Changes in Morphology and Function of Adrenal Cortex in Mice Fed a High-Fat Diet

Home , Adrenal cortex, Adrenal steroid, GLI1

ORIGINAL ARTICLE Changes in morphology and function of adrenal cortex in mice fed a high-fat diet

MM Swierczynska1,2,4, I Mateska1,2, M Peitzsch3, SR Bornstein2, T Chavakis2, G Eisenhofer3, V Lamounier-Zepter2,5 and S Eaton1,5

BACKGROUND/OBJECTIVES: Obesity is a major risk factor for the development of type 2 diabetes and other debilitating diseases. Obesity and diabetes are intimately linked with altered levels of adrenal steroids. Elevated levels of these hormones induce insulin resistance and cause cardiovascular diseases. The mechanisms underlying obesity-related alterations in adrenal steroids are still not well understood. Here, we investigated how diet-induced obesity affects the morphology and function of the mouse adrenal cortex. METHODS: We fed animals either a high-fat diet (HFD) or a normal diet (60% kcal from fat or 10% kcal from fat, respectively) for 18 weeks. We then assessed various aspects of adrenal gland morphology and function, as well as basal plasma concentrations of steroid hormones and ACTH. RESULTS: We show that adrenal glands of mice fed a HFD release more corticosterone and aldosterone, resulting in higher plasma levels. This increase is driven by adrenal cortical hyperplasia, and by increased expression of multiple genes involved in steroidogenesis. We demonstrate that diet-induced obesity elevates Sonic hedgehog signaling in Gli1-positive progenitors, which populate the adrenal capsule and give rise to the steroidogenic cells of the adrenal cortex. Feeding animals with a HFD depletes Gli1-positive progenitors, as the adrenal cortex expands. CONCLUSIONS: This work provides insight into how diet-induced obesity changes the biology of the adrenal gland. The association of these changes with increased Shh signaling suggests possible therapeutic strategies for obesity-related steroid hormone dysfunction. International Journal of Obesity (2015) 39, 321–330; doi:10.1038/ijo.2014.102

INTRODUCTION with corticosterone sufﬁces to increase adiposity, a phenotype 17 Obesity constitutes a major risk factor for the development of type that resolves after cessation of treatment. Thus, abnormal levels 2 diabetes, cardiovascular diseases, non-alcoholic steatohepatitis of adrenal steroids may contribute to the development of obesity and liver failure, certain forms of cancer and sleep-breathing and its comorbidities. However, the mechanisms that underlie disorders.1 It has become an epidemic in western countries, and elevated levels of adrenal steroids in obesity are still not well its prevalence is continuously increasing.2 Thus, obesity is understood. currently one of the most important public health problems. Aldosterone synthesis is controlled mostly by systemically 6 Obesity is associated with alterations in plasma cortisol and circulating angiotensin II. On the other hand, secretion of aldosterone levels.3,4 Cortisol and other glucocorticoids promote glucocorticoids, and to lesser extent aldosterone, is controlled differentiation of pre-adipocytes to adipocytes, and may con- mainly by the hypothalamic–pituitary–adrenal (HPA) axis, and tribute to increased body fat mass. Moreover, they inhibit glucose directly induced by the pituitary hormone ACTH.18,19 Increased uptake by peripheral tissues, and stimulate gluconeogenesis in the steroid hormone levels in obesity result, at least in part, from liver.5 Aldosterone, on the other hand, is the major regulator of hyperactivity of the HPA axis at the central level.20 Obese people blood volume and blood pressure homeostasis.6 Obese people also appear to have higher adrenal sensitivity to ACTH.8 However, tend to have elevated levels of aldosterone and increased rates synthesis of steroid hormones in obese patients might also be of glucocorticoid production and clearance.6–9 Similarly, animal affected by other systemic and paracrine factors.18,19,21 models of obesity have elevated levels of aldosterone and The adrenal gland undergoes remodeling in response to chronic corticosterone, the major glucocorticoid in rodents.10–14 Interest- changes in physiological demand for steroid hormones.22 This ingly, increased levels of cortisol and aldosterone have been process might involve the Gli1-positive progenitor cells that linked to insulin resistance and development of cardiovascular populate the adrenal capsule and can give rise to all steroidogenic – diseases.6,15 A potential causal link between elevated cortisol and lineages.23 26 These Gli1-positive cells respond to Sonic hedgehog aldosterone levels and obesity-related diseases may be inferred (Shh) secreted by cells residing in peripheral cortex. Therefore, it from observations that patients with disorders primarily caused by has been proposed that Shh signaling might have a role in increased secretion of adrenal steroids are more prone to develop adrenocortical maintenance and remodeling in response to obesity and its comorbidities.16 Accordingly, treatment of mice physiological stimuli.27,28

1Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany; 2Department of Internal Medicine III, University Hospital Carl Gustav Carus, Dresden University of Technology, Dresden, Germany and 3Department of Clinical Chemistry and Laboratory Medicine, University Hospital Carl Gustav Carus, Dresden University of Technology, Dresden, Germany. Correspondence: Dr S Eaton, Max Planck Institute of Molecular Cell Biology and Genetics, Pfotenhauerstrasse 108, 01307 Dresden, Germany. E-mail: [email protected] 4Current address: Biozentrum, University of Basel, Basel, Switzerland. 5These authors contributed equally to this work. Received 7 January 2014; revised 9 May 2014; accepted 15 May 2014; accepted article preview online 12 June 2014; advance online publication, 8 July 2014 Adrenal cortex in mice fed a high-fat diet MM Swierczynska et al 322 The ob/ob mouse and the Zucker (fa/fa) rat, which become Biochemical analysis obese due to the loss of leptin signaling, have been used to study Plasma cholesterol, triglycerides and ACTH levels were analyzed using dysregulation of steroid hormones in obesity. In these models, Cholesterol Quantitation Kit (Sigma-Aldrich), EnzyChrom Triglyceride Assay increased plasma steroid hormone levels are associated with Kit (Bioassays, Hayward, CA, USA), and ACTH (Mouse/Rat) ELISA (Abnova increased conversion of 11-dehydrocorticosterone to cortico- GMBH, Heidelberg, Germany) as instructed by the manufacturers. sterone in peripheral tissues, increased sensitivity to angiotensin II, hyperactivity of the HPA axis, enlargement of the adrenal cortex Steroid hormones measurements and increased release of steroid hormones by the adrenal Aldosterone, corticosterone and their precursors were measured in mouse gland.10–12,29,30 However, as leptin directly inhibits secretion of plasma and supernatants of organ cultures by liquid chromatography- steroid hormones by adrenal glands31,32 these monogenic models tandem mass spectrometry. An AB Sciex QTRAP 5500, equipped with may not accurately reflect the changes that occur in diet-induced atmospheric pressure chemical ionization source, coupled to the Acquity obesity. Feeding C57BL/6 mice with a high-fat diet (HFD) has been UPLC system (Waters, Milford, MA, USA) was used. Samples were prepared by offline solid phase extraction supported by positive pressure as shown to accurately reflect various aspects of obesity in 34 14,33 described elsewhere. Analytes were detected by the triple quadrupole humans. However, the effects of diet-induced obesity on mass spectrometer in multiple reaction-monitoring scan mode using adrenal function/morphology have not been studied in detail. positive atmospheric pressure chemical ionization. Analyte concentrations Here, we investigated the effect of diet-induced obesity on were quantified using ratios of analyte peak area to the respective internal various aspects of adrenal morphology and function, including the standard peak area for which deuterated analogs were used. activity of Shh pathway and its impact on Gli1-positive capsular progenitor cells. We show that feeding animals a HFD elevates Immunofluorescence adrenal production of aldosterone, corticosterone and their PFA-fixed glands were cryopreserved in 30% sucrose (AppliChem GMBH, precursors, thereby elevating their levels in plasma. We demon- Darmstadt, Germany)/phosphate-buffered saline overnight at 4 °C, strate that increased steroidogenic output in obese animals results embedded in TFM medium (Triangle Biomedical Sciences, Durham, NC, from adrenal cortex hyperplasia, and from increased expression of USA), and frozen at − 80 °C. Each adrenal was cut into 8-μm-thick serial genes controlling various aspects of steroidogenesis. Finally, we sections. Special care was taken to process every gland in a uniform show that the expansion of the adrenal gland in obese animals is fashion. accompanied by depletion of Gli1-positive capsular progenitor Adrenal cryosections were incubated overnight at 4 °C with appropriate cells, with a concomitant upregulation of the Shh signaling primary antibodies, washed and incubated for 1 h at room temperature pathway in individual cells. Our work provides insight into how with the respective secondary antibody together with DAPI (1:5000; Roche, Penzberg, Germany). Antibodies used are listed in Supplementary Table 1. diet-induced obesity changes the biology of the adrenal gland, and suggests that Shh pathway activity might contribute to these changes. Image acquisition and analysis Z-series images were acquired with Plan Apochromat × 10/0.45 M27 or Plan Apochromat × 20/0.8 objectives using LSM700 inverted confocal MATERIALS AND METHODS microscope (Zeiss, Jena, Germany). Laser power, photomultiplier gain and Animals pinhole size were set for each antibody individually, and kept constant for all subsequent image acquisitions. Animal work was approved by the Ethical Committee of the Land- For analysis, we chose sections located in similar regions of each adrenal esdirektion Dresden. C57BL/6J mice were from Harlan Laboratories (Horst, gland, based on position in the series of collected sections. Analysis of Netherlands) or in-house husbandry of the Biomedical Services Facility LacZ tm2Alj GFP capsule thickness, cortex and medulla areas, and anti-GFP staining (BMS; Dresden, Germany). Heterozygous Gli (Gli1 /J) and Shh 35 tm6Amc intensity were done with Fiji software . Nuclear densities and numbers, (Shh /J) mice were obtained from the Jackson Laboratory (Bar Cyp11B2-positive area and Gli1-positive cells were quantified using Harbor, ME, USA) and were back-crossed to a C57BL/6J background at the CellProfiler software36 (analysis pipelines are available upon request). BMS for three generations. Eight-week-old mice were singly caged, and In analyses based on StAR, Cyp11B2, GFP and DAPI, a z-projection of the allowed to feed ad libitum on a HFD or a normal diet (ND) (60% kcal from maximum intensity from the acquired z-stacks was performed for each fat or 10% kcal from fat, respectively; Research Diets Inc., New Brunswick, image. Gli1-positive cells were enumerated in every third stack, to avoid NJ, USA) for the period of 18 weeks. Their body weight was recorded double counting of individual cells. Quantifications were performed from weekly. Animals were always killed at around 0900 hours. After killing in adrenals of eight (capsule thickness/cortex area/medulla area), six (nuclear CO2, animal blood was collected into EDTA tubes, and plasma was density), four (Gli1-positive cell numbers, Cyp11B2-positive area) or three separated by centrifugation 1500 g (20 min, 4 °C), and subsequently snap- (anti-GFP staining intensity) mice per group. frozen in liquid nitrogen until further analysis. Adrenal glands were excised, cleaned from the surrounding fat tissue and weighed. Subsequently, they were used for organ cultures, or preserved in RNAlater (Qiagen, Hilden, Quantitative real-time PCR Germany) for RNA extraction, or fixed for 30 min in 4% paraformaldehyde Adrenal RNA extraction with RNeasy Mini Kit (Qiagen), and subsequent (PFA; Sigma-Aldrich, St Louis, MO, USA)/phosphate-buffered saline (PBS) cDNA synthesis with M-MLV Reverse Transcriptase (Promega GMBH, for immunofluorescence. For each type of analysis, equal numbers of left Mannheim, Germany) were performed according to the manufacturers’ and right adrenals were used. instructions. The expression of all genes was determined using ABsolute QPCR SYBR Green Mix (Thermo Fisher Scientific, Waltham, MA, USA), and Organ cultures the Mx 3000P system (Agilent Technologies, Santa Clara, CA, USA). Primer sequences used and amplification protocols are listed in Supplementary Adrenal glands were collected in phosphate-buffered saline + 200 U ml − 1 Table 2. Gene expression was normalized to β-actin expression. penicillin and 200 mg ml − 1 streptomycin (Life Technologies, Carlsbad, CA, USA). Glands were halved along the shorter axis, halves were placed in separate wells in a 24-well plate, in 1 ml of DMEM/F-12 medium (Life Statistical analysis Technologies)+1% insulin–transferrin–selenium mixture (ITS-X; Life Changes in body mass, nuclear number and nuclear density were analyzed Technologies)+2.5% Nu-Serum (BD Biosciences, Franklin Lakes, NJ, USA), using two-way ANOVA. To assess the correlation between adrenal weight and incubated for 1 h at 37 °C in 5% CO2/95% air. Then, media were and body mass Spearman’s rank correlation coefficient was calculated. exchanged to DMEM/F-12 medium+1% ITS-X, and glands were incubated Otherwise, data were analyzed with the independent Mann–Whitney for further 3 h. Media were collected and frozen in − 20 °C until further U-test. Statistical analysis was performed with Graphpad Prism software analysis. (GraphPad Software Inc., La Jolla, CA, USA).

International Journal of Obesity (2015) 321 – 330 © 2015 Macmillan Publishers Limited Adrenal cortex in mice fed a high-fat diet MM Swierczynska et al 323 RESULTS compared with ND counterparts (Figure 2d). The reduction of Mice fed a HFD gained more weight compared with mice fed a capsule thickness in HFD animals was further confirmed by ND, starting from the 1st week of the experiment. After 18 weeks the analogous analysis of cryosections stained with anti-SF-1 on a diet, HFD mice were ~ 30% heavier than ND animals antibody, a well-established nuclear marker of steroidogenic cells (Supplementary Figure 1A), and had markedly elevated circulating (Supplementary Figure 2). In contrast, the cortex of animals fed a cholesterol and triglyceride levels (Supplementary Figures 1B and C). HFD was 37% larger than in animals fed a ND (Figure 2e). There was no difference in medulla size between the two groups of animals (Figure 2f). Thus, the enlargement of adrenal gland in High-fat diet upregulates adrenal steroidogenesis obese animals is caused solely by the expansion of the cortex. fi fi We rst quanti ed the effect of HFD on circulating steroid The adult mouse adrenal cortex is divided into two distinct hormone levels. We measured not only corticosterone and zones. The outer zona glomerulosa (ZG) produces mainly aldosterone but also their precursors pregnenolone and proges- aldosterone, and is characterized by the expression of aldosterone fi terone; 11-deoxycorticosterone was not quanti ed because its synthase Cyp11B2. The inner zona fasciculata (ZF) produces mainly fi levels were below the lower limit of quanti cation of the protocol. corticosterone, and does not express Cyp11B2. To test whether Mice fed HFD had increased morning (basal) levels of all measured the enlargement of adrenal cortex in animals fed a HFD resulted hormones; median pregnenolone concentrations were increased from the expansion of ZG, ZF or both, we stained adrenal by 59%, progesterone by 25%, corticosterone by 320% and cryosections with antibody against Cyp11B2. Quantifying the area aldosterone by 100% (Figure 1a). This suggests that HFD positive for Cyp11B2 (Figures 3a and b) revealed that the size of upregulates steroidogenesis. Individual measurements are listed the ZG was the same in ND and HFD adrenal glands (Figure 3c). in Supplementary Table 3. Thus, enlargement of the adrenal cortex on a HFD is due entirely Plasma hormone levels reflects not only the rate of synthesis to the expansion of corticosterone-producing ZF. but also peripheral binding and degradation, all of which can be The expansion of the ZF and adrenal cortex in obese animals affected by a great number of factors. Moreover, pregnenolone could be explained either by an increase in cell number and progesterone can be synthesized by multiple tissues. To (hyperplasia), cell size (hypertrophy) or both. To distinguish these specifically examine adrenal steroidogenesis, we quantified possibilities we quantified the number of nuclei and their packing steroid hormone released into supernatants by adrenal glands density in the outer region of the adrenal gland, spanning 200 μm explanted from HFD and ND animals in in vitro cultures. Glands from the border toward the inside of the gland (Figure 3d). isolated from obese animals released at least 58% more 11- This region includes the capsule, ZG and the larger portion deoxycorticosterone, corticosterone and aldosterone to the media of ZF. We assumed that cell hypertrophy would manifest than glands of control animals (Figure 1b). Pregnenolone was not as a decrease in packing density of the nuclei. The enlarged quantified because its levels were below the lower limit of adrenal cortex from animals fed a HFD contained more quantification. Thus, the elevated plasma corticosterone and nuclei than cortex from ND animals (Figure 3e). However, there aldosterone levels in obese animals result, at least in part, from was no difference in the density of the nuclei in the adrenal increased production by the adrenal gland. On the other hand, cortex of HFD and ND mice (Figure 3f). Thus, the increased size elevated levels of progesterone cannot be accounted for by of the adrenal cortex in HFD animals results from cortical increased adrenal production. HFD may affect progesterone hyperplasia. production and metabolism by other tissues. ACTH levels were not different between the HFD and ND mice Diet-induced obesity is associated with changes in expression of (Figure 1c). Thus, increased ACTH production may not account for multiple genes involved in steroidogenesis the elevated plasma steroids in HFD animals. We wondered whether feeding animals the HFD might increase not only the size of the adrenal cortex but also its biosynthetic High-fat diet induces morphological changes of adrenal gland capacity. To address this, we used qPCR to measure levels of We investigated whether increased levels of plasma steroids mRNA for a variety of proteins involved in steroidogenesis, might reflect enlargement of the adrenal gland. Indeed, mice fed normalizing them to mRNA for actin. A complication inherent in HFD had heavier adrenal glands compared with ND mice this approach is that the cortex accounts for a larger fraction of (Figure 2a). Interestingly, while adrenal mass did not correlate the adrenal in HFD animals. Thus, even if cortical cells did not with body weight in ND mice (rs = 0.09271; P = 0.6815), it was increase transcription of a particular gene, its transcript would significantly correlated in mice fed HFD (rs = 0.4365; P = 0.0329), appear to be over-represented in total adrenal mRNA of HFD mice. even though the variations in body mass were comparable in both To control for this, we used our measurements of capsule, cortex groups of animals. Adrenal mass in HFD animals correlated even and medullar thickness to estimate their relative contributions to more strongly with weight gain (rs = 0.5286; P = 0.0079), although the volume of the adrenal on normal and -fat diets, approximating these parameters remained uncorrelated in normally fed animals the adrenal as a sphere. The predicted increase in adrenal volume (rs = − 0.02235; P = 0.9214). Thus, adrenal enlargement due to diet- based on these thickness measurements corresponds precisely to induced obesity cannot be explained by a general compensation the observed increase in adrenal mass on a HFD (1.34-fold). These mechanism that maintains proportionality between the adrenal calculations showed that the expansion of the cortex alone would and body mass. Rather, it is a specific consequence of a HFD that account for an ~ 8% increase in relative expression of cortex- strongly correlates with the development of obesity. specific genes (Supplementary Figure 3). qPCR showed that Enlargement of the adrenal might result from changes in the median StAR mRNA levels increased ~ 200% in adrenals of HFD relative sizes of the adrenal capsule, cortex and medulla. There- mice (Figure 4a). Median mRNA levels of enzymes responsible for fore, we compared the sizes of each zone in animals fed HFD and pregnenolone, progesterone, 11-deoxycorticosterone and corti- ND. We stained adrenal cryosections with the nuclear stain DAPI, costerone synthesis (CYP11A, 3β-hydroxysteroid dehydrogenase, and with antibody against StAR, a marker of steroidogenic cells. CYP21 and CYP11B1, respectively) were all significantly elevated by The StAR-negative region in the gland periphery was defined as 440% in obese animals (Figure 4b). Although the ZG represents a the adrenal capsule. We assumed that the StAR-positive area smaller fraction of the adrenal in mice fed a HFD, relative levels of corresponds to the adrenal cortex, while the StAR-negative central CYP11B2 mRNA (normalized to actin) were comparable in HFD and region of the gland corresponds to the medulla (Figures 2b and c). ND animals (Figure 4b). The fact that we did not observe any The adrenal capsule from HFD animals was ~ 20% thinner difference in normalized Cyp11B2 mRNA levels between ND and

Figure 1. High-fat diet elevates basal levels of circulating steroid hormones and adrenal steroid production without affecting basal plasma ACTH levels. (a) Morning (basal) plasma concentrations of indicated steroid hormones. (b) Levels of indicated steroid hormones released by explanted adrenals in in vitro cultures. Each point represents the average measurement from both halves of the same adrenal gland. (c) Basal levels of plasma ACTH. Black line indicates median. *P ⩽ 0.05; **P ⩽ 0.01; ***P ⩽ 0.001 (Mann–Whitney U-test).

HFD adrenals suggests that Cyp11B2 expression is upregulated in increases the size of the cortex but also increases the expression the ZG cells. No signiﬁcant increase was observed in the level of of several steroidogenic enzymes. These transcriptional changes SF-1 mRNA, which is expressed in both ZG and ZF cells (Figure 4c). may further contribute to increased secretion of steroid hormones These data show that feeding animals with a HFD not only from adrenals of HFD mice.

Figure 2. Mice fed a HFD have bigger adrenal glands due to enlargement of adrenal cortex. (a) Total adrenal mass (left+right adrenal). (b, c) Show adrenal glands stained with anti-StAR antibody (green) and DAPI (magenta) from mice fed (b) ND and (c) HFD. (Insets) High magniﬁcation of the capsule area. Scale bars indicate 100 μm. (d–f) Sizes of adrenal zones: (d) capsule thickness, (e) cortex area and (f) medulla area. Black lines indicate the median, and signiﬁcance was calculated using Mann–Whitney U-test. *P ⩽ 0.05; **P ⩽ 0.01.

Figure 3. Mice fed a HFD have bigger adrenal glands due to adrenal cortex hyperplasia. (a, b) show adrenal glands stained for Cyp11B2 from mice fed (a) ND and (b) HFD. (Insets) High magnification of the ZG area. Scale bars indicate 100 μm. The signal adjacent to the medulla is caused by non-specific secondary antibody binding. (c) Quantification of Cyp11B2-positive areas in animals fed ND and HFD. (d) Cartoon depicting the basis of nuclear number and nuclear density analysis. Based on nuclear DAPI staining, CellProfiler software divided the outer region of the adrenal gland into 20 doughnut-shape zones, each 10 μm thick. These zones cover most of the adrenal cortex area. (e) Quantification of the number of nuclei in the region of the adrenal cortex indicated in (d). (f) Quantification of nuclear packing density in the region of the adrenal cortex indicated in (d). In (c) black lines indicate the median, and significance was calculated using Mann–Whitney U-test. In (e, f), n = 6, error bars indicate ± s.e.m., and significance was calculated using two-way ANOVA. ****P ⩽ 0.0001.

Figure 4. High-fat diet induces changes in expression of multiple genes involved in steroidogenesis. (a–g) Real-time qPCR quantiﬁcation of mRNA expression of various genes affecting the synthesis of steroid hormones in adrenal glands: (a) StAR, (b) steroidogenic enzymes, (c) SF-1, (d) HMG-CoA reductase, (e) lipoprotein receptors, (f) AT1R–angiotensin 2 receptor and (g) MCR2–ACTH receptor. Normalizing gene: β-actin. Black lines represent median. *P ⩽ 0.05; **P ⩽ 0.01 (Mann–Whitney U–test).

Figure 5. High-fat diet depletes Shh-responsive adrenal progenitor cells, but increases Shh pathway activity in individual cells. (a, b) Anti-beta- galactosidase staining of adrenal glands isolated from heterozygous Gli1LacZ mice fed (a) ND and (b) HFD. (Insets) High magnification of the capsule area. (c) Quantification of Gli1-positive cells in adrenals isolated from heterozygous Gli1LacZ mice on a ND or a HFD. (d) Sketch of Shh signaling. In the absence of Shh, its receptor Patched (Ptch1) inhibits activity of the seven-pass transmembrane protein Smoothened (Smo). The transcription of Shh-target genes is off. Binding of Shh to Ptch1 alleviates Smo repression and activates a downstream signaling cascade, which results in transcription of target genes. These include the transcription factor Gli1, Ptch1 itself, and Hedgehog Interacting Protein (HHIP). (e) Real-time qPCR quantification of mRNA expression of Shh target genes in adrenal glands, normalized to β-actin. Relative expression of Shh target genes in the whole adrenal does not change in proportion to the reduced number of Gli1-positive/Shh-responsive cells. Taking into account the reduced number of Gli1-positive cells, expression of these targets is increased in the HFD group. (f, g) Anti-GFP staining of adrenal glands isolated from heterozygous ShhGFP mice fed (f) ND and (g) HFD. (Insets) High magnification of the capsule and cortex area. (h) Quantification of anti-GFP staining intensity in adrenal glands isolated from ShhGFP animals fed a ND and a HFD. Staining intensity was quantified starting from the edge of the gland, 156 μm towards the medulla. This region covers the adrenal capsule and most of the cortex. (i) Real-time qPCR quantification of Shh mRNA expression in adrenal glands of mice fed a ND and a HFD. In (a, b) and (f, g) scale bars indicate 100 μm. In both types of staining bright spots adjacent to the medulla reflect non-specific secondary antibody binding. In (c, e, i) black lines indicate the median, and significance was calculated using Mann–Whitney U-test. In (h), n = 3, thin dotted lines indicate ± s.e.m. *P ⩽ 0.05.

International Journal of Obesity (2015) 321 – 330 © 2015 Macmillan Publishers Limited Adrenal cortex in mice fed a high-fat diet MM Swierczynska et al 327 Steroid hormones are synthesized from cholesterol obtained of mRNA for 3-hydroxy-3-methyl-glutaryl-CoA (HMG-CoA) reductase, either by de novo synthesis, or taken up from the circulating a rate-limiting enzyme in the cholesterol biosynthesis, were lipoprotein particles. Availability of cholesterol strongly influences comparable in animals fed ND and HFD (Figure 4d). Moreover, the rate of steroidogenesis. Therefore, we investigated whether there were no significant changes in mRNA levels of LDL receptor the adrenal glands of obese animals might increase cholesterol and SR-B1 (Figure 4e), the major receptors responsible for adrenal production and/or the efficiency of lipoprotein uptake. The levels lipoprotein uptake. Thus, the adrenal cortex of HFD mice does not

© 2015 Macmillan Publishers Limited International Journal of Obesity (2015) 321 – 330 Adrenal cortex in mice fed a high-fat diet MM Swierczynska et al 328 appear to increase the efficiency of lipoprotein uptake or DISCUSSION cholesterol biosynthesis, at least through transcriptional We have shown that a HFD causes hyperplasia of the adrenal mechanisms. cortex. Although homeostatic mechanisms may exist to maintain To investigate whether feeding animals HFD might increase proportionality between adrenal size and body mass, these cannot adrenal sensitivity to angiotensin II or ACTH, we looked for account for adrenal enlargement on a HFD. We have shown that changes in expression of their cognate receptors AT1R and MCR2, adrenal size correlates significantly with the body mass only in respectively. The expression of AT1R was reduced by 30% in HFD mice fed a HFD, but not in animals fed a ND. Thus, adrenal mice (Figure 4f), which possibly reflects the reduced proportion of enlargement in obese animals cannot be explained as a ZG to the whole gland in these animals. However, levels of MCR2 compensatory mechanism that corrects for the increased body mRNA did increase by 40% in the adrenals of HFD mice size. Rather, it appears to be directly related to the development of (Figure 4g). These observations suggest that HFD may increase obesity. the sensitivity of the adrenal to ACTH, but not to angiotensin. Adrenal hyperplasia occurs when glands are chronically stimulated with ACTH.37 Although we did not observe differences High-fat diet affects capsular Gli1-positive progenitor cells in basal ACTH levels between ND and HFD animals, we cannot The adrenal capsule is populated by Gli1-positive progenitor cells exclude that adrenal enlargement in HFD animals results from that give rise to all steroidogenic lineages. To investigate how HPA axis hyper-activation and higher 24-hour integrated ACTH these cells were affected by feeding with a HFD, we first compared levels. Abdominal obesity is accompanied by the dissociation of 8 the number of Gli1-positive in the capsules of mice fed different ACTH and glucocorticoid pulsatility. Therefore, HFD mice might diets, using heterozygous Gli1LacZ mice, which express beta- have increased ACTH production during a different time of the galactosidase under the control of Gli1 promoter (Figures 4a and b). day, that is, in the evening when the levels of ACTH are the 38 These GliLacZ animals did not differ from their wild-type counter- highest in rodents. We also observed a significant upregulation parts in body weight, adrenal mass or plasma parameters of the ACTH receptor MCR2 mRNA in HFD animals. This might (Supplementary Figure 4). Interestingly, despite the larger size of suggest increased HPA axis activity, as MCR2 is a direct target of 39 their adrenals, HFD mice had 39% fewer Gli1-positive cells per ACTH signaling. adrenal section compared with ND mice (Figure 5c). Taking into What might be the other factors that promote adrenal consideration the total surface area of the capsule, this reduction hyperplasia upon feeding a HFD? Homeostatic maintenance of must be even stronger. Gli1-positive cells in the adrenal capsule the adrenal cortex, and its remodeling in response to various have been shown to give rise to columns of differentiated ZG and environmental stimuli occur due to proliferation of cells in the 40 ZF cells in the adrenal cortex. This may suggest that cortex periphery of the gland—in the capsule or subcapsular region. enlargement in HFD mice depletes Gli1-positive progenitors. Gli1-positive/Shh-responsive cells that reside in the capsule 24,26 Gli1, Ptch1 and HHIP are well-established direct transcriptional constitute one of the adrenal progenitor cells populations. targets of canonical Shh signaling (Figure 5d). To investigate Interestingly, we show that HFD causes a depletion of Gli1-positive whether depletion of Gli1-positive progenitors was associated progenitors, but increases Shh pathway activity in these cells. with changes in Shh signaling, we performed qPCR on whole Thus, it is possible that increased Shh signaling has a role in HFD- adrenal mRNA to compare the expression of these targets induced adrenal hyperplasia. The extent to which Gli1-positive in HFD and control animals. HFD adrenals are enlarged 1.34-fold, progenitors divide symmetrically or asymmetrically to produce but contain o61% of the Shh-responsive cells found in ND mice steroidogenic cells has never been studied. The depletion of Gli1- (Figure 5d). Thus, if expression of Shh target genes in individual positive capsular cells during cortical hyperplasia may suggest that cells does not change on the HFD, their expression normalized to such a balance between symmetric and asymmetric divisions is actin expression would be o46% (0.61/1.34) of the value in ND altered in animals fed a HFD. animals. However, we observed no significant decrease in either It would be extremely interesting to know what causes Shh Gli1 or HHIP expression in HFD mice, and median Ptch1 expression pathway activation in adrenal glands of obese animals. Our data decreased only by ~ 30% (Figure 5e). This suggests that, although suggest that increased production of Shh ligand by subcapsular Shh-responsive cells in the adrenal capsule are fewer in number in cells does not account for this. However, disturbances in HFD mice, the activity of the Shh pathway in individual cell lipoprotein levels, which are associated with obesity,41 might increases. To determine whether increase in Shh pathway activity affect the form in which Shh is secreted by the cortical cells. in capsular cells resulted from elevated Shh expression, we We have previously shown that Shh can be secreted from performed qPCR to quantify relative Shh mRNA levels and mammalian cells in two distinct forms, as a cholesterol-free quantified the anti-GFP staining in heterozygous ShhGFP mice monomer and the cholesterol-modified lipoprotein-associated fed different diets. These mice express a ShhGFP fusion protein form. Each of these forms possesses distinct signaling properties.42 from the endogenous locus. Although ShhGFP mice were smaller Thus, it is possible that increased Shh pathway activity in the compared with their WT counterparts, they responded to a HFD in capsular cells results from the shift in the form of secreted Shh. comparable fashion (Supplementary Figure 5). Albeit the peak Interestingly, we have also demonstrated that Shh circulates anti-GFP staining intensity was modestly reduced in the cortex of systemically on lipoproteins in mammals,42 although how this pool animals fed a HFD (Figure 5h), we observe no difference in the size might change in obesity and contribute to the signaling in adrenal of the GFP-positive region between ND and HFD adrenals (Figures capsule is not known. Furthermore, it has been demonstrated that 5f and g). Thus, although HFD increases the size of the adrenal lipoprotein particles, oxysterols, vitamin D and endocannabinoids – cortex, it does not increase the size of the Shh-expressing region. directly affect Shh pathway activity.43 46 Thus, dramatic changes in Therefore, Shh-expressing cells comprise a smaller fraction of HFD lipid metabolism associated with obesity might directly increase Shh adrenals. Consistent with this, qPCR analysis shows a propor- pathway activity in the capsule cells. Endocannabinoids are tionate reduction in relative Shh mRNA levels in adrenals of WT particularly interesting in this respect because they affect the animals fed a HFD (Figure 5i). These data suggest that HFD has pathwaybothpositivelyandnegatively,46 and the endocannabinoid little effect on the size of the Shh-expressing region, or on the system is known to be involved in the development of obesity.47 amount of Shh expressed in this region. Thus, upregulation of the Finally, we cannot exclude that the increased expression of some Shh-signaling pathway in the capsular progenitor cells does not Shh-target genes in HFD adrenals results from changes in result from the increased Shh expression in the adjacent non-canonical signaling. Recently, it has been demonstrated that cortical cells. Delta-like homolog 1 (Dlk1) directly affects Gli1 transcription in the

International Journal of Obesity (2015) 321 – 330 © 2015 Macmillan Publishers Limited Adrenal cortex in mice fed a high-fat diet MM Swierczynska et al 329 28 rat adrenal capsule via activation of Erk1/2 kinases. It would be 12 Naeser P. Function of the adrenal cortex in obese-hyperglycemic mice (gene very interesting to know whether this pathway is affected in obese symbol ob). Diabetologia 1974; 10: 449–453. animals. 13 Northcott CA, Fink GD, Garver H, Haywood JR, Laimon-Thomson EL, McClain JL We have shown that diet-induced obesity not only induces et al. The development of hypertension and hyperaldosteronism in a rodent 153 – hyperplasia but also increases transcription of genes important for model of life-long obesity. Endocrinology 2012; : 1764 1773. steroidogenesis. It has been shown that transcription of steroido- 14 Fraulob JC, Ogg-Diamantino R, Fernandes-Santos C, Aguila MB, 48 Mandarim-de-Lacerda CA. A mouse model of metabolic syndrome: insulin genic genes depends on SF-1, and can be upregulated by ACTH. resistance, fatty liver and non-alcoholic fatty pancreas disease (NAFPD) in C57BL/6 However, the transcriptional changes seen on a HFD appear to be mice fed a high fat diet. J Clin Biochem Nutr 2010; 46:212–223. partially independent of these mechanisms. A number of factors 15 Roberge C, Carpentier AC, Langlois MF, Baillargeon JP, Ardilouze JL, Maheux P are known to produce similar transcriptional changes in tissue et al. Adrenocortical dysregulation as a major player in insulin resistance and culture cells. For example, it has been shown that adipocytes onset of obesity. Am J Physiol Endocrinol Metab 2007; 293: E1465–E1478. secrete factors that stimulate steroidogenesis in NCI-H295R 16 Thomson SP, Stump CS, Kurukulasuriya LR, Sowers JR. Adrenal steroids and the cells.49–51 These factors, including Wnt-proteins, positively affect metabolic syndrome. Curr Hypertens Rep 2007; 9:512–519. steroidogenesis by upregulating StAR expression.50 IGF-1 also 17 Cassano AE, White JR, Penraat KA, Wilson CD, Rasmussen S, Karatsoreos IN. 52 Anatomic, hematologic, and biochemical features of C57BL/6NCrl mice upregulates StAR expression in NCI-H295R cells, and its levels 62 – increase in obesity.53 Our observations now provide a system to maintained on chronic oral corticosterone. Comp Med 2012; : 348 360. 18 Bornstein SR, Engeland WC, Ehrhart-Bornstein M, Herman JP. Dissociation of ACTH test the importance of these factors in vivo in diet-induced obesity. and glucocorticoids. Trends Endocrinol Metab 2008; 19:175–180. Furthermore, while it is clear that central dysregulation of the HPA 19 Willenberg HS, Schinner S, Ansurudeen I. New mechanisms to control aldosterone axis and perturbed peripheral metabolism contribute to elevated synthesis. Horm Metab Res Metab 2008; 40:435–441. levels of adrenal steroids in obesity, our work highlights 20 Wang M. The role of glucocorticoid action in the pathophysiology of the the importance of the local changes in adrenal gland function. Metabolic Syndrome. Nutr Metab 2005; 2:3. A deeper understanding of mechanisms governing these local 21 Lamounier-Zepter V, Ehrhart-Bornstein M, Bornstein SR. Metabolic syndrome and processes will allow for development of better therapeutic the endocrine stress system. Horm Metab Res 2006; 38: 437–441. strategies for obesity-related steroid hormone dysfunction. 22 Wolkersdorfer GW, Bornstein SR. Tissue remodelling in the adrenal gland. Biochem Pharmacol 1998; 56: 163–171. 23 Huang CC, Miyagawa S, Matsumaru D, Parker KL, Yao HH. Progenitor cell CONFLICT OF INTEREST expansion and organ size of mouse adrenal is regulated by sonic hedgehog. Endocrinology 2010; 151: 1119–1128. The authors declare no conflict of interest. 24 King P, Paul A, Laufer E. Shh signaling regulates adrenocortical development and identifies progenitors of steroidogenic lineages. Proc Natl Acad Sci USA 2009; 106: – ACKNOWLEDGEMENTS 21185 21190. 25 Ching S, Vilain E. Targeted disruption of Sonic Hedgehog in the mouse adrenal We gratefully acknowledge Celso Gomez-Sanchez, Ken Morohashi and Peter King for leads to adrenocortical hypoplasia. Genesis 2009; 47:628–637. providing antibodies. We are grateful to Marc Bickle from the Technology 26 Wood MA, Acharya A, Finco I, Swonger JM, Elston MJ, Tallquist MD et al. Development Studio (MPI-CBG), for designing and implementing pipelines for the Fetal adrenal capsular cells serve as progenitor cells for steroidogenic and image analysis. We thank Julia Jarrels and Britta Schilling from the DNA Microarray stromal adrenocortical cell lineages in M. musculus. Development 2013; 140: Facility (MPI-CBG) for their assistance in real-time quantitative PCR analyses. We 4522–4532. acknowledge Biomedical Services (MPI-CBG) for their help in animal husbandry and 27 Laufer E, Kesper D, Vortkamp A, King P. Sonic hedgehog signaling during adrenal handling. We are grateful to Susann Malik for technical assistance. This work was development. Mol Cell Endocrinol 2012; 351:19–27. supported by DFG grant, KFO 252. 28 Guasti L, Cavlan D, Cogger K, Banu Z, Shakur A, Latif S et al. Dlk1 upregulates Gli1 expression in male rat adrenal capsule cells through the activation of beta1 integrin and ERK1-2. Endocrinology 2013; 154: 4675–4684. REFERENCES 29 McGinnis R, Walker J, Margules D, Aird F, Redei E. Dysregulation of the 1 Kopelman PG. Obesity as a medical problem. Nature 2000; 404:635–643. hypothalamus-pituitary-adrenal axis in male and female, genetically obese (ob/ 2 King B, Jiang Y, Su X, Xu J, Xie L, Standard J et al. Weight control, endocrine ob) mice. J Neuroendocrinol 1992; 4:765–771. hormones and cancer prevention. Exp Biol Med (Maywood) 2013; 238: 30 Naeser P. In vitro release of corticosteroids from adrenal glands of obese- 502–508. hyperglycemic mice (gene symbol ob). Acta Physiol Scand 1974; 92:175–180. 3 Krug AW, Ehrhart-Bornstein M. Adrenocortical dysfunction in obesity and the 31 Bornstein SR, Uhlmann K, Haidan A, Ehrhart-Bornstein M, Scherbaum WA. metabolic syndrome. Horm Metab Res 2008; 40: 515–517. Evidence for a novel peripheral action of leptin as a metabolic signal to the adrenal 4 Pasquali R, Vicennati V, Gambineri A. Adrenal and gonadal function in obesity. gland: leptin inhibits cortisol release directly. Diabetes 1997; 46: 1235–1238. J Endocrinol Invest 2002; 25: 893–898. 32 Kruse M, Bornstein SR, Uhlmann K, Paeth G, Scherbaum WA. Leptin down- 5 Qi D, Rodrigues B. Glucocorticoids produce whole body insulin resistance with regulates the steroid producing system in the adrenal. Endocr Res 1998; 24: changes in cardiac metabolism. Am J Physiol Endocrinol Metab 2007; 292: 587–590. E654–E667. 33 Buettner R, Scholmerich J, Bollheimer LC. High-fat diets: modeling the metabolic 6 Sowers JR, Whaley-Connell A, Epstein M. Narrative review: the emerging clinical disorders of human obesity in rodents. Obesity 2007; 15:798–808. implications of the role of aldosterone in the metabolic syndrome and resistant 34 Hinchliffe E, Carter S, Owen LJ, Keevil BG. Quantitation of aldosterone in human hypertension. Ann Intern Med 2009; 150:776–783. plasma by ultra high performance liquid chromatography tandem mass spec- 7 Pasquali R, Vicennati V. Activity of the hypothalamic-pituitary-adrenal trometry. J Chromatogr B Anal Technol Biomed Life Sci 2013; 913–914:19–23. axis in different obesity phenotypes. Int J Obes Relat Metab Disord 2000; 24: 35 Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T et al. S47–S49. Fiji: an open-source platform for biological-image analysis. Nat Methods 2012; 9: 8 Pasquali R, Vicennati V, Cacciari M, Pagotto U. The hypothalamic-pituitary-adrenal 676–682. axis activity in obesity and the metabolic syndrome. Ann NY Acad Sci 2006; 1083: 36 Kamentsky L, Jones TR, Fraser A, Bray MA, Logan DJ, Madden KL et al. Improved 111–128. structure, function and compatibility for CellProfiler: modular high-throughput 9 Glass AR, Burman KD, Dahms WT, Boehm TM. Endocrine function in human image analysis software. Bioinformatics 2011; 27:1179–1180. obesity. Metab Clin Exp 1981; 30:89–104. 37 Miller WL, Auchus RJ. The molecular biology, biochemistry, and physiology of 10 Muller-Fielitz H, Lau M, Johren O, Stellmacher F, Schwaninger M, Raasch W. Blood human steroidogenesis and its disorders. Endocr Rev 2011; 32:81–151. pressure response to angiotensin II is enhanced in obese Zucker rats and is 38 Lightman SL, Conway-Campbell BL. The crucial role of pulsatile activity of the HPA attributed to an aldosterone-dependent mechanism. Br J Pharmacol 2012; 166: axis for continuous dynamic equilibration. Nat Rev Neurosci 2010; 11:710–718. 2417–2429. 39 Beuschlein F, Fassnacht M, Klink A, Allolio B, Reincke M. ACTH-receptor expres- 11 Livingstone DE, Jones GC, Smith K, Jamieson PM, Andrew R, Kenyon CJ et al. sion, regulation and role in adrenocortial tumor formation. Eur J Endocrinol 2001; Understanding the role of glucocorticoids in obesity: tissue-specific alterations of 144:199–206. corticosterone metabolism in obese Zucker rats. Endocrinology 2000; 141: 40 Wood MA, Hammer GD. Adrenocortical stem and progenitor cells: unifying model 560–563. of two proposed origins. Mol Cell Endocrinol 2011; 336: 206–212.

© 2015 Macmillan Publishers Limited International Journal of Obesity (2015) 321 – 330 Adrenal cortex in mice fed a high-fat diet MM Swierczynska et al 330 41 Klop B, Elte JW, Cabezas MC. Dyslipidemia in obesity: mechanisms and potential 48SewerMB,DammerEB,JagarlapudiS. Transcriptional regulation of targets. Nutrients 2013; 5: 1218–1240. adrenocortical steroidogenic gene expression. Drug Metab Rev 2007; 39: 42 Palm W, Swierczynska MM, Kumari V, Ehrhart-Bornstein M, Bornstein SR, Eaton S. 371–388. Secretion and signaling activities of lipoprotein-associated hedgehog and non- 49 Schinner S, Willenberg HS, Krause D, Schott M, Lamounier-Zepter V, Krug AW et al. sterol-modified hedgehog in flies and mammals. PLoS Biology 2013; 11: e1001505. Adipocyte-derived products induce the transcription of the StAR promoter and 43 Nachtergaele S, Mydock LK, Krishnan K, Rammohan J, Schlesinger PH, Covey DF stimulate aldosterone and cortisol secretion from adrenocortical cells through the et al. Oxysterols are allosteric activators of the oncoprotein Smoothened. Nat Wnt-signaling pathway. Int J Obes 2007; 31: 864–870. Chem Biol 2012; 8:211–220. 50 Krug AW, Vleugels K, Schinner S, Lamounier-Zepter V, Ziegler CG, 44 Khaliullina H, Panakova D, Eugster C, Riedel F, Carvalho M, Eaton S. Patched Bornstein SR et al. Human adipocytes induce an ERK1/2 MAP kinases-mediated regulates Smoothened trafficking using lipoprotein-derived lipids. Development upregulation of steroidogenic acute regulatory protein (StAR) and an 2009; 136: 4111–4121. angiotensin II-sensitization in human adrenocortical cells. Int J Obes 2007; 45 Bijlsma MF, Spek CA, Zivkovic D, van de Water S, Rezaee F, Peppelenbosch MP. 31: 1605–1616. Repression of smoothened by patched-dependent (pro-)vitamin D3 secretion. 51 Ehrhart-Bornstein M, Lamounier-Zepter V, Schraven A, Langenbach J, PLoS Biol 2006; 4:e232. Willenberg HS, Barthel A et al. Human adipocytes secrete mineralocorticoid- 46 Khaliullina H, Bilgin M, Sampaio JL, Shevchenko A, Eaton S. Lipoproteins releasing factors. Proc Natl Acad Sci USA 2003; 100: 14211–14216. carry endocannabinoids that inhibit the Hedgehog pathway. bioRxiv 2013; 52 Lavoie HA, King SR. Transcriptional regulation of steroidogenic genes: STARD1, doi:10.1101/000570. CYP11A1 and HSD3B. Exp Biol Med (Maywood) 2009; 234: 880–907. 47 Di Marzo V. The endocannabinoid system in obesity and type 2 diabetes. 53 Calle EE, Kaaks R. Overweight, obesity and cancer: epidemiological evidence and Diabetologia 2008; 51:1356–1367. proposed mechanisms. Nat Rev Cancer 2004; 4:579–591.

Supplementary Information accompanies this paper on International Journal of Obesity website (http://www.nature.com/ijo)