1152/ajpendo.90377.2008)
Qu WM, Huang ZL, Xu XH, Aritake K, Eguchi N, Nambu F, Narumiya S, Urade Y & Hayaishi O 2006 Lipocalin-type prostaglandin D synthase produces prostaglandin D2 involved in regulation of physiological sleep. PNAS 103 17949–17954. (doi:10.1073/pnas.0608581103)
Quinkler M, Bujalska IJ, Tomlinson JW, Smith DM & Stewart PM 2006 Depot-specific prostaglandin synthesis in human adipose tissue: a novel possible mechanism of adipogenesis. Gene 380 137–143. (doi:10.1016/j.gene.2006.05.026)
Ragolia L, Palaia T, Hall CE, Maesaka JK, Eguchi N & Urade Y 2005 Accelerated glucose intolerance, nephropathy, and atherosclerosis in prostaglandin D2 synthase knock-out mice. Journal of Biological Chemistry 280 29946–29955. (doi:10.1074/jbc.M502927200)
Saleem S, Shah ZA, Urade Y & Dore S 2009 Lipocalin-prostaglandin D synthase is a critical beneficial factor in transient and permanent focal cerebral ischemia. Neuroscience 160 248–254. (doi:10.1016/j.neuroscience.2009.02.039)
Sasaguri T & Miwa Y 2004 Prostaglandin J2 family and the cardiovascular system. Current Vascular Pharmacology 2 103–114. (doi:10.2174/1570161043476384)
Sasayama D, Hori H, Iijima Y, Teraishi T, Hattori K, Ota M, Fujii T, Higuchi T, Amano N & Kunugi H 2011 Modulation of cortisol responses to the DEX/CRH test by polymorphisms of the interleukin-1β gene in healthy adults. Behavioral and Brain Functions 7 23. (doi:10.1186/1744-9081-7-23)
Shin AC, MohanKumar SM, Sirivelu MP, Claycombe KJ, Haywood JR, Fink GD & MohanKumar PS 2010 Chronic exposure to a high-fat diet affects stress axis function differentially in diet-induced obese and diet-resistant rats. International Journal of Obesity 34 1218–1226. (doi:10.1038/ijo.2010.34)
Taba Y, Sasaguri T, Miyagi M, Abumiya T, Miwa Y, Ikeda T & Mitsumata M 2000 Fluid shear stress induces lipocalin-type prostaglandin D(2) synthase expression in vascular endothelial cells. Circulation Research 86 967–973. (doi:10.1161/01.RES.86.9.967)
Tanaka R, Miwa Y, Mou K, Tomikawa M, Eguchi N, Urade Y, Takahashi-Yanaga F, Morimoto S, Wake N & Sasaguri T 2009 Knockout of the l-pgds gene aggravates obesity and atherosclerosis in mice. Biochemical and Biophysical Research Communications 378 851–856. (doi:10.1016/j.bbrc.2008.11.152)
Urade Y, Kitahama K, Ohishi H, Kaneko T, Mizuno N & Hayaishi O 1993 Dominant expression of mRNA for prostaglandin D synthase in leptomeninges, choroid plexus, and oligodendrocytes of the adult rat brain. PNAS 90 9070–9074. (doi:10.1073/pnas.90.19.9070)
