11�-HSD1 inhibitors still improve metabolicphenotype in male 11�-HSD1 knock-out micesuggesting off-target mechanisms

Erika Harno1, Elizabeth C. Cottrell2, Alice Yu3, Joanne DeSchoolmeester3,Pablo Morentin Gutierrez4, Mark Denn4, John G. Swales4, Fred W. Goldberg3,Mohammad Bohlooly-Y5, Harriet Andersén5, Martin J. Wild4,Andrew V. Turnbull3, Brendan Leighton3, Anne White1,2

1Neuroscience Research Group, Faculty of Life Sciences, University of Manchester, Manchester M13 9PTUK; 2Centre for Endocrinology & Diabetes, Faculty of Medical and Human Sciences, University ofManchester, Manchester M13 9PT UK; 3CVGI iMed, AstraZeneca, Alderley Park, SK10 4TG, UK.; 4DMPK,AstraZeneca, Alderley Park, SK10 4TG, UK.; 5Transgenic RAD, Discovery Sciences, AstraZeneca, MölndalS-431 83, Sweden.

The enzyme 11�-hydroxysteroid dehydrogenase type 1 (11�-HSD1) is a target for novel type 2diabetes and obesity therapies based on the premise that lowering tissue glucocorticoids will havepositive effects on body weight, glycaemic control and insulin sensitivity. An 11�-HSD1 inhibitor(compound C) inhibited liver 11�-HSD1 by �90%, but only led to small improvements in metabolicparameters in high fat fed (HFF) male C57Bl/6J mice. A four-fold higher concentration producedsimilar enzyme inhibition, but did additionally reduce body weight (17%), food intake (28%) andglucose (22%). We hypothesised that at the higher doses it might be accessing the brain. However,when we developed male brain-specific 11�-HSD1 knock-out mice, they had similar body weightand fat pad mass as well as glucose and insulin responses when HFF compared to HFF Nestin-Crecontrols. We then found that administration of compound C to male global 11�-HSD1 knock-outmice elicited improvements in metabolic parameters, suggesting “off-target” mechanisms. Basedon patent literature, we synthesised another 11�-HSD1 inhibitor (MK-0916) from a different chem-ical series and showed it too had similar “off-target” body weight and food intake effects at highdoses.

In summary, a significant component of the beneficial metabolic effects of these 11�-HSD1 inhib-itors occurs via 11�-HSD1-independent pathways, and only limited efficacy is achievable fromselective 11�-HSD1 inhibition. These data challenge the concept that inhibition of 11�-HSD1 islikely to produce a “step-change” treatment for diabetes and/or obesity.

As rates of metabolic syndrome and its component con-ditions of obesity, type 2 diabetes and hypertension

continue to rise (1), there is an increasing need to findimproved therapies to treat these disorders. Glucocorti-coids are implicated as causal in promoting both obesityand insulin resistance, the latter of which is a key stage inthe progression to type 2 diabetes. Exposure to excessglucocorticoids, as occurs in Cushing’s syndrome, drives

hyperphagia, body weight gain, hyperlipidemia and insu-lin resistance.

Circulating glucocorticoids are derived at least in partby intracellular regeneration of active steroids (cortisol inhumans, corticosterone in rodents) from inactive metab-olites [cortisone/11-dehydrocorticsterone (11-DHC)] bythe enzyme 11�-hydroxysteroid dehydrogenase type 1(11�-HSD1). In obese human subjects, circulating cortisollevels do not correlate with body mass index (BMI), or

ISSN Print 0013-7227 ISSN Online 1945-7170Printed in U.S.A.Copyright © 2013 by The Endocrine SocietyReceived July 1, 2013. Accepted October 11, 2013.

Abbreviations:

E N E R G Y B A L A N C E - O B E S I T Y

doi: 10.1210/en.2013-1613 Endocrinology endo.endojournals.org 1

Endocrinology. First published ahead of print October 29, 2013 as doi:10.1210/en.2013-1613

Copyright (C) 2013 by The Endocrine Society

glucose and insulin concentrations (2) as there is increasedcortisol clearance (3). However increased tissue 11�-HSD1 expression and activity has been demonstrated, no-tably in metabolic tissues including liver and adipose tissue(4–7). This has led to the widely held belief that elevated11�-HSD1 in tissues may be contributing to metabolicdisease (8, 9).

Several elegant studies have highlighted the role of 11�-HSD1 in metabolic syndrome. Mice with global knock-out of 11�-HSD1 (GKO) have lower body weight whenfed high fat diet (HFD), less visceral fat and lower fastingglucose, accompanied by improved glucose tolerance (10,11). Conversely, overexpression of 11�-HSD1 in adiposetissue of mice causes hyperphagia and visceral obesity, andwhen fedHFDthesemice exhibit insulin-resistantdiabetes(12). This defining study provided some of the first evi-dence suggestingacausative linkbetweenelevatedadipose11�-HSD1 and insulin resistance.

Evidence from these knock-out and transgenic studiestogether with studies in humans suggested that decreasingcortisol by inhibition of 11�-HSD1 would be an attractivetarget for new therapeutic agents. As a result many phar-maceutical and biotechnology companies and some aca-demic groups set up programs to develop 11�-HSD1 in-hibitors as a potential therapy for type 2 diabetes. Inpreclinical studies with C57Bl/6J mice fed a HFD, bene-ficial effects of 11�-HSD1 inhibition were observed, in-cluding reduced body weight, food intake, fasting glucoseand insulin (13–17). More recently, Phase IIb clinical trialswith 11�-HSD1 inhibitors resulted in improved glucosehomeostasis and decreased body weight in type 2 diabeticsubjects (18, 19). However, only high doses of 11�-HSD1inhibitors (and very high levels of 11�-HSD1 inhibition)improve glycaemic control in humans and even then theyonly have modest effects (18, 19).

Another inhibitor of 11�-HSD1 (compound C discov-ered by AstraZeneca) is highly effective in reducing en-zyme activity both in vitro and in mouse studies. However,significant beneficial effects on the metabolic phenotypewere only seen when high doses of inhibitor were used. Wetherefore explored whether these compounds were havingtheir beneficial effects by CNS inhibition of 11�-HSD1which required the higher doses of inhibitor to access theCNS or whether high dose inhibitor administrationcaused “off-target” effects. Our data suggest that a sig-nificant component of the beneficial effects of 11�-HSD1inhibitor administration on body weight and glycaemiccontrol occurs via 11�-HSD1-independent mechanisms,and call into question the validity of this enzyme as a drugtarget for the treatment of type 2 diabetes and obesity.

Materials and Methods and Materials

Animals and genotypingThe HSD11B1 gene targeting vector was prepared from a

129/Sv BAC clone (ResGen, Invitrogen). DNA fragments of5.2kb, 1.1kb and 2.6kb (for 5� homology, deletion and 3� ho-mology regions respectively), were cloned into a modified loxPfloxed PGKneo plasmid, linearized and electroporated into R1mouse embryonic (ES) cells. The deletion fragment includes exon2. Candidate ES cell clones were screened by PCR and confirmedby Southern blotting to identify homologous recombination.Targeted ES cells were injected into C57BL/6J blastocysts, andembryos implanted into pseudopregnant B6CBA female mice.Chimeric animals were mated with C57BL/6 to produce agoutiheterozygous animals (F1).

For generation of 11�-HSD1 global knock-out mice (GKO),heterozygous F1 males were bred with Rosa26-Cre knock-inmice (kindly provided by Dr. Philippe Soriano (20) and back-crossed onto a C57BL/6 background at AstraZeneca) to deletethe loxP floxed region including exon 2 and the neo cassette, andGKO mice for studies were obtained by intercrossing. For gen-erating 11�-HSD1 conditional (floxed) mice, the heterozygousF1 mice were bred with EIIa-Cre knock-in mice (C57BL/6 back-ground) to delete the loxP flanked PGKneo cassette.

Offspring were genotyped at weaning by PCR using specificprimers to detect wild-type allele (1.6kb fragment) and null allele(0.6kb fragment), respectively forward: CAGGTGAGTAC-CACTGTGTCTCATTT; reverse: AGTCCGCCTGCAAA-GAGATAGATG. For conditional mice, primers used detectedthe wild-type allele (315bp) or floxed allele (440bp); forward:GATCTGCCCACCTTTGCCTTCT and reverse: CCAT-GAGCTTTCCCGCCTTGACA respectively.

Cre-loxP technology was used to generate mice with CNS-specific disruption of the HSD11B1 gene (BKO). Male miceheterozygous for the conditional HSD11B1 allele and Cre re-combinase under the control of rat nestin promoter (CharlesRiver) were mated with female homozygous conditionalHSD11B1 mice. The resulting Cre-positive 11�-HSD1lox/lox off-spring (BKO) and Cre-negative 11�-HSD1lox/lox (WTf/f) weredetermined by genotyping at weaning. The primers used were:forward: GCTACTTCTTTTCAACCCCTAAAA and reverse:CTGGATAGTTTTTACTGCCAGACC giving a band size of420bp in Cre-positive (BKO) mice and no band in WTlittermates.

C57BL/6J mice used for inhibitor characterization studieswere bred in-house (AstraZeneca, Alderely Park, UK). All workwas carried out in accordance with the U.K. Home Office Ani-mals (Scientific Procedures) Act 1986 and approved by localEthical Review Committee.

Compound DetailsCompound C (2-[(2S,6R)-2,6-dimethylmorpholin-4-yl]-N-

(5-hydroxy-2-adamantyl)-4-[(2R)-oxolan-2-yl]pyrimidine-5-carboxamide) has an IC50 potency of 0.07 �M in mouse as mea-sured by homogeneous time-resolved fluorescence (HTRF). It isin the same chemical series as previously published compounds(21, 22). MK-0916 (3-[1-(4-chlorophenyl)-3-fluorocyclobutyl]-4,5-dicyclopropyl-1,2,4-triazole) has been synthesized from pat-ent literature and its effects have been previously characterized

2 11�-HSD1 inhibitors in KO mice Endocrinology

[Patents WO2003104207, WO2005073200 andWO2007038452; (23, 24).

Experimental designMice were fed standard chow (RM1, Special Diet Services

Ltd) or HFD (60% kcal from fat, D12492, Research Diets) adlibitum for up to 30 weeks to induce obesity. During this periodand throughout the studies the mice were handled on a regularbasis and were therefore acclimatized to handling in advance ofany procedures. Administration of small molecule 11�-HSD1inhibitors to mice was via compound in diet for up to 25 days. Forinhibitor studies, food intake was measured prior to treatmentphase, to calculate the amount of compound required for dietincorporation and in all studies during treatment phase to ex-amine changes in food intake with compound or gene deletion.For determination of fasting glucose, mice were fasted for up to18h prior to tail venesection and glucose measurements weremade using an Accu-check glucometer (Roche). At the end of thestudy, animals were narcosed under CO2/O2 and death con-firmed by cervical dislocation. Mesenteric, inguinal (as a mea-sure of subcutaneous) and epididymal adipose tissues were ex-cised and weighed. Adipose tissues, liver and brain were snapfrozen for subsequent mRNA analysis and enzyme activitymeasurements.

Compound levels were measured in plasma and brain afterextraction in acetonitrile by LC-MS/MS (Thermo TSQ Vantage;ThermoScientific) to determine exposure levels. Free plasma lev-els of compounds were calculated based on the measured com-pound concentrations in plasma then corrected using a constantfree fraction (fu) for each compound. These were generated inhouse in vitro using previously published methodology (25). Freedrug levels in brain were calculated using a constant free fractionin brain (fub) obtained with similar methodology as fu. For com-pound C, fu and fub were 0.43 and 0.27 respectively. For com-pound MK-0916, fu and fub were 0.08 and 0.06 respectively. Inthe HFD experiments, brain concentrations were predicted at thedifferent time points based on measured free compound concen-trations in plasma and a constant [free]brain/[free]plasma ratio foreach compound determined in previous experiments (data notshown).

To represent the level of cover for each compound, the re-sulting free compound concentrations in plasma and brain weredivided by their respective in vitro IC50 values. IC50 values usedwere 0.07 �M and 0.03 �M for Compounds C and MK-0916respectively.

In vitro potency (IC50) assessmentThe conversion of cortisone to cortisol by 11�-HSD1 oxo-

reductase was measured using a cortisol competitive HTRF assay(CisBio International). The assay incubation was carried out inblack 384 well plates (Matrix) consisting of cortisone (160 nM),glucose-6-phosphate (1 mM), NADPH (100 �M), glucose-6-phosphate dehydrogenase (12.5 �g/ml), EDTA (1 mM), assaybuffer (K2HPO4/KH2PO4, 100 mM) pH 7.5, recombinant 11�-HSD1 (1.5 �g/ml) plus test compound in a total volume of 20�l.The assay was incubated for 25 minutes at 37°C and the reactionstopped by the addition of 10�l of 0.5 mM glycyrrhetinic acidplus cortisol-XL665. Anticortisol Cryptate was then added tothe plates and incubated for 2h at room temperature. Fluores-cence was measured and the 665nm:620nm ratio calculated us-

ing an Envision plate reader. This data was used to calculate IC50

values (Origin 7.5).

11�-HSD1 activity assayEnzyme activity (knock-out studies) and target engagement

(compound studies) were determined ex vivo by measuring con-version of 3H-cortisone to 3H-cortisol. Tissues were taken fromknock-out mice or from mice treated with 11�-HSD1 inhibitorsas appropriate. Briefly, tissues were incubated for 10 (liver), 45(brain) or 60min (adipose) in the presence of 3H-cortisone (20nmol/l, 1 �Ci/ml, specific activity 1.97GBq/mmol; Perkin El-mer). Radio-labeled steroids were extracted using ethyl acetateand separated using reversed phase HPLC.

Quantitative mRNA analysisTotal RNA was extracted from brain, hypothalamus, liver

and epididymal adipose tissue using either Trizol reagent (brainand adipose tissues; Invitrogen) or RNeasy mini kit (hypothal-amus and liver; Qiagen), according to the manufacturer’s in-structions. Contaminating genomic DNA was removed using aTurboDNAse kit (Ambion). RNA quantity and integrity wasdetermined (ND1000, NanoDrop Technologies), before one-step qRT-PCR was carried out using Taqman© RNA-to-CT™1-Step Kit (Applied Biosystems) for measurement of HSD11B1mRNA expression (Applied Biosystems). Specific Taqmanprobes were used to amplify a region of exon 2 and 3 of theHSD11B1 gene (Mm01313990 m1), and values were normal-ized to levels of Hypoxanthine-guanine phosphoribosyltrans-ferase (HPRT; Mm00446968 m1), using standard curveanalysis.

Oral Glucose Tolerance TestMice were fasted for 16 hours before baseline blood glucose

was measured (Accu-chek glucometer, Roche) and blood sam-ples were collected for insulin measurement (Chrystal ChemELISA) by tail venesection. Mice were administered 10ml/kg of20% glucose by oral gavage (0mins) and glucose measurementstaken at 20, 40, 60 and 90mins. Blood samples were collected forinsulin measurements at 20 and 40mins.

Assays of circulating analytesTail venesection blood samples from free-running mice were

taken at the previously determined peak and nadir of cortico-sterone. Corticosterone was measured in these samples by a com-mercially available kit (Cayman Chemical). Terminal plasmasamples from fed or overnight fasted mice were used to measureplasma lipids and leptin. NEFAs (WAKO Chemicals), triglycer-ides (Sigma) and leptin (Millipore) were measured by commer-cially available kits.

Dual X-ray Absorptiometry ScanMice were anesthetized using halothane and were scanned in

a Lunar PIXImus Dual X-ray absorptiometry scanner (DEXA,GE Healthcare). The associated software calculated the % fatbased on the scan.

StatisticsData are expressed as mean�SEM. Statistical analysis was

performed using Prism 5 (GraphPad). Unpaired Student’s t test,

doi: 10.1210/en.2013-1613 endo.endojournals.org 3

1-way or 2-way ANOVAs with Bonferonni post hoc tests wereused as appropriate. HSD11B1 gene expression data was notnormally distributed, and so were analyzed by Mann-Whitney ttest. P � .05 was considered significant.

Results

Inhibition of 11�-HSD1 improves the metabolicphenotype of high fat diet fed mice, but only atvery high doses

Compound C is a novel pyrimidine 11�-HSD1 inhib-itor (Figure 1A), related to other compounds in the samechemical series which have previously been published (21,22). This 11�-HSD1 inhibitor has an IC50 of 0.07 �Mwith a plasma free fraction of 0.43. Mice, fed HFD for up

to 30 weeks to induce obesity, were given compound C at50 mg/kg/d in the diet for 20 days. Steady-state levels ofunbound compound in plasma were 44–102 times abovethe IC50 potency (Supplemental Figure 1B) and resulted ina � 90% reduction in 11�-HSD1 enzyme activity in liver(Supplemental Figure 1A). This dose of compound did notsignificantly lower body weight (Figure 1B & C) or foodintake (Figure 1D), although the mice tended to gain lessweight than the control mice on HFD. However, there wasa small, but significant, reduction in fasting glucose (Fig-ure 1E) and leptin (Figure 1F) compared to controls.

Using a higher dose of compound (200 mg/kg/d), weachieved a steady-state unbound compound level inplasma of 299–588 times above IC50 potency (Supple-mental Figure 1B). This higher dose of compound again

Figure 1. The small-molecule 11�-HSD1 inhibitor, compound C, lowers body weight, food intake and glucose in obese mice. (A) Structure ofcompound C. (B) Percentage body weight change over 20 days treatment. (C) Percentage body weight change on day 20. (D) Cumulative foodintake over first 20 days of treatment. (E) Fasting glucose levels on treatment day 21. (F) Fed leptin levels in terminal plasma taken on treatmentday 25. Open circles � vehicle mice, open squares � 50 mg/kg/d compound C, closed squares � 200 mg/kg/d compound C. Data is expressed asmean�SEM, n � 7–10, *P � .05, ***P � .001.

4 11�-HSD1 inhibitors in KO mice Endocrinology

inhibited 11�-HSD1 by more than 90% and did not raisecirculating corticosterone (data not shown). However itreduced body weight (by 17%; Figure 1B & C), and foodintake (Figure 1D). Fasting glucose (Figure 1E) and circu-lating leptin (Figure 1F) were also reduced in treated com-pared to untreated HFF mice and leptin levels were lowerwhen compared to the lower dose of compound as well.Together, these data indicate that compound C has pos-itive effects on the metabolic phenotype in obese mice,albeit only at a very high dose.

Brain 11�-HSD1 deletion does not reduce bodyweight or food intake

Given that our selected compound inhibited liver 11�-HSD1 maximally at both doses tested, but only reducedbody weight and food intake at the higher dose, we con-sidered whether the higher dose was acting in the CNS.Glucocorticoids are known to have an adverse effect onenergy balance and administration of glucocorticoids di-rectly into the brain can both stimulate food intake (26)and inhibit peripheral glucose uptake and insulin sensi-tivity (27, 28). Furthermore, 11�-HSD1 is expressedwithin the hypothalamus (Figure 2A and ref. 10), a keybrain region mediating these effects of central glucocor-ticoids. We therefore examined whether the compoundcould penetrate the brain and found that free inhibitorconcentrations in brain tissue were approximately 11 and67 times IC50 at the low and high dose, respectively (Sup-plemental Figure 1C). These findings support the hypoth-esis that our 11�-HSD1 inhibitor might be acting in theCNS. In order to determine the contribution of CNS 11�-HSD1 inhibition directly, we generated a CNS-specificknock-out mouse by crossing mice with exon 2 of theHSD11B1 gene flanked by loxP sites (floxed mice) with aNestin-Cre strain. Brain 11�-HSD1 knock-out (BKO)mice had a greater than 87% reduction in enzyme activityand a greater than 98% reduction in mRNA expression inthe brain (Figure 2A & B). No effects of gene deletion wereseen in liver, but there was a 50% reduction in mRNAexpression in epididymal adipose tissue, although this didnot lead to a reduction in 11�-HSD1 enzyme activity (Fig-ure 2A & B), a more physiologically relevant measure offunction. Deletion of 11�-HSD1 specifically in the braindid not alter circulating corticosterone concentrations(Figure 2C). In the course of our phenotyping studies, wedetermined that Nestin-Cre mice (designated WTnes) ex-hibit an intrinsic metabolic phenotype (Supplemental Fig-ure 2 and ref 29 & 30), and consequently all metabolicstudies were performed using Nestin-Cre mice as controls.

Both BKO and WTnes mice gained weight similarlywhen fed HFD (Figure 2D), and body fat content was notdifferent between genotypes at the end of 12 weeks HFD

feeding, as measured using dual X-ray absorptiometry(DXA; Figure 2E). Glucose metabolism and insulin sen-sitivity were also not improved in BKO mice (Figure 2J-L),and both genotypes had comparable epididymal, mesen-teric and inguinal fat pad weights (Figure 2F-H). Foodintake was increased in BKO mice compared to WTnes

mice (Figure 2I), but was significantly lower (�15%) inboth these genotypes compared to WTf/f and C57BL/6Jmice (Supplemental Figure 2C and data not shown).

Global 11�-HSD1 knock-out mice on a C57BL/6Jbackground have no improvement in metabolicparameters

Data obtained from BKO mice did not support the hy-pothesis that a component of the beneficial metabolic ef-fects of our inhibitors occurred via a CNS-mediated mech-anism, and led us to consider whether in fact thecompound may be acting via other “off-target” mecha-nisms. In order to test this, we generated and phenotypedmice with a global knock-out of 11�-HSD1 (GKO). Thesemice had reduced levels of 11�-HSD1 mRNA expression(�97%; Figure 3A) and enzyme activity (�92%; at thelimit of detection; Figure 3B). Global deletion of 11�-HSD1 did not cause a rebound increase in circulating cor-ticosterone from the adrenal gland at either peak or nadir(Figure 3D). The effect of 11�-HSD1 inhibitors was thentested in this model. Using this approach, any residualeffects of 11�-HSD1 inhibition could be attributed to non-11�-HSD1 mediated or “off-target” actions.

Mice with global knock-out of 11�-HSD1 did notcounteract body weight gain when fed a HFD and wereessentially similar to WT controls (Figure 3D). This lack ofeffect of whole-body 11�-HSD1 deletion was seen in 5separate studies. In addition, there were no differences inepididymal, mesenteric or inguinal fat pad weights (Figure3E-G) and no change in percentage body fat (Figure 3H).The dyslipidaemia in HFD-fed mice was not improved inGKO mice in that NEFA (Figure 3I) and triglyceride (Fig-ure 3J) levels were similar between GKO and WT mice.There was a marginal, but significant, increase in foodintake in the GKO mice (�6%) compared to WT controls(Figure 3K). In addition, there was no difference betweenWT and GKO mice in terms of glucose tolerance (Figure3L) or insulin sensitivity (Figure 3M) during an OGTT.The stress involved in an OGTT may increase corticoste-rone and therefore increase glucose levels, however instudies carried out by our group in the absence of glucose,multiple repeated blood samples taken over 90min did notincrease glucose levels (data not shown).

doi: 10.1210/en.2013-1613 endo.endojournals.org 5

Body weight loss produced by 11�-HSD1 inhibitorsis predominantly “off-target”

The lack of metabolic improvement in GKO mice cou-pled with the requirement for very high concentrations of

inhibitor to improve metabolic parameters in C57BL/6Janimals further supported the idea that most the effectsseen with compound C were through engagement withanother unknown target. In order to address the possibil-

Figure 2. Brain 11�-HSD1 knock-out mice have no reduction in body weight, fat mass, food intake or glucose tolerance. (A) 11�-HSD1 mRNAexpression in brain, hypothalamus, epididymal adipose and liver, (B) enzyme activity in brain, epididymal adipose and liver tissues in brain 11�-HSD1 knock-out mice (BKO) and wild-type littermates (WTf/f). (C) Circulating corticosterone concentrations in BKO and WTf/f mice after 18 weekshigh fat feeding (HFF). (D) Body weight over 12 weeks HFF in BKO and Nestin-Cre wild-type mice (WTnes). (E) Percentage body fat measured byDEXA scan after 12 weeks HFF. (F) Epididymal, (G) mesenteric and (H) subcutaneus fat pad masses after 14 weeks HFF. (I) Average food intake (g/wk) over 12 weeks HFF. (J) Glucose and (K) insulin excursion during an oral glucose tolerance test (OGTT) after 13 weeks HFF. (L) HOMA-IR after13 weeks HFF. Open squares � WTnes mice, closed squares � BKO mice. Data is expressed as mean�SEM, n � 6–16, *P � .05, **P � .01,***P � .001 vs. appropriate WT strain.

6 11�-HSD1 inhibitors in KO mice Endocrinology

ity of “off-target” effects, we administered compound C toour GKO mouse strain which had been fed HFD.

Obese WT and GKO mice were administered com-pound C (100 mg/kg/d) in their diet for 25 days. At thisdose, the compound reduced 11�-HSD1 enzyme activityin WT mice to a similar level to that seen in the GKO mice(Supplemental Figure 3A). Free plasma and brain levels of

compound were similar between WT and GKO mice (Sup-plemental Figure 3B & C).

WT mice on HFD which were given 100 mg/kg/d com-pound C had approximately 10% lower body weight thancontrols by the end of the treatment period (Figure 4A).However most of this observed weight loss was also seenin mice where 11�-HSD1 had been deleted globally (Fig-

Figure 3. Global 11�-HSD1 knock-out mice do not have an improved metabolic phenotype compared to WT controls. (A) 11�-HSD1 mRNAexpression and (B) enzyme activity in brain, epididymal adipose and liver tissues in global 11�-HSD1 knock-out mice (GKO) and wild-typelittermates (WT). (C) Circulating corticosterone after 18 weeks high fat feeding (HFF). (D) Body weight over 18 weeks HFF. (E) Epididymal (F)mesenteric and (G) subcutaneous fat pad weights after 12 weeks HFF. (H) Percentage body fat measured by DEXA scan after 12 weeks HFF. (I)Plasma nonesterified fatty acids (NEFA) and (J) plasma triglycerides in fasted mice after 20 weeks HFF. (K) Average food intake (g/d) over 2 weeksin HFF mice. (L) Glucose and (M) insulin excursion during an OGTT after 13 weeks HFF. Open triangles � WT mice, closed triangles � GKO mice.Data is expressed as mean�SEM, n � 6–12, *P � .05, ***P � .001 vs. WT.

doi: 10.1210/en.2013-1613 endo.endojournals.org 7

ure 4B). This suggests that much of the body weight lossis “off-target” (Figure 4C) although analysis by two-wayANOVA did indicate an interaction between groups, mak-ing the data more difficult to interpret, so we cannot dis-count that there might be a small effect of 11�-HSD1 in-hibition. There was also a trend towards a reduction infood intake (Figure 4D) and leptin (Figure 4E) in bothgenotypes treated with inhibitor, but this was not signif-

icant, nor was there any effect on fasting glucose in thisstudy (Figure 4F).

The effects seen with compound C in GKO mice may bespecific to this compound or generic to 11�-HSD1 inhib-itors. To test this, we administered a compound from adistinct chemical series (MK-0916; ref 23), which inhib-ited the enzyme in vitro and examined its metabolic effectsin HFD-fed WT and GKO mice. MK-0916 has been used

Figure 4. Positive metabolic effects of the 11�-HSD1 inhibitor, compound C, appear to be “off target”. (A) Body weight change in wild-type(WT) mice with compound C treatment. (B) Body weight change in global 11�-HSD1 knock-out (GKO) mice with compound C treatment. (C)Percentage body weight change on day 25 of treatment. (D) Cumulative food intake during the 25 day treatment. (E) Fed leptin levels in terminalplasma taken on treatment day 25 (F) Fasting glucose levels on treatment day 20. Open square � vehicle treated WT, open circle � 100 mg/kg/dcompound C treated WT, closed square � vehicle treated GKO, closed circle � 100 mg/kg/d compound C treated GKO. Data expressed asmean�SEM, n-5–7 mice per group, *** P � .001 vs. high fat diet treated mice.

8 11�-HSD1 inhibitors in KO mice Endocrinology

to treat patients in clinical trials (19, 31). It inhibits pe-ripheral 11�-HSD1 and can also access the CNS at highdoses. It was synthesized based on patent literature (Figure5A).

MK-0916 (10 mg/kg/d) given to mice in HFD reducedenzyme activity in liver by � 90% in WT mice (Supple-mental Figure 4A). Free concentrations of the compoundwere found at similar levels in WT and GKO mice both inplasma (Supplemental Figure 4B) and in brain (Supple-mental Figure 4C). Over the 21 day treatment period with

MK-0916, mice lost more weight compared to their HFD-fed controls (2-way ANOVA; Figure 5B). This weight lossdid not occur in GKO mice (Figure 5C) indicating it wasvia an 11�-HSD1-specific mechanism. At this dose ofcompound, no reduction in food intake (Figure 5D) or fedglucose (Figure 5E) was observed in either genotype.

As there was a relatively small weight loss with 10 mg/kg/d of MK-0916, we administered a higher dose of thiscompound to both WT and GKO mice, to get a moreclinically relevant body weight loss. Again at this dose,

Figure 5. Body weight and food intake effects of lower dose 11�-HSD1 inhibitor, MK-0916, are “on target”. (A) Structure of MK-0916 (B) Bodyweight change in wild-type (WT) and (C) global 11�-HSD1 knock-out (GKO) mice with 10 mg/kg/d MK-0916 treatment. (D) Cumulative foodintake during the 22 day treatment. (E) Fasting glucose levels on treatment day 16. Open square � vehicle treated WT, open circle � 10 mg/kg/dMK-0916 treated WT, closed square � vehicle treated GKO, closed circle � 10 mg/kg/d MK-0916 treated GKO Data expressed as mean�SEM,5–8. *P � .05 vs. vehicle treatment for that genotype.

doi: 10.1210/en.2013-1613 endo.endojournals.org 9

11�-HSD1 was inhibited by more than 90% (Supplemen-tal Figure 4D) and free compound levels did not differbetween the WT and GKO mice in plasma (SupplementalFigure 4E) and brain (Supplemental Figure 4F). With thishigher dose of inhibitor, HFD-fed WT mice lost approx-imately 12% of their body weight (Figure 6A & C). Asimilar magnitude of body weight loss was also observedin GKO mice (Figure 6B & C), indicating that the bodyweight loss was “off-target”. Treatment with MK-0916caused a reduction in food intake in both WT and GKO

mice (Figure 6D), indicating that this was also an “off-target” effect. No decrease in fed glucose levels (Figure 6E)were observed in either WT or GKO mice. There was atrend towards an improvement in insulin sensitivity, asmeasured by HOMA-IR in both the WT and the GKOmice treated with MK-0916. The magnitude of improve-ment in insulin sensitivity appeared greater in the WTmice, but neither genotype reached significance (Figure6F).

Figure 6. Body weight and food intake effects of higher dose 11�-HSD1 inhibitor, MK-0916, are “off target”. (A) Body weight change in WT and(B) global 11�-HSD1 knock-out (GKO) mice with 30 mg/kg/d MK-0916 treatment. (C) Percentage body weight change on day 21. (D) Cumulativefood intake during the 22 day treatment. (E) Fasting glucose levels on treatment day 16. (F) Insulin resistance as measured by HOMA-IR attreatment day 16. Open square � vehicle treated WT, open circle � 30 mg/kg/d MK-0916 treated WT, closed square � vehicle treated GKO,closed circle � 30 mg/kg/d MK-0916 treated GKO. Data expressed as mean�SEM, n � 6–8. ** P � .01, ***P � .001 vs. vehicle treatment forthat genotype.

10 11�-HSD1 inhibitors in KO mice Endocrinology

Discussion

Inhibition of 11�-HSD1 has been targeted by many phar-maceutical companies as a promising new therapy to treattype 2 diabetes and obesity, primarily because of the an-ticipated effects in lowering glucose and reducing bodyweight in obese individuals. However, despite numerouspublished preclinical studies over the last decade, andmany compounds entering clinical development, no 11�-HSD1 inhibitor has made it to Phase III clinical trials. Withthe high levels of attrition of Phase I/II development com-pounds it is not easy to determine whether the lack ofsuccess is due to lack of effectiveness at the target orwhether it is due to the individual compounds. For 11�-HSD1, while the original studies with knock-out modelswere encouraging, the studies presented here and morerecent data from other groups (32, 33) suggest that thereis little improvement in metabolic phenotype in the ab-sence of 11�-HSD1. In parallel, our studies with 11�-HSD1 inhibitors have produced similar profiles to otherpreclinical studies in HFF C57Bl/6 mice in that high dosesof inhibitors are needed to successfully reduce blood glu-cose and body weight. We have addressed this disconnectdirectly in our present studies, by examining the effects of11�-HSD1 inhibition in mice lacking this enzyme, andshown that although there are some small improvementsin metabolic parameters with low doses of compounds,the effects seen with high doses of two inhibitors arelargely due to “off-target” mechanisms. We therefore be-lieve that it is of importance to reconsider the true poten-tial of 11�-HSD1 inhibition in the development of noveltherapies for treatment of type 2 diabetes and obesity.

A conundrum appears to exist in studies with 11�-HSD1 inhibitors where low concentrations of compoundsmaximally inhibit the enzyme, but higher concentrationsare needed to have more marked pharmacodynamic ef-fects. In a Phase IIa study in patients with type 2 diabetes,MK-0916 (a compound we studied here) failed to lowerfasting plasma glucose even with approximately 85% en-zyme inhibition in liver, but did deliver a modest degree ofweight loss (19). Notably this profile was very similar tothat observed by the low dose of MK-0916 in the currentpreclinical study, which delivered its modest weight losseffects through on-target mechanisms. In a Phase IIb clin-ical trial in type 2 diabetic patients, INCB13739 producedlowering of HbA1c, fasting plasma glucose andHOMA-IR as well as a modest decrease in body weight,but only at higher doses (100 mg and above). Interestingly,a 50 mg dose of inhibitor already reached 90% inhibitionof 11�-HSD1 in the liver (18, 34, 35), but was withouteffects on HbA1c or body weight. Thus, clinical data in-dicate that only very high doses of inhibitors, giving very

high levels of enzyme inhibition, are significantlyefficacious.

We demonstrated that although inhibition of 11�-HSD1 with small molecule inhibitors is very effective, ourselected compound (compound C) at the lower dose (50mg/kg/d) was not able to produce a relevant improvementin metabolic phenotype. As a caveat, high fat feeding inmice decreases the levels of 11�-HSD1 in adipose tissue(36), and therefore it may be more difficult to demonstrateinhibition of the enzyme. Nevertheless, a four-fold in-crease in the dose of compound C (200 mg/kg/d) also re-sulted in greater than 90% inhibition of the enzyme, butadditionally gave much larger reductions in body weight,food intake and fasting glucose levels. The need for veryhigh doses of inhibitor is not restricted to our studies asnoted above. This led us to consider that exposures ofperipheral tissues to the drug were not the major deter-minant of efficacy, and achieving high CNS drug expo-sures might be required.

It was apparent that compound C was able to penetratethe brain, as at the high dose the CNS levels reached valuesgreater than those required to inhibit enzyme activity. Fur-thermore expression of 11�-HSD1 has been localized tomany regions of the brain, one of which is the arcuatenucleus of the hypothalamus (10, 37, 38), an area impor-tant for the control of both food intake and glucose ho-meostasis. Within the CNS, glucocorticoids have beenshown to stimulate appetite (39, 40), and to induce hepaticinsulin resistance (27). This suggested to us that decreasingactive glucocorticoids within the CNS might produce ben-eficial effects in terms of reducing food intake and im-proving glucose homeostasis. Our generation and pheno-typing of mice lacking 11�-HSD1 specifically within theCNS has shown that these animals are almost indistin-guishable compared with their Nestin-Cre controls, indi-cating that expression of this enzyme within the brain con-tributes minimally to the regulation of body weight andglucose metabolism. However, the BKO mice did haveincreased food intake compared to Nestin-Cre controlmice, but both these genotypes have significantly lowerfood intake than 11�-HSD1 floxed or C57Bl/6J controlmice and therefore the increased food intake in BKO mice,compared to WTnes mice, is most likely due to the back-ground strains. Nevertheless, as mice with the Nestin-Creconstruct have a metabolic phenotype (Supplemental Fig-ure 2 & refs 29, 30) we cannot discount the possibility thatthe presence of Nestin-Cre may be masking a subtle low-ering of body weight, food intake and/or glucose levelscaused by knock-out of 11�-HSD1 in the brain.

The lack of beneficial metabolic effect when 11�-HSD1was knocked out in the CNS suggests that the requirementfor a high dose of inhibitor to have a metabolic effect is not

doi: 10.1210/en.2013-1613 endo.endojournals.org 11

because of a need to access the brain for efficacy. Ouralternative hypothesis was that the greater metabolic ef-fects of high doses of 11�-HSD1 inhibitors could be me-diated via non-11�-HSD1 ie, “off-target” mechanisms.To test this we evaluated the effects of compounds inglobal 11�-HSD1 knock-out (GKO) mice. The GKOmouse we developed did not display metabolic improve-ments (no effect on body weight gain, adiposity or insulinsensitivity) and even had increased food intake. In con-trast, the original GKO phenotype showed improvementsin glucose tolerance and reduced body weight gain onHFD (11) and this was maintained when the genotype wasbackcrossed onto C57BL/6 mice. These mice were shownto have increased core body temperature, which wasthought to be responsible, at least in part, for the reducedbody weight (41). This may explain the increased foodintake without alterations in body weight in the presentstudy. However more recently there has been some dis-parity between published studies, with some not findingthe same degree of metabolic protection of GKO (32, 33)which is in accordance with our own data presented here,and other recent studies showing some small degree ofmetabolic improvement (42). Why differences are seen isunclear, however it may be due to small differences instrain background or housing conditions.

Our findings from the GKO studies indicate that witha low dose of 11�-HSD1 inhibitor there are some 11�-HSD1-dependent decreases in body weight, as seen withMK-0916. These reductions in body weight are unlikely tobe relevant for a novel type 2 diabetes and obesity therapy,the target for many of these inhibitors entering clinicaltrials. However, at higher doses, the 11�-HSD1 inhibitorsgave greater levels of efficacy. Therefore, in this, and pos-sibly other, preclinical mouse studies, these larger im-provements are likely to be partially due to “off-target”mechanisms. It is however unexpected that two com-pounds with completely different chemistry would givesimilar “off-target” effects. A potential explanation forthis may be that high doses of compounds induce nausea,however the mice were closely monitored for symptomsand showed no ill effects while on the treatment.

There has been a huge drive to develop pharmacolog-ical approaches to inhibiting 11�-HSD1 as a new therapyfor type 2 diabetes and obesity. However, despite this todate, the development of all these inhibitors has beenhalted prior to Phase III clinical trials. There may be anumber of reasons for this, including potential side effectsin relation to perturbation of the HPA axis (43). However,the data reported herein demonstrate that selective inhi-bition of 11�-HSD1 can only deliver small improvementsin glycaemic control and body weight and that the largerimprovements observed in preclinical trials are likely to be

“off-target”. Indeed they support the hypothesis that onlylimited effects are likely clinically, suggesting that the clin-ical effects observed to date may be the best that can beexpected from selective inhibition of 11�-HSD1. Recentpreclinical evidence suggesting a role for 11�-HSD1 in thedevelopment of atherosclerotic lesions (44, 45) may openup additional aspects of cardiovascular disease that maybe targeted by 11�-HSD1 inhibition. Nevertheless it ap-pears that at present, the selective inhibition of 11�-HSD1will only lead to small improvements in glucose controland obesity and therefore cannot be considered to be astrong therapeutic target for producing a “step-change”treatment of obesity or type 2 diabetes.

Acknowledgments

The authors would like to thank AstraZeneca and the Universityof Manchester Biomedical Research Centre for their funding andsupport during the studies. Grateful thanks are also given toDavid Tucker, Lynne Vicary and the transgenic team for theproduction and genotyping of the 11�-HSD1 KO mice, to BillBrown, Elizabeth Kelly, Laura Hayward and the IPG-AnimalScience & Welfare support team for the generation of experi-mental results, to Gemma Convey, Julie Cook and Usha Chau-han for analytical support and to Elizabeth Pilling and PeterCeuppens for Statistical Support.

Address all correspondence and requests for reprints to: Cor-respondence to: awhite@manchester.ac.uk, Faculty of Life Sci-ences, & Faculty of Medical and Human Sciences, AV Hill Build-ing, University of Manchester, Manchester, M13 9PT, UK.,E-mail: awhite@manchester.ac.uk, Telephone: �44 (0) 161 2755178.

This work was supported by .Disclosure Summary: EH and AW receive grant funding from

AstraZeneca. AY, JDeS, PMG, MD, JGS, FWG, MB-Y, HA,AVT and BL are employees of AstraZeneca. ECC has nothing todeclare.

References

1. Reaven GM. Why Syndrome X? From Harold Himsworth to theinsulin resistance syndrome. Cell metabolism. 2005;1(1):9–14.

2. Pavlatou MG, Vickers KC, Varma S, Malek R, Sampson M, Rema-ley AT, Gold PW, Skarulis MC, Kino T. Circulating cortisol-asso-ciated signature of glucocorticoid-related gene expression in subcu-taneous fat of obese subjects. Obesity. 2013;21(5):960–967.

3. Walker BR. Cortisol–cause and cure for metabolic syndrome? Di-abetic medicine : a journal of the British Diabetic Association. 2006;23(12):1281–1288.

4. Rask E, Olsson T, Soderberg S, Andrew R, Livingstone DE, JohnsonO, Walker BR. Tissue-specific dysregulation of cortisol metabolismin human obesity. The Journal of clinical endocrinology and me-tabolism. 2001;86(3):1418–1421.

5. Valsamakis G, Anwar A, Tomlinson JW, Shackleton CH, McTer-

12 11�-HSD1 inhibitors in KO mice Endocrinology

nan PG, Chetty R, Wood PJ, Banerjee AK, Holder G, Barnett AH,Stewart PM, Kumar S. 11beta-hydroxysteroid dehydrogenase type1 activity in lean and obese males with type 2 diabetes mellitus. TheJournal of clinical endocrinology and metabolism. 2004;89(9):4755–4761.

6. Westerbacka J, Corner A, Kannisto K, Kolak M, Makkonen J, Kor-sheninnikova E, Nyman T, Hamsten A, Fisher RM, Yki-Jarvinen H.Acute in vivo effects of insulin on gene expression in adipose tissuein insulin-resistant and insulin-sensitive subjects. Diabetologia.2006;49(1):132–140.

7. Sjostrand M, Jansson PA, Palming J, de Schoolmeester J, Gill D, ReesA, Sjogren L, Persson T, Eriksson JW. Repeated measurements of11beta-HSD-1 activity in subcutaneous adipose tissue from lean,abdominally obese, and type 2 diabetes subjects–no change follow-ing a mixed meal. Hormone and metabolic research � Hormon- undStoffwechselforschung � Hormones et metabolisme. 2010;42(11):798–802.

8. Morton NM. Obesity and corticosteroids: 11beta-hydroxysteroidtype 1 as a cause and therapeutic target in metabolic disease. Mo-lecular and cellular endocrinology. 2010;316(2):154–164.

9. Gathercole LL, Stewart PM. Targeting the pre-receptor metabolismof cortisol as a novel therapy in obesity and diabetes. The Journal ofsteroid biochemistry and molecular biology. 2010;122(1–3):21–27.

10. Densmore VS, Morton NM, Mullins JJ, Seckl JR. 11 beta-hydrox-ysteroid dehydrogenase type 1 induction in the arcuate nucleus byhigh-fat feeding: A novel constraint to hyperphagia? Endocrinology.2006;147(9):4486–4495.

11. Kotelevtsev Y, Holmes MC, Burchell A, Houston PM, Schmoll D,Jamieson P, Best R, Brown R, Edwards CR, Seckl JR, Mullins JJ.11beta-hydroxysteroid dehydrogenase type 1 knockout mice showattenuated glucocorticoid-inducible responses and resist hypergly-cemia on obesity or stress. Proceedings of the National Academy ofSciences of the United States of America. 1997;94(26):14924–14929.

12. Masuzaki H, Paterson J, Shinyama H, Morton NM, Mullins JJ,Seckl JR, Flier JS. A transgenic model of visceral obesity and themetabolic syndrome. Science (New York, NY. 2001;294(5549):2166–2170.

13. Morgan SA, Sherlock M, Gathercole LL, Lavery GG, Lenaghan C,Bujalska IJ, Laber D, Yu A, Convey G, Mayers R, Hegyi K, Sethi JK,Stewart PM, Smith DM, Tomlinson JW. 11beta-hydroxysteroid de-hydrogenase type 1 regulates glucocorticoid-induced insulin resis-tance in skeletal muscle. Diabetes. 2009;58(11):2506–2515.

14. Veniant MM, Hale C, Komorowski R, Chen MM, St Jean DJ, FotschC, Wang M. Time of the day for 11beta-HSD1 inhibition plays a rolein improving glucose homeostasis in DIO mice. Diabetes obesitymetabolism. 2009;11(2):109–117.

15. Wan ZK, Chenail E, Xiang J, Li HQ, Ipek M, Bard J, Svenson K,Mansour TS, Xu X, Tian X, Suri V, Hahm S, Xing Y, Johnson CE,Li X, Qadri A, Panza D, Perreault M, Tobin JF, Saiah E. Efficacious11beta-hydroxysteroid dehydrogenase type I inhibitors in the diet-induced obesity mouse model. Journal of medicinal chemistry.2009;52(17):5449–5461.

16. Wang SJ, Birtles S, de Schoolmeester J, Swales J, Moody G, HislopD, O’Dowd J, Smith DM, Turnbull AV, Arch JR. Inhibition of11beta-hydroxysteroid dehydrogenase type 1 reduces food intakeand weight gain but maintains energy expenditure in diet-inducedobese mice. Diabetologia. 2006;49(6):1333–1337.

17. Hermanowski-Vosatka A, Balkovec JM, Cheng K, Chen HY, Her-nandez M, Koo GC, Le Grand CB, Li Z, Metzger JM, Mundt SS,Noonan H, Nunes CN, Olson SH, Pikounis B, Ren N, Robertson N,Schaeffer JM, Shah K, Springer MS, Strack AM, Strowski M, Wu K,Wu T, Xiao J, Zhang BB, Wright SD, Thieringer R. 11beta-HSD1inhibition ameliorates metabolic syndrome and prevents progres-sion of atherosclerosis in mice. The Journal of experimental medi-cine. 2005;202(4):517–527.

18. Rosenstock J, Banarer S, Fonseca VA, Inzucchi SE, Sun W, Yao W,

Hollis G, Flores R, Levy R, Williams WV, Seckl JR, Huber R. The11-beta-hydroxysteroid dehydrogenase type 1 inhibitorINCB13739 improves hyperglycemia in patients with type 2 diabe-tes inadequately controlled by metformin monotherapy. Diabetescare. 2010;33(7):1516–1522.

19. Feig PU, Shah S, Hermanowski-Vosatka A, Plotkin D, Springer MS,Donahue S, Thach C, Klein EJ, Lai E, Kaufman KD. Effects of an11beta-hydroxysteroid dehydrogenase type 1 inhibitor, MK-0916,in patients with type 2 diabetes mellitus and metabolic syndrome.Diabetes obesity metabolism. 2011;13(6):498–504.

20. Soriano P. Generalized lacZ expression with the ROSA26 Cre re-porter strain. Nature genetics. 1999;21(1):70–71.

21. Goldberg FW, Leach AG, Scott JS, Snelson WL, Groombridge SD,Donald CS, Bennett SN, Bodin C, Gutierrez PM, Gyte AC. Free-Wilson and structural approaches to co-optimizing human and ro-dent isoform potency for 11beta-hydroxysteroid dehydrogenasetype 1 (11beta-HSD1) inhibitors. Journal of medicinal chemistry.2012;55(23):10652–10661.

22. Scott JS, Gill AL, Godfrey L, Groombridge SD, Rees A, Revill J,Schofield P, Sorme P, Stocker A, Swales JG, Whittamore PR. Op-timisation of pharmacokinetic properties in a neutral series of 11be-ta-HSD1 inhibitors. Bioorganic medicinal chemistry letters. 2012;22(21):6756–6761.

23. Bauman DR, Whitehead A, Contino LC, Cui J, Garcia-Calvo M, GuX, Kevin N, Ma X, Pai LY, Shah K, Shen X, Stribling S, Zokian HJ,Metzger J, Shevell DE, Waddell ST. Evaluation of selective inhibi-tors of 11beta-HSD1 for the treatment of hypertension. Bioorganicmedicinal chemistry letters. 2013;23(12):3650–3653.

24. Wright DH, Stone JA, Crumley TM, Wenning L, Zheng W, Yan K,Yang AY, Sun L, Cilissen C, Ramael S, Hermanowski-Vosatka A,Langdon RB, Gottesdiener KM, Wagner JA, Lai E. Pharmacoki-netic/pharmacodynamic studies of the 11beta-HSD1 inhibitor MK-0916 in healthy subjects. British journal of clinical pharmacology.2013; doi: 10.1111/bcp. 12131.

25. Buttar D, Colclough N, Gerhardt S, MacFaul PA, Phillips SD, Plow-right A, Whittamore P, Tam K, Maskos K, Steinbacher S, Steuber H.A combined spectroscopic and crystallographic approach to prob-ing drug-human serum albumin interactions. Bioorganic medicinalchemistry. 2010;18(21):7486–7496.

26. Zakrzewska KE, Cusin I, Stricker-Krongrad A, Boss O, Ricquier D,Jeanrenaud B, Rohner-Jeanrenaud F. Induction of obesity and hy-perleptinemia by central glucocorticoid infusion in the rat. Diabetes.1999;48(2):365–370.

27. Yi CX, Foppen E, Abplanalp W, Gao Y, Alkemade A, la Fleur SE,Serlie MJ, Fliers E, Buijs RM, Tschop MH, Kalsbeek A. Glucocor-ticoid signaling in the arcuate nucleus modulates hepatic insulinsensitivity. Diabetes. 2012;61(2):339–345.

28. Cusin I, Rouru J, Rohner-Jeanrenaud F. Intracerebroventricular glu-cocorticoid infusion in normal rats: induction of parasympathetic-mediated obesity and insulin resistance. Obesity research. 2001;9(7):401–406.

29. Harno E, Cottrell EC, White A. Metabolic pitfalls of CNS cre-basedtechnology. Cell metabolism. 2013;18(1):21–28.

30. Briancon N, McNay DE, Maratos-Flier E, Flier JS. Combined neuralinactivation of suppressor of cytokine signaling-3 and protein-ty-rosine phosphatase-1B reveals additive, synergistic, and factor-spe-cific roles in the regulation of body energy balance. Diabetes. 2010;59(12):3074–3084.

31. Shah S, Hermanowski-Vosatka A, Gibson K, Ruck RA, Jia G, ZhangJ, Hwang PM, Ryan NW, Langdon RB, Feig PU. Efficacy and safetyof the selective 11beta-HSD-1 inhibitors MK-0736 and MK-0916 inoverweight and obese patients with hypertension. J Am Soc Hyper-tens. 2011;5(3):166–176.

32. Michailidou Z, Turban S, Miller E, Zou X, Schrader J, Ratcliffe PJ,Hadoke PW, Walker BR, Iredale JP, Morton NM, Seckl JR. In-creased angiogenesis protects against adipose hypoxia and fibrosisin metabolic disease-resistant 11beta-hydroxysteroid dehydroge-

doi: 10.1210/en.2013-1613 endo.endojournals.org 13

nase type 1 (HSD1)-deficient mice. The Journal of biological chem-istry. 2012;287(6):4188–4197.

33. Nixon M, Wake DJ, Livingstone DE, Stimson RH, Esteves CL, SecklJR, Chapman KE, Andrew R, Walker BR. Salicylate downregulates11beta-HSD1 expression in adipose tissue in obese mice and in hu-mans, mediating insulin sensitization. Diabetes. 2012;61(4):790–796.

34. Hawkins M, Hunter D, Kishore P, Schwartz S, Hompesch M, HollisG, Levy R, Williams B, Huber R. INCB13739, a selective inhibitorof 11�-hydroxysteroid degydrogenase type 1 (11�HSD1), im-proved insulin sensitivity and lowers plasma cholesterol over 28days in patients with type 2 diabetes mellitus. Diabetes. 2008;57(Suppl 1):A99–A100.

35. Hollis G, Huber R. 11beta-Hydroxysteroid dehydrogenase type 1inhibition in type 2 diabetes mellitus. Diabetes obesity metabolism.2011;13(1):1–6.

36. Morton NM, Ramage L, Seckl JR. Down-regulation of adipose11beta-hydroxysteroid dehydrogenase type 1 by high-fat feeding inmice: a potential adaptive mechanism counteracting metabolic dis-ease. Endocrinology. 2004;145(6):2707–2712.

37. Yau JL, Seckl JR. 11beta-hydroxysteroid dehydrogenase type I in thebrain; thickening the glucocorticoid soup. Molecular psychiatry.2001;6(6):611–614.

38. Bisschop PH, Dekker MJ, Osterthun W, Kwakkel J, Anink JJ,Boelen A, Unmehopa UA, Koper JW, Lamberts SW, Stewart PM,Swaab DF, Fliers E. Expression of 11beta-hydroxysteroid dehydro-genase type 1 in the human hypothalamus. Journal of neuroendo-crinology. 2013;25(5):425–432.

39. Savontaus E, Conwell IM, Wardlaw SL. Effects of adrenalectomy onAGRP, POMC, NPY and CART gene expression in the basal hy-

pothalamus of fed and fasted rats. Brain research. 2002;958(1):130–138.

40. Beaulieu S, Gagne B, Barden N. Glucocorticoid regulation of pro-opiomelanocortin messenger ribonucleic acid content of rat hypo-thalamus. Molecular endocrinology (Baltimore, MD. 1988;2(8):727–731.

41. Morton NM, Paterson JM, Masuzaki H, Holmes MC, Staels B,Fievet C, Walker BR, Flier JS, Mullins JJ, Seckl JR. Novel adiposetissue-mediated resistance to diet-induced visceral obesity in 11 be-ta-hydroxysteroid dehydrogenase type 1-deficient mice. Diabetes.2004;53(4):931–938.

42. Wamil M, Battle JH, Turban S, Kipari T, Seguret D, de Sousa Pei-xoto R, Nelson YB, Nowakowska D, Ferenbach D, Ramage L,Chapman KE, Hughes J, Dunbar DR, Seckl JR, Morton NM. Novelfat depot-specific mechanisms underlie resistance to visceral obesityand inflammation in 11 beta-hydroxysteroid dehydrogenase type1-deficient mice. Diabetes. 2011;60(4):1158–1167.

43. Harno E, White A. Will treating diabetes with 11beta-HSD1 inhib-itors affect the HPA axis? Trends in endocrinology and metabolism:TEM. 2010;21(10):619–627.

44. Iqbal J, Macdonald LJ, Low L, Seckl JR, Yau CW, Walker BR,Hadoke PW. Contribution of endogenous glucocorticoids and theirintravascular metabolism by 11beta-HSDs to postangioplasty neo-intimal proliferation in mice. Endocrinology. 2012;153(12):5896–5905.

45. Garcia RA, Search DJ, Lupisella JA, Ostrowski J, Guan B, Chen J,Yang WP, Truong A, He A, Zhang R, Yan M, Hellings SE, Garga-lovic PS, Ryan CS, Watson LM, Langish RA, Shipkova PA, CarsonNL, Taylor JR, Yang R, Psaltis GC, Harrity TW, Robl JA, GordonDA. 11beta-hydroxysteroid dehydrogenase type 1 gene knockoutattenuates atherosclerosis and in vivo foam cell formation in hyper-lipidemic apoE(-)/(-) mice. PloS one. 2013;8(2):e53192.

14 11�-HSD1 inhibitors in KO mice Endocrinology