Inhibition of 11β-HSD1 activity in vivo limits glucocorticoid exposure to human adipose tissue and decreases lipolysis

Abstract

The pathophysiological importance of glucocorticoids (GC) is exemplified by patients with Cushing’s syndrome who develop hypertension, obesity and insulin resistance. At a cellular level, availability of GC to the glucocorticoid and mineralocorticoid receptors (GR and MR) is controlled by the isoforms of 11β-hydroxysteroid dehydrogenase. In liver and adipose tissue, 11β-HSD1, converts endogenous, inactive cortisone to active cortisol, but also catalyses the bio-activation of the synthetic prednisone to prednisolone.

7 healthy males were investigated before and after a single dose of the 11β-HSD inhibitor, carbenoxolone (CBX) (100mg) and after 72hrs of treatment (100mg 3x/day). Inhibition of 11β-HSD1 was monitored using five different mechanistic biomarkers and each demonstrated reduced 11β-HSD1 activity after CBX administration. After both a single dose and 72hrs of treatment with CBX, cortisol and prednisolone generation decreased as did the urinary THF+5αTHF:THE ratio. Using adipose tissue microdialysis, we observed decreased interstitial fluid cortisol availability with CBX treatment. Furthermore, a functional consequence of 11β-HSD1 inhibition was observed, namely decreased prednisone-induced glycerol release into adipose tissue interstitial fluid indicative of inhibition of GC mediated lipolysis.

In conclusion, CBX is able to rapidly inhibit the generation of active glucocorticoid in human adipose tissue. Importantly, limiting GC availability in vivo has functional consequences including decreased glycerol release.

The phenotypic similarities between patients with cortisol excess, Cushing’s syndrome and those with obesity have highlighted the important role that glucocorticoids (GC) play in the control of body composition and metabolism. However, obesity is not a state of circulating cortisol excess (1). GC availability to bind to the glucocorticoid and mineralocorticoid receptors (GR and MR) is controlled by the isoenzymes of 11β-hydroxysteroid dehydrogenase. Two isoforms have been identified; 11β-HSD2 is located in mineralocorticoid target tissues (kidney, placenta) that inactivates cortisol to cortisone and serves to protect the mineralocorticoid receptor (MR) from occupation by cortisol for which it shares similar affinity as aldosterone (2) and 11β-HSD1 in GC target tissues (liver, adipose and muscle). 11β-HSD1 is a bidirectional endolumenal enzyme that in vivo acts predominantly as an oxo-reductase generating active cortisol from inactive cortisone and thus amplifies GC action locally (3). Activity is highly regulated by a number of factors including pro-inflammatory cytokines and growth factors (4-6), is crucially dependent upon cofactor (NADPH) availability (7-9) and is dysregulated in obesity (10;11). Although levels of expression and activity of 11β-HSD1 have been much debated in the published literature (11-16), the most fundamental and pertinent observation is that selective 11β-HSD1 inhibition (selective in that it inhibits 11β-HSD1 and not 11β-HSD2) improves glucose tolerance and insulin sensitivity in rodent models (17-19). However, currently these compounds are not available for use in clinical studies. Liquorice derivatives, glycyrrhizic acid and its hydrolytic product, glycyrrhetinic acid (GE) are potent inhibitors of both 11β-HSD1 and 2 (20-23) causing hypertension and hypokalaemia as a consequence of impaired inactivation of cortisol (through inhibition of 11β-HSD2), allowing it to bind and activate the MR. The hemisuccinate derivative of GE, carbenoxolone (CBX) also inhibits 11β-HSD1 and 2. It has been used in several clinical studies and improves whole body insulin sensitivity and decreases glucose production through inhibition of glycogenolysis (24;25). However, concern has been expressed that it may not be able to access adipose tissue and this has important implications for selective 11β-HSD1 inhibition as a potential therapeutic strategy in humans (26).

GC have potent effects upon adipocyte biology. As well as inducing adipocyte differentiation (27), they inhibit omental pre-adipocyte proliferation (13) and induce lipolysis (28). Whilst the mechanism of action has not been completely defined it is likely that GC act to increase the activity of hormone sensitive lipase (HSL) to hydrolyse triacylglycerol to diacylglycerol with the eventual release of free fatty acids and glycerol (28). 11β-HSD1 is more highly expressed in omental preadipocytes compared to subcutaneous and it is believed to have a fundamental role to promote adipocyte differentiation and to limit pre-adipocyte proliferation (13;29). Further evidence as to the importance of 11β-HSD1 has been provided by rodent models. Over-expression specifically within adipose tissue recapitulates many of the features of the metabolic syndrome including central obesity, hypertension and dyslipidaemia (30;31). Similarly, over–expression specifically within the liver causes features of the metabolic syndrome without obesity (32). Mice with targeted deletion of HSD11B1 resist diet induced obesity and develop relative insulin sensitivity (33). More recently transgenic animals have been created that over-express 11β-HSD2 which is not normally expressed in adipose tissue; local inactivation of GC by this mechanism protects against the adverse metabolic consequences of diet induced obesity (34).

On this background, we have conducted a detailed clinical study to directly compare and assess markers of 11β-HSD1 inhibition in vivo . In addition, as a proof of concept experiment using the non-selective inhibitor CBX (selective 11β-HSD1 inhibitors are not currently available for clinical use), we have investigated whether 11β-HSD1 inhibition can limit the availability of both cortisol and prednisolone (derived from the synthetic, inactive GC, prednisone that is dependent upon 11β-HSD1 reductase activity to generate bioactive prednisolone) in serum and adipose tissue interstitial fluid. Furthermore, we have hypothesised that decreasing GC availability (both in serum and adipose interstitial fluid) through 11β-HSD1 inhibition, has functional consequences to limit adipose tissue lipolysis.

Patients and Methods

The study was approved by South Birmingham local research ethics committee and all subjects gave their informed, written consent. 7 healthy male volunteers (mean age 28±3years, mean BMI 24.5±1.3kg/m²) were recruited following local advertisement and underwent 2 separate weeks of investigation as described below. Each week of investigation was separated by a washout period of at last 8 weeks.

Clinical protocol

Week 1

The aims of the week 1 protocol were firstly to demonstrate that inhibition of 11β-HSD1 limits prednisolone generation in serum and adipose tissue, and secondly to assess the functional impact of this upon adipose tissue lipolysis. Subjects were investigated in the fasted state, water being available ad libitum. At 08.30h an adipose tissue microdialysis catheter (CMA60, CMA microdialysis, Stockholm, Sweden) was inserted into the subcutaneous adipose tissue under local anaesthetic (2mL of 1% lidocaine) 10cm lateral to the umbilicus. The CMA60 catheters are 0.5mm in diameter and 50mm in length, the distal 30mm consisting of a semi-permeable membrane with a 20kDa cut-off. Following a flush sequence (15μL/min for 5 min), microdialysis was performed at a rate of 0.5μL/min and continued for 30 minutes prior to sample collection (commenced at 09.00h, hourly aliquots collected for 5 hours). Baseline blood samples were taken at 10.00h (t=0mins) and 10mg prednisone (Aventis Pharma ltd, Paris, France) was administered orally. Blood samples were then taken at the following times; t=20, 40, 60, 80, 100, 120, 140, 160, 180, 240mins for biochemical analysis as described below. Cannulae were then removed and the subjects allowed to eat and drink. On the following day, subjects were then re-investigated as described above 2hrs after a single 100mg dose of carbenoxolone (Tokiwa Phytochemicals, Chiba, Japan) taken at 08.00h (2 hours before the start of the cortisol generation profile) and additionally after 72hrs of continuous carbenoxolone treatment (100mg 3 x per day – 08.00h, 14.00h and 22.00h). To confirm inhibition of 11β-HSD activity, 24-hr urine collections were performed during the second full day of CBX treatment (24 to 48hrs), analyzed by GC/MS (see below) and compared with samples collected off all glucocorticoid and CBX treatment (at least 8 weeks). Blood pressure was measured on each day of the protocol (average of three readings, measured supine after 10 minutes rest using Dynamap®, Critikon, Tampa, USA)

Week 2

In order to compare prednisolone generation with the more recognised marker of 11β-HSD1 activity, cortisol generation from oral cortisone, and to show that inhibition of 11β-HSD1 activity can limit GC availability to adipose tissue, subjects were re-admitted to the research facility after an 8-week washout period. Subjects were dexamethasone suppressed (1mg dexamethasone at 23.00h the preceding night and a further dose of 0.5 mg at 09.00h) and at 08.30h, in the fasted state a CMA60 microdialysis catheter was inserted and microdialysis performed as described above. Baseline blood samples were again taken at 10.00h (t=0mins) and an oral dose of 25mg cortisone acetate (Aventis Pharma ltd, West Malling, UK) administered. Blood samples were then taken at the following times, t=20, 40, 60, 80, 100, 120, 140, 160, 180, 240mins for biochemical analysis as described below. As in week 1, investigations were repeated 2hrs after a single dose and after 72hrs continuous treatment with carbenoxolone. A 24-hr urine collection for corticosteroid metabolite was gain taken on the second day of CBX treatments and analysed by GC/MS. Blood pressure was also measured as described above.

Biochemical assays

Serum

Blood counts, urea, creatinine and electrolytes, cholesterol, triglycerides, liver chemistry and plasma glucose were measured using standard laboratory methods (Roche Modular system, Roche Ltd, Lewes, UK). Serum cortisol was assayed using a chemiluminescent immunoassay (Bayer Advia Centaur, Bayer Diagnostics, Newbury, UK) with inter-assay coefficients of variation of 10.2% at 76nmoI/L, 7.7% at 528nmol/L and 7.4% at 882nmol/L. Cortisone was assayed after extraction from serum followed by radioimmunoassay (RIA) of the extract with ¹²⁵I-Cortisone and Sac-Cel® (IDS ltd., Tyne and Weir, UK) second antibody separation. The coefficient of variation for 10 consecutive assays was less than 15% for values between 50 and 100nmol/L and less than 10% for values over 100nmol/L. Prednisolone levels were analyzed by HPLC as described previously (35). Briefly, 100μL of fludrocortisone (40μmol/L) were added to 1ml serum, mixed with dichloromethane (10mL), and the aqueous phase discarded. 1mL of 0.1M sodium hydroxide was added and after mixing, the aqueous phase again discarded. A similar wash step was repeated twice, replacing sodium hydroxide with filtered degassed water. Following addition of 160μL of 1g/L polyethylene glycol (molecular weight 20,000) the dichloromethane layer was evaporated to dryness and resuspended in 200μL HPLC mobile phase. 100μL of aqueous steroid standard or reconstituted extracted steroid were injected onto a 150 × 4.6mm column packed with 5μm ODS-2 (LUNA; Phenomenex, Macclesfield, UK). Mobile phase of tetrahydrofuran/water/methanol in the ratio 23:65:2 and containing 100mmol/L ammonium acetate was pumped at 1.0mL/min. Full resolution of all steroids at absorption at 254nm was achieved. Using this assay, CV for within-day estimates of imprecision were 4.4%. FFAs were measured in serum by the acyl-CoA synthase and acyl-CoA oxidase methods (Wako Chemicals, Neuss, Germany). DHEA was measured using a commercially available conventional 2-site radio-immuno assay (Diagnostic Systems Laboratories Inc., Webster, TX) and DHEAS measured using a coat-a-count radio-immuno assay (Diagnostic Products Corporation, Los Angeles, CA). Both assays were performed as per the manufacturers’ guidelines.

Stock solutions of carbenoxolone (MW 446.36) and working solutions of internal standards (IS, PF-00603176) (0.25μg/mL) were prepared in methanol and stored at - 20°C. Calibration standards (5 to 2500ng/mL) and quality control (QC) samples (10, 300, 1500ng/mL) were prepared by spiking 5μL of the appropriate standard or QC working solutions into 50μL of blank plasma. Aliquots of 0.15mL acetonitrile:methanol (1:1) and 10μL of working IS solution were added to all standards, quality control samples and 50μL of unknown samples. Samples were vortexed for 2 minutes and then centrifuged at 1900 × g (3600 rpm) for 10 minutes at 5°C. 10μL of the supernatant from each tube was then injected into a Sciex API 4000 mass spectrometer equipped with an HPLC system (Agilent 1100 Binary Pump) using a Synergi Hydro-RP column (30 × 2.0mm, 4μ, Phenomenex, CA). The mass spectrometer was operated under the following conditions: electro-spray in positive ion mode, source temperature = 450°C. Data were collected under MRM mode of three ion transactions: m/z 571.3⟶453.2 for carbenoxolone and 272.5⟶135.2 for PF-00603176 (IS), at a dwell time of 150 milliseconds with declustering potential at 126 and 65 eV, collision energy at 27 and 61 eV, and exit potential at 20 and 12 eV, respectively. Due to the limitations in sample volume obtained from microdialysis, hourly samples from individuals across the GC generation profiles were pooled. Pooled samples from individuals were analyzed separately. The lower limit of detection of the assay for these samples was set at 1.0ng/mL (2.2nM).

Microdialysate

Microdialysate samples were collected in microvials and exchanged hourly. Cortisol was measured using a commercially available colourimetric competitive ELISA (R and D systems, Minneapolis, MN). The minimum detectable dose range for the assay was 0.1-0.3 nmol/L with intra-assay CVs of 6-9%.

Samples from week 1 were analyzed using a mobile photometric, enzyme-kinetic analyzer (CMA 600, Sweden) for concentrations of glucose, pyruvate, lactate, and glycerol. Carbenoxolone in microdialysate samples was measured by an HPLC method as described above.

Urinary corticosteroid metabolites

Urinary corticosteroid metabolite analysis was performed by GC/MS as described previously (36). The sum of total cortisol metabolites (THF (tetrahydrocortisol), THE (tetrahydrocortisone), 5α-THF, α-cortolone, cortisone (E), cortisol (F), β-cortolone, β-cortol, α-cortol) provides a reflection of cortisol secretion rate. More specifically, the ratio of tetrahydro-metabolites of cortisol (THF + 5αTHF) to those of cortisone (THE) provides a reflection of 11β-HSD1 activity when considered with the ratio of urinary free cortisol (UFF) to cortisone (UFE) which more accurately reflects renal 11β-HSD2 activity. The activities of 5α and 5β-reductases can be inferred from measuring the ratio of THF:5αTHF.

Statistical analysis

Power and sample size calculations were performed as part of the study design and based upon predicted changes in urinary steroid metabolite ratios. Data are presented as mean±S.E. unless otherwise stated. Area under the curve (AUC) analysis was performed using the trapezoidal method. For comparison of single variables between weeks one and two of the clinical protocol, paired t-tests have been used. Where repeated samples have been taken (either during an individual investigation or for comparison of the same investigation during a single week of the clinical protocol) repeated measures ANOVA has been used. All analysis was performed using the SigmaStat 3.1 software package (Systat Software, Inc. Point Richmond, CA).

Results

Baseline characteristics and biochemical parameters before and during carbenoxolone treatment are presented in Table 1.

Table 1.

Clinical, anthropometric and biochemical characterization of 7 healthy male subjects before, 2hrs after a single dose of carbenoxolone (CBX, 100mg), and after 72hrs of CBX treatment (300mg per day). The investigations were repeated on 2 separate occasions separated by at least 8-weeks (week 1 and week 2)

	Week 1			Week 2
	Baseline	2hrs post single dose CBX (100mg)	4 days CBX (300mg/d)	Baseline	2hrs post single dose CBX (100mg)	4 days CBX (300mg/d)
Variable (local reference range)
Age (years)	28±3
Height (m)	1.8±0.02
Weight (kg)	78.9±2.7
BMI (kg/m2)	24.5±1.3
Systolic BP (mmHg)	127±6	116±7	126±6	137±7	137±13	127±3
Diastolic BP(mmHg)	68±5	60±3	71±3	77±5	70±4	71±3
Na (134-146mmol/L)	141±0.9	139±0.8	140±0.8	139±1	139±1	139±1
K (3.4-5.2mmol/L)	4.0±0.2	3.9±0.1	3.8±0.1 c	4.2±0.1	4.2±0.1	4.2±0.1
Urea (3.2-7.6mmol/L)	4.8±0.6	4.6±0.5	3.8±0.3 a,b	4.8±0.6	4.7±0.6	4.4±0.7
Creatinine (60-126μmol/L)	96±4.6	99±3.2	95±4.4	95±5	92±5	90±4 a
Glucose (mmol/L)	4.2±0.2	4.4±0.1	4.6±0.1	4.6±0.2	4.7±0.3	4.8±0.2
Cholesterol (mmol/L)	4.6±0.8	4.6±0.8	4.3±0.7	4.9±0.8	4.8±0.9	4.8±0.8
TG (mmol/L)	1.4±0.5	1.1±0.3	1.0±0.3	0.9±0.4	1.2±0.4	1.0±0.4
Alkaline phosphatase (70-320U/L)	159±22	152±21	142±17	153±21	146±23	145±21
Asp. transaminase (5-43U/L)	22±3	22±3	26±5	26±5	23±3	34±7 a,b
Bilirubin (1-22μmol/L)	18±8	16±6	15±5	17±8	17±8	12±5
Haemoglobin (13.5-18.0g/dL)	14.7±0.6	14.8±0.9	13.4±0.4	14.3±0.5	13.9±0.4	13.3±0.4 a,b
WCC (4.0-11.0 x109/L)	5.4±0.7	5.4±0.6 c	4.8±0.5	6.1±0.9	6.6±0.3	5.8±0.7
Platelets (150-450 x109/L)	222±24	200±13	193±16	192±16	203±19	205±20
DHEAS (2.2-15.2μmol/L)	5.7±0.6	5.0±0.4	5.2±0.4
DHEA (4.9-43.0nmol/L))	28.7±4.0	23.8±2.7	27.7±2.0
Cortisol (180-550nmol/L)	410±51	340±38	412±36
Carbenoxolone (μg/mL)		8.8±1.1	42.9±4.2 b

^ap<0.05 vs. baseline

^bp<0.05 vs. single dose

^cp<0.05 vs. week 2

Analysis of urinary corticosteroid metabolites by GC/MS including absolute values are presented in table 2. Urine collections on CBX were performed 24-48hrs after the administration of glucocorticoid and therefore are unlikely to have been influenced by cortisone, dexamethasone or prednisone administration. Total cortisol metabolite production decreased following CBX administration. Consistent with inhibition of 11β-HSD2, the UFF/UFE ratio increased significantly. Despite this, the THF+5αTHF/THE ratio decreased following CBX treatment as did the ratio of cortols/cortolones consistent with inhibition of 11β-HSD1 activity. There was no alteration in the activities of 5α or 5β-reductase. Results of the urinary GC/MS analysis after CBX treatment were similar in weeks 1 and 2 (table 2).

Table 2.

Urinary corticosteroid metabolite analysis performed by GC/MS on 24hr urine samples from 7 healthy male volunteers before and after 48hrs continuous CBX treatment (300mg/day). The investigations were repeated on 2 separate occasions separated by at least 8-weeks (week 1 and week 2).

		Week 1	Week 2
	Baseline	48hr CBX (300mg/d)	48hr CBX (300mg/d)
Corticosteroid metabolites (μg/24hrs)
Total cortisol metabolites	12535±2686	8942±995 a	6106±1084ab
THE	4200±743	3161±517	2575±464
5a THF	2025±480	823±187a	709±193a
THF	2469±117	1261±239a	1139±269a
UFE	150±32	95±14	62±14
UFF	76±16	84±16	48±10
α-cortolone	1699±172	892±166a	642±63 a
β-cortolone	887±83	769±129	533±54 a
α-cortol	509±63	183±33a	155±29a
β-cortol	696±110	302±65a	243±58 a
Corticosteroid metabolite ratios
UFF:UFE	0.52±0.05	0.84±0.06 a	0.78±0.05 a
(THF+5αTHF)/THE	0.98±0.12	0.66±0.15 a	0.67±0.06 a
THF/5αTHF	1.93±0.72	1.72±0.35	2.04±0.43
Cortols/cortolones	0.44±0.06	0.30±0.03 a	0.32±0.04 a

^ap<0.05 vs. baseline

^bp<0.05 vs. week 1

Prednisolone and cortisol generation

Following oral administration of either prednisone (10mg) or cortisone acetate (25mg), all subjects were able to generate prednisolone and cortisol respectively (figure 1). A single dose of CBX decreased prednisolone generation (AUC: 1671±64 (baseline) vs. 1491±61nmol/L.hr (CBX), p<0.05), however cortisol generation was not different (AUC: 1617±125 (baseline) vs. 1499±103nmol/L.hr (CBX), p=ns). After 72hrs CBX treatment, inhibition of prednisolone generation was maintained (AUC: 1443±80nmol/L.hr p<0.05 vs. baseline), but cortisol generation decreased significantly (AUC: 1203±76nmol/L.hr, p<0.05 vs. baseline) (figure 1). Peak prednisolone concentrations did not decrease after a single dose of CBX (574±26 vs. 554±42nmol/L, p=ns), but fell following 72hrs of CBX administration (475±24nmol/L, p<0.05). Similarly, a single dose of CBX did not effect peak cortisol concentrations, but these fell after 72hrs of CBX treatment (peak cortisol: 579±25 (baseline) vs. 546±33 (single dose CBX) p=ns, 465±21nmol/L (72hrs CBX), p<0.05).

Figure 1.

Clinical study protocol. Week 1 and 2 protocols were performed in 7 healthy adult male volunteers who were on no regular medications and had not received GC therapy within the last 12 months. Weeks 1 and 2 of the protocol were separated by at least 8 weeks, and in the final week of this wash-out period, a repeat 24hr urine collection was performed off all medication.

Cortisone concentrations following dexamethasone suppression were undetectable and rapidly increased following oral cortisone acetate (36±13, t=20mins; 74±10, t=40mins; 72±10, t=60mins; 78±7, t=120mins; 57±4nmol/L, t=240mins). Cortisone appearance in the serum was similar before and after single dose of CBX (AUC cortisone: 241±27 vs. 223±27nmol/L.hr (single dose CBX), p=ns), but decreased after 72hrs of treatment 167±19nmol/L.hr, p<0.005 vs. baseline and p<0.01 vs. single dose) consistent with inhibition of renal 11β-HSD2 limiting cortisone re-generation from cortisol.

Adipose tissue interstitial fluid cortisol concentrations (as measured by adipose microdialysis) were low, but detectable within the limits of the assay following dexamethasone suppression (0.5±0.2nmol/L). Following oral administration of cortisone acetate, cortisol concentrations in adipose tissue interstitial fluid increased rapidly (1.7±0.5, t=60mins; 6.6±1.0, t=120mins; 8.0±2.0, t=180mins; 6.7±2.3nmol/L, t=240mins) (figure 2A). Both a single dose and 72 hrs of treatment with CBX decreased cortisol concentrations following oral cortisone within adipose tissue interstitial fluid (peak cortisol: 8.9±2.1 (baseline), 5.6±1.1 (single dose CBX), 5.3±0.6nmol/L (72 hrs CBX, p=0.07). In addition, AUC for cortisol generation fell significantly (AUC: 19.9±3.7 (baseline), 11.2±1.9 (single dose CBX), p<0.005, 10.9±1.5nmol/L.hr (72 hrs CBX), p<0.01) (figure 2A and B) indicative of inhibition of 11β-HSD1 activity.

Figure 2.

A single dose of carbenoxolone, 100mg (grey triangles, hashed bars), and 72hrs treatment, 300mg/day (filled squares and bars), decrease cortisol generation from oral cortisone acetate, 25mg (A. generation profile and B. Area under the curve analysis). Similarly, prednisolone generation form oral prednisone, 10mg is also decreased (C. generation profile and D. AUC analysis) (baseline samples, white circles and bars).

Carbenoxolone concentration in serum and interstitial fluid

Serum CBX concentrations as measured by HPLC were detectable 2hours after a single 100mg dose (19.6±2.5μmol/L), but fell significantly within 4hours (10.3±1.6μmol/L, p<0.005). After 72hrs of CBX treatment, CBX levels increased significantly (96.2±9.4μmol/L p<0.001 vs. single dose) and showed no decrease within 4 hours of a morning dose (103.6±10.7μmol/L, p=ns). Although concentrations were >5000-fold less than those observed in serum, we were able to measure CBX within adipose tissue interstitial fluid. Following a single dose of CBX, the concentration within adipose interstitial fluid was at the lower limit of detection of the CBX assay (2.0±1.0nmol/L). However, after 72hrs of CBX treatment, levels increased to 19.7±15.0nmol/L, well within the detectable range and in excess of the molar concentrations of cortisol. Prior to CBX administration CBX levels were undetectable.

Glucocorticoid induced lipolysis

Prednisone administration (week 1) caused a significant rise in serum FFA levels in all subjects: (73±18, t=0mins vs. 175±35μmol/L, t=240mins, p<0.05). Whilst we were unable to detect a difference in serum FFA concentrations following CBX administration (either single dose or 72hrs treatment), peak prednisolone concentration (both on and off CBX) correlated positively with FFA generation (R=0.49, p<0.05).

Adipose tissue interstitial fluid glycerol concentrations significantly increased following oral prednisone (63.0±23.5, t=0mins vs. 223.9±24.0μmol/L, t=240mins, p<0.01). Both a single dose, and 72hr treatment with CBX decreased glycerol concentrations following oral prednisone (peak glycerol: 240.9±23.1 (baseline), 193.3±17.3 (single dose CBX), p<0.05, 174.4±29.1μmol/L (72hr CBX), p<0.05). In addition, AUC for glycerol release decreased (AUC: 700.9±63.2 (baseline), 585.2±54.9 (single dose CBX), 534.5±91.1μmol/L.hr (72hrs CBX), p=0.06) (figure 2C and D). There were no significant changes in adipose tissue interstitial fluid concentrations of glucose, lactate or pyruvate (data not shown).

Discussion

In this paper we have shown that CBX inhibits the generation of bioactive cortisol and prednisolone through inhibition of 11β-HSD1. Furthermore, inhibition can be measured not only in serum, but also in adipose tissue interstitial fluid as well as in urine. Finally, and most importantly, decreasing GC availability is functionally important in human adipose tissue, limiting GC induced lipolysis.

The changes in serum biochemistry during this study were small. However, the decrease in urea and potassium during week 1 (and not week 2) may reflect the additional mineralocorticoid action of prednisolone with consequent fluid retention. Although bioequivalent doses of cortisol and prednisolone have similar potencies for MR activation, the 10mg dose of prednisone and 25mg of cortisone acetate are not bioequivalent; indeed the former represents an approximate doubling of mineralocorticoid action. This mechanism may also contribute to the observed decrease in haemoglobin concentration (haemodilution), but the contribution of repeated blood sampling over a relatively short period of time maybe a factor.

Urinary corticosteroid metabolite analysis has traditionally been the most widely used method for the accurate clinical assessment of 11β-HSD activity. In previous studies, CBX increased the UFF:UFE ratio by 53-102% reflecting inhibition of renal 11β-HSD2 (37;38) and our data (45-62%) are consistent with this; the decrease in the THF+5αTHF:THE ratio is less marked (28-34%), but again our values are comparable (32-33%). Inhibition of 11β-HSD2 by CBX will inhibit the conversion of cortisol to cortisone and therefore selective 11β-HSD1 inhibitors are likely to cause a more significant decrease in the THF+5αTHF:THE ratio. Overall, the data presented validate the use of urinary GC/MS to assess 11β-HSD1 inhibition.

11β-HSD1 and 2 are able to interconvert cortisol and cortisone as well as prednisone and prednisolone and activity of both enzymes is inhibited by GE (39-42). Using in vitro systems, enzyme kinetics are similar for both substrates (39). During the cortisol generation profile (week 2), subjects have endogenous cortisol production suppressed using dexamethasone. GCs are potent up-regulators of 11β-HSD1 activity and expression (43) and as a result baseline measurements of activity may be artificially high. CBX not only acts as a competitive inhibitor of 11β-HSD1 but also decreases mRNA expression (21), and this may counteract the GC induction on subsequent days. Inhibition of 11β-HSD2 prolongs cortisol and prednisolone half-life (44) and therefore would increase total cortisol and prednisolone generation. In agreement with this, we observed decreased total cortisol metabolite production following CBX treatment, suggestive of a predominant action of CBX upon 11β-HSD2. Selective 11β-HSD1 inhibitors would be predicted to have a more dramatic impact upon these profiles in comparison with CBX and as such the generation of both prednisolone and cortisol may be useful biomarkers.

CBX has been used in a small number of studies to determine the clinical impact of 11β-HSD inhibition. These studies have focussed principally on organ systems known to express 11β-HSD1 (liver, fat, brain, eye, bone). In healthy adults, CBX improves whole body insulin sensitivity (24) and in patients with type 2 diabetes glucose production rates are decreased through a reduction in glycogenolysis (25). In healthy volunteers, CBX decreases intra-ocular pressure (37) and more recent studies have demonstrated improvements in cognitive function in elderly men and those with type 2 diabetes (45). As an agent that causes acquired apparent mineralocorticoid excess through its inhibition of 11β-HSD2, CBX is not viable as a long-term therapeutic strategy and much attention has now been focussed upon the development of selective 11β-HSD1 inhibitors (46).

Due to the emerging body of evidence suggesting a role for 11β-HSD1 in the pathogenesis of obesity, metabolic syndrome and insulin resistance, rodent studies have used selective inhibitors in models of these conditions. The results have been impressive with dramatic reductions in insulin resistance, improvements in glucose tolerance and lipid profiles, and decreased atherogenesis (17-19). To-date however, there are no published data on clinical studies in humans.

A key concept underpinning pre-receptor GC metabolism through 11β-HSDs is tissue specificity of response. Within adipose tissue interstitial fluid, a single dose of CBX is as effective as 72hrs continuous treatment at limiting cortisol availability. Compared with the parallel serum measurements, which are believed to most accurately reflect hepatic activity (10;11), inhibition of activity was more marked in the microdialysate samples (45% vs. 25%). Furthermore, cortisol production in the microdialysate was inhibited after a single dose of CBX when serum cortisol was unaffected until 72hr treatment. This disconnect in the inhibition measured in microdialysate and serum is consistent with local inhibition in adipose. This hypothesis is further strengthened by the observation that we were able to measure CBX within adipose tissue interstitial fluid. Although CBX concentrations in the serum were 5,000 fold greater than those in adipose tissue interstitial fluid, these concentrations are still in excess of interstitial fluid GC concentrations, and therefore functionally important competitive inhibition seems plausible. However, the current study does not allow us to categorically determine whether the changes observed in adipose tissue represent local inhibition or are a reflection of changes in circulatory GC levels. Interestingly as adipose tissue does not express 11β-HSD2 (43), it may also provide the ‘cleanest’ system for looking at 11β-HSD1 inhibition by CBX. A single previous study has used adipose microdialysis to examine cortisol metabolism in adipose tissue (26). The methodology employed was different, and tritiated substrate was introduced via the microdialysis catheter. In our study, microdialysis was simply used to sample adipose tissue interstitial fluid to determine the expose of adipocytes to extra-cellular GC before and after CBX treatment. Although differences were observed between lean and obese individuals in the previous study, inhibition with CBX could not be demonstrated (26). Substrate delivery, metabolism and product recovery by this methodology is extremely complex and subject to a large number of important variables including steroid distribution within lipid and tissue blood flow. The authors’ conclusion was that CBX could not access adipose tissue interstitial fluid, however, our data suggests that this now seems unlikely.

The action of GC to promote lipolysis is well described both in vitro and in vivo (28;47). We have been able to show for the first time that activity of 11β-HSD1 is a critical regulator of GC induced lipolysis in vivo. Whilst peak prednisolone following oral prednisone correlated significantly with FFA release in to the serum, we were unable to show differences in FFA generation before or after CBX. This may well reflect rapid re-esterification of released FFA, but may also reflect the fact that changes in serum prednisolone (before and after CBX) were small (albeit significant) perhaps reflecting concomitant inhibition of renal 11β-HSD2. The adipose tissue microdialysis data, using prednisone as an 11β-HSD1 substrate reveal that CBX decreased peak glycerol release indicative of inhibition of GC mediated lipolysis, although the AUC analysis achieved only a borderline significant result (p=0.06). Unfortunately, due to the small volumes of microdialysate samples and lack of a sensitive and specific assay method, direct measurement of interstitial fluid prednisolone was not possible. The role of FFA in the control of insulin sensitivity is well described (48), and it is possible that the improvement in insulin sensitivity in rodents treated with selective 11β-HSD1 inhibitors may arise at least in part as a result of decreased adipose tissue derived FFA. In addition, FFA have been reported to regulate 11β-HSD1 in one study (49) although not in another (50).

To conclude, urinary GC/MS, serum prednisolone as well as serum and interstitial adipose cortisol generation are useful biomarkers of 11β-HSD1 inhibition. At a tissue specific level, limitation of GC availability to human adipose tissue through inhibition of 11β-HSD1 has functional consequences including decreased lipolysis.

Figure 3.

A single dose of carbenoxolone, 100mg (grey triangles, hashed b0ars), and 72hrs treatment, 300mg/day (filled squares and bars) limit cortisol availability within adipose tissue interstitial fluid following oral cortisone acetate, 25mg as measured by microdialysis (A. generation profile and B. AUC analysis). Following oral prednisone, 10mg, glycerol release into adipose tissue interstitial fluid increases (C). CBX treatment (single dose, grey triangles, hashed bars and 72hrs filled squares and bars) decreases prednisone induced glycerol release. (C. glycerol release, D. AUC analysis) (baseline samples, white circles and bars).