Human receptor kinetics and lung tissue retention of the enhanced-affinity glucocorticoid fluticasone furoate

Fluticasone furoate (FF) – USAN approved name, a new topically active glucocorticoid has been recently identified. The aim of this study was to characterise the binding affinity of this compound to the human lung glucocorticoid receptor in relation to other glucocorticoids. Additionally, we sought to determine the binding behaviour of fluticasone furoate to human lung tissue. The glucocorticoid receptor binding kinetics of fluticasone furoate revealed a remarkably fast association and a slow dissociation resulting in a relative receptor affinity (RRA) of 2989 ± 135 with reference to dexamethasone (RRA: 100 ± 5). Thus, the RRA of FF exceeds the RRAs of all currently clinically used corticosteroids such as mometasone furoate (MF; RRA 2244), fluticasone propionate (FP; RRA 1775), ciclesonide's active metabolite (RRA 1212 – rat receptor data) or budesonide (RRA 855). FP and FF displayed pronounced retention in human lung tissue in vitro. Lowest tissue binding was found for MF. There was no indication of instability or chemical modification of FF in human lung tissue. These advantageous binding attributes may contribute to a highly efficacious profile for FF as a topical treatment for inflammatory disorders of the respiratory tract.


Background
A new topically active glucocorticoid, fluticasone furoate (FF, GW685698X), has been recently identified ( Figure 1) and is being progressed for the treatment of respiratory diseases. Fluticasone furoate (FF) shares structural similarities with fluticasone propionate (FP) with the exception of the substitution of the 17-α hydroxyl group. While this position is esterified with propionic acid in FP, FF carries a 2-furoate ester moiety.
For topically applied glucocorticoids, it is favorable to combine high local efficacy with low systemic exposure. An enhanced affinity for lung tissue may prolong residence time in the lung and minimise systemic effects. Therefore, a high receptor affinity and a high retention in the target tissue should be paralleled by rapid and complete hepatic metabolism of the glucocorticoid to inactive derivatives. We previously described the receptor binding affinity of FP and MF as well as their retention in lung tissue in vitro [1][2][3][4]. Both FP and MF have high affinities for the human lung glucocorticoid receptor. The relative receptor affinity (RRA) of FP is about 1800 compared to the reference compound dexamethasone (RRA= 100), the RRA of MF is about 2250.
The aim of this study was to characterise the binding affinity of the novel compound FF to the glucocorticoid receptor in relation to other glucocorticoids. Therefore, we isolated human lung glucocorticoid receptors from human lung tissue and determined the binding affinity of Plasma samples were obtained from healthy volunteers who gave informed consent. Samples were used immediately for metabolism studies to retain full enzymatic activity. For desorption and other experiments, plasma samples were shock frozen in liquid nitrogen and stored at -70°C until usage.

Preparation of lung cytosol for receptor binding experiments
Human lung tissue was deep frozen immediately after resection and stored in liquid nitrogen. Frozen tissue was pulverized and homogenized in three aliquots buffer solution A with an Ultra Turrax mixer (Janke and Kunkel, Staufen, Germany) in an ice bath. Thereafter the diluted cytosol was centrifuged for 1 hr at 105,000 × g at 4°C (Ultracentrifuge L8-55 M, Beckman Instruments Irvine, California). The cytosol was stored in aliquots at -70°C. The protein concentration of the cytosol was determined according to the method of Lowry et al. [6]. Concentration of glucocorticoid receptors in the cytosol was 30-60 fmol/mg protein.

Kinetics of receptor binding of glucocortiocids
The receptor binding experiments were performed according to the procedure described earlier [1] based on [7][8][9].  compound were added to 200 µL of cytosol, were mixed in glass vials and incubated for 18 to 20 h at 0-4°C. The assay for total binding (B t in [mol/L]) was carried out accordingly, but the unlabelled glucocorticoid was replaced by buffer solution G. To determine the total [ 3 H]glucocorticoid concentration (T), 20 µL of the mixture were used for scintillation counting. After incubation, 200 µL of each incubation mixture were added to 200 µL suspension of activated charcoal (2 % Norit A in buffer solution G), incubated for 10 min on ice and centrifuged for 5 min between 0-4°C. For scintillation counting 200 µL of the supernatant were used. Scintillation counting was performed with a Rackbeta 1214 LKB from Wallac (Freiburg, Germany) using Emulsifier-Safe™ from Packard Bioscience (Groningen, Netherlands).
Receptor concentration (R 0 ) of the cytosol was calculated by the method of Scatchard [10] according the equation: with B S being the specific binding of the labelled dexamethasone in [mol/L], H being the unbound labelled glucocorticoid, and K D being the equilibrium dissociation rate constant. B S and H were indirectly determined using the equations: The Scatchard plot revealed the equilibrium dissociation rate constant K D (slope of the straight line) and the receptor number R 0 in mol receptors per mg total protein of the cytosol (interception of the straight line with the x-axis).

B. Determination of association rate constants k Ass (= k 1 )
For the determination of the association rate constant, the cytosol was incubated with different concentrations of [ 3 H]-glucocorticoid in the absence and presence of excess unlabelled glucocorticoid. For the assay of non-specific binding, 10 parts of cytosol, 1 volume part of [ 3 H]-glucocorticoid (1.2 × 10 -7 mol/L) and 1 volume part of cold glucocorticoid (1.2 × 10 -4 mol/L) were mixed in glass vials and incubated at 20°C. The assay for total binding was carried out accordingly, but the unlabelled glucocorticoid (1.2 × 10 -4 mol/L) was replaced by buffer G. To determine the total [ 3 H]-glucocorticoid concentration, aliquots of the incubation mixtures were used for scintillation counting. At intervals, 200 µL incubation mixture were mixed with 200 µL suspension of Norit A, incubated for 10 min on ice and centrifuged for 5 min between 0-4°C. For scintillation counting 200 µL of the supernatant were used.
The association rate constant (k Ass = k 1 ) of the cytosol was calculated according the equation: with G t being the concentration of unbound labelled glucocorticoid at time t, R t being the concentration of free receptors at time t, G 0 being the concentration of unbound labelled glucocorticoid at time t = 0, R 0 being the concentration of free receptors at time t = 0 and t being the time of incubation. G 0 and G t were indirectly determined using the equations: To linearize the calculated data points a Z t -value was calculated for each time point of measurement taking the dilution factor of the cytosol and the receptor concentration into account: The Z t -values were plotted against time t and a linear regression was performed. The slope of the straight line (k Ass = k 1 ) and the coefficient of correlation r were calculated based on a minimum of four data points. The coefficient of correlation was always higher than r = 0.975.

C. Determination of dissociation rate constants k Diss (= k -1 )
For determination of the dissociation rate constant, 10 volume parts cytosol and 1 volume part [ 3 H]-glucocorticoid solution (6 × 10 -7 mol/L) were incubated for 18-20 h between 0-4°C (mixture 1). To determine the non-specific binding, 10 volume parts of cytosol, 1 volume part of [ 3 H]-glucocorticoid solution (6 × 10 -7 mol/L) and 1 part of unlabelled glucocorticoid (3 × 10 -4 mol/L) were incubated for 18-20 h between 0-4°C (mixture 2). Incubation mixtures were subsequently brought to a temperature of 20°C. One volume part of unlabelled glucocorticoid (3 × 10 -4 mol/L) was added to mixture 1. At intervals 200 µL each of the mixtures 1 and 2 were mixed with 200 µL Norit A suspension, incubated at 0-4°C for 10 min and thereafter centrifuged for 5 min at 0-4°C. The supernatant was used for scintillation counting. The first order rate constant was calculated according: with B s,t being the specific binding of the labelled compound at time t, B s,0 being the specific binding of the labelled compound at time t = 0. Since the specific bind- ing was determined indirectly (see A) the equation can be rewritten as: The B s,t -values were plotted semi-logarithmical against time t and a linear regression was performed. The slope of the straight line (k Diss = k -1 ) and the coefficient of correlation r were calculated based on a minimum of six data points. The coefficient of correlation was always higher than r = 0.975. Equilibrium dissociation constant (K D ) was calculated for each glucocorticoid based on association and dissociation rate constants: Relative receptor affinities (RRA) for glucocorticoids (GC) were calculated with reference to dexamethasone (Dexa):

Stability of fluticasone furoate (FF) in fresh human lung tissue in vitro
Fluticasone furoate (FF) (0.3 µg/mL) was incubated in 10 ml Krebs-Ringer-HEPES buffer with lung tissue pieces at 37°C shielded from light in a thermostatically controlled shaking water bath GFL 1083 (Burgwedel, Germany). Incubations were performed in the presence and absence of dichlorvos (1 mg/mL). Over 24 hours, samples of 1.0 mL tissue-free supernatant were taken and immediately stored at -20°C until analysis. The incubation medium was replenished by buffer which was pre-temperated to 37°C. In case of incubations with dichlorvos the medium used for replenishment contained the esterase inhibitor.

Adsorption of glucocorticoids to lung tissue
Lung tissue was washed in Krebs-Ringer-HEPES buffer (pH 7.4) and sliced into pieces of 1 mm 3 . For each binding experiment approximately 0.5 g of lung tissue was used. Adsorption of glucocorticoids (0.3 µg/mL) to human lung tissue was determined as described earlier [3]. Briefly, lung tissue pieces were suspended under gentle shaking for 1 h at 37°C in 20 ml Krebs-Ringer-HEPES buffer containing 0.3 µg/ml of the glucocorticoid. 2.0 mL samples were taken and stored at -20°C until analysis. The volume withdrawn was replaced with fresh buffer of 37°C. Only glass lab ware was used for these experiments to avoid any non-specific binding effects of the highly lipophilic compounds to plastic material. For control, blank samples with glucocorticoid-containing buffer, but no tissue, were incubated under the same experimental conditions (1 h at 37°C, in Krebs-Ringer-HEPES buffer) and analyzed for non-specific adsorption of the glucocorticoids to the glass tubes.

Desorption of glucocorticoids from lung tissue
Desorption of glucocorticoids to human lung tissue was determined as described earlier [3]. Briefly, lung tissue (1.0 g) was saturated with glucocorticoids for 1 h at 37°C by shaking in 40 mL Krebs-Ringer-HEPES buffer containing 0.3 µg/mL of the respective glucocorticoid. After incubation tissue was washed with 2 mL buffer and transferred into 10.0 mL human plasma (37°C). Again, only glass lab ware was used for these experiments to exclude any nonspecific binding effects of the highly lipophilic compounds to plastic material. Samples of 1.0 mL were taken at defined time points. The volume was replaced with fresh plasma at 37°C. Samples were stored at -20°C until further analysis.

Sample preparation, HPLC conditions and data analysis
Samples of 1.0 mL (tissue desorption/stability) or 2.0 mL (tissue adsorption) were mixed with 0.1 mL internal standard solution and extracted twice with 3 mL diethylether for 30 min, using a roller mixer, followed by centrifugation (20°C) for 5 min. The organic phase was separated and evaporated to dryness under a gentle stream of nitrogen at 25°C. The resulting residue was reconstituted in 0.2 mL mobile phase. Internal standard (IS) was amcinonide 3 µg/mL (tissue binding studies) or dexamethasone 3 µg/mL (stability studies). Linearity was given from 10-500 ng/mL glucocorticoid, coefficients of correlation of the calibration curves were at least 0.99.
The HPLC system was a Waters HPLC (Milford, MA) consisting of a 1525 binary pump, an 717plus autosampler and 2487 dual wavelength absorbance detector set at the detection wavelength of 254 nm. Data collection and integration were accomplished using Breeze™ software version 3.2. Analysis was performed on a Symmetry C 18 column (150 × 4.6 mm I.D., 5 µm particle size, Waters, MA). Typically, 20 µL of sample were injected and separated at a flow rate of 1 mL/min. Gradient elution was performed using water (containing 0.2 % (v/v) acetic acid) and ACN, starting at 60:40 (v/v) water/ACN increasing linearly to 29:71 (v/v) water/ACN by 30 min. The assay was accurate and reproducible. The lower limit of quantitation was 10 ng/mL for all glucocorticoids except ciclesonide (20 ng/mL).

Determination of the relative retention time k' of glucocorticoids
Relative retention times k' or chromatographic capacity factors log (k'), respectively, of all new generation glucocorticoids in comparison with older glucocorticoids were determined by a HPLC method based on a former report × 100 [5]. Briefly, to calculate k' the HPLC retention time on a C 18 reversed-phase column of an individual glucocorticoid was related to the retention time of an internal standard (dexamethasone-21-isonicotinate). Therefore, 10 µL of the respective glucococorticoid and the internal standard at a concentration of each 10 µg/mL in methanol were chromatographed under identical conditions (column and HPLC system described above). The sample was injected and separated at a flow rate of 0.7 mL/min. The mobile phase consisted of methanol, water, ACN and acetic acid at 40:20:5:0.2 (v/v).

Statistical analysis
Mean and mean deviation of the mean were calculated for all data. Data sets were analysed by one-way ANOVA with post-hoc Bonferroni's multiple comparison test. Statistical significance was defined as a significance level of p ≤ 0.05. Due to the very limited sample number a pre-test was performed to test the normal distribution of the residuals. Therefore, the residuals of each data group were calculated and the ratio of range to standard deviation was analysed according to David et al. [11]. Only when the results were between the lower and upper critical limits tabulated by Pearson and Stephens [12] a normal distribution of the residuals was assumed at a significance level of p ≤ 0.05 and a subsequent ANOVA analysis was performed. On one data set a reciprocal transformation was performed for normal distribution of the residuals and subsequent ANOVA analysis. Due to the limited number of data p values should be interpreted very cautiously.

Receptor binding kinetics and relative receptor affinity of fluticasone furoate (FF)
The receptor binding kinetics to the human lung glucocorticoid receptor revealed that the association kinetics of fluticasone furoate (FF) was distinctly different from those of fluticasone propionate (FP) and mometasone furoate (MF) (

Correlation between glucocorticoid lipophilicy and receptor affinity
The chromatographic capacity factor log (k') reveals an excellent correlation to the partition coefficient in 1-octanol-water [13,14] which is regarded as a typical parameter of compound lipophilicity. When the lipophilicity of a glucocorticoid is expressed as its relative retention time k' at a reversed-phase HPLC column and correlated with the relative receptor affinity of the respective compound, a significant relationship is observed (Figure 2). Potential fitting of the data according to the equation: y = c * x b .
(with c and b representing constants) revealed a coefficient of correlation of r = 0.982. This relationship is statistically significant (p < 0.0001). All glucocorticoids esterified at C21 display higher lipophilicity. However, these compounds have little or no binding affinity to the glucocorticoid receptor. They are either inactive metabolites such as beclomethasone-21-monopropionate (21-BMP) or inactive pro-drugs such as ciclesonide or beclomethasone-17,21-dipropionate which need to be activated by hydrolysis of the C21 ester [15,16].

Stability of FF in freshly isolated human lung tissue
The stability of FF in the presence in human lung tissue was monitored over a period of 24 hours at an incubation temperature of 37°C ( Figure 3). The incubations were performed in the presence and absence of the esterase inhibitor diclorvos. Indications of instability of the compound are either decreased compound concentrations in the supernatant in the absence of dichlorvos and/or the appearance of new peaks in the HPLC chromatograms. The initial decrease of FF concentration indicated the binding to the lung tissue pieces. Over the incubation period, concentrations of FF in the tissue supernatant were slightly higher in the absence of dichlorvos. No new peaks were observed in the HPLC chromatograms. No statistically significant differences were revealed between concentrations of FF in the presence and absence of the esterase inhibitor diclorvos at any of the single time points. Thus, non-specific esterase-catalyzed hydrolysis of FF did not occur in the presence of human lung tissue.
The enzymatic integrity of the lung tissue was demonstrated in a simultaneously performed control experiment with beclomethasone-17,21-dipropionate (BDP). The results of these control experiments were identical to those described previously [3]. In the absence of dichlorvos BDP concentrations in the supernatant rapidly decreased and the main metabolite beclomethasone-17monopropionate (17-BMP) was detectable at high concentrations. Dichlorvos inhibited the decomposition of BDP and delayed the formation of 17-BMP up to 10 hours of incubation (data not shown).

Lung tissue binding affinity of fluticasone furoate (FF)
The binding affinity of FF in comparison with MF and FP to human lung tissue was determined in separate adsorption and desorption experiments. Control experiments for non-specific binding to incubation vials were performed in parallel with the respective glucocorticoid-containing buffer solutions under identical conditions. These control experiments revealed no non-specific binding of FF or FP to the glass incubation vials (Figure 4). A decrease in MF concentrations over 480 min at 37°C was paralleled by formation of the degradation product 9,11-epoxy MF as described earlier [3]. Thus, MF did not display non-specific binding to glass, but did show chemical instability.
Adsorption of FF to human lung tissue in vitro occurred rapidly and was complete after about 20 min (data not shown). After 60 min incubation with the glucocorticoidcontaining buffer at 37°C highest tissue binding was seen for FF (4.18 ± 0.16 ng/mg) and this was statistically significantly higher compared to FP (3.39 ± 0.06 ng/mg; p ≤ 0.001) and MF (3.65 ± 0.15 ng/mg; p ≤ 0.01) ( Figure 5, left columns). FF also showed greater binding to human  nasal tissue compared with FP in a single experiment with tissue pooled from 3 donors (data not shown).
The desorption of the glucocorticoids from lung tissue into human plasma revealed differences between the compounds ( Figure 5, right columns). After 60 min highest concentrations of FP (1.55 ± 0.13 ng/mg) and FF (1.21 ± 0.23 ng/mg) were still present in the tissue. Remaining concentrations of FP and FF were not statistically significantly different. As reported previously [3], MF was rapidly redistributed from the lung tissue into human plasma and consequently lowest concentrations of mometasone furoate were detected in the tissue (0.57 ± 0.15 ng/mg). This was statistically significantly lower compared to FF (p ≤ 0.01) and FP (p ≤ 0.001).

Discussion
Fluticasone furoate (FF) is a newly developed glucocorticoid for topical application. In the present investigation we characterized the receptor binding kinetics and the binding affinity to human lung tissue of FF in comparison with other latest generation glucocorticoids. We found that FF exhibited the highest ever described relative receptor affinity (RRA) of a topical glucocorticoid. The RRA of FF (2989 ± 135) exceeds the receptor affinities of all currently used corticosteroids such as mometasone furoate (MF; RRA = 2244 ± 142), fluticasone propionate (FP; RRA = 1775 ± 130), the active beclomethasone-17,21-dipropionate (BDP) metabolite beclomethasone-17-monopropionate (17-BMP; RRA = 1345 ± 125), ciclesonide's active principle (des-Cicle; RRA = 1212, rat receptor data) and budesonide (RRA = 855). Together with the compound's high retention in human lung tissue FF incorporates attributes that are suitable for topical anti-inflammatory therapy.
The substitution pattern of the steroidal D-ring is important for the affinity to the glucocorticoid receptor as well as for receptor selectivity [17]. For example, the D-ring substitution confers on MF highly potent glucocorticoid receptor binding affinity [4,18]. We deduced that D-ring modifications of MF were so favourable for high affinity binding to the glucocorticoid receptor that metabolic hydroxylation at the 6β position or loss of chlorine at the 9 position did not result in complete loss of ligand-binding properties.
One characteristic of the MF D-ring substitution pattern, the furoate moiety, is also present in FF. Consistent with the notion that the esterification of the 17α-OH by furoylation augments affinity we a found remarkably high RRA for FF that exceeds the RRA of e.g. FP by more than 60 %. This result is supported by recent X-ray crystal structure data of FF co-crystallized with the glucocorticoid receptor [19]. These data show the 17α-furoate ester fully Control experiment for non-specific adsorption of glucocor-ticoids to incubation vials Figure 4 Control experiment for non-specific adsorption of glucocorticoids to incubation vials. The respective compounds were incubated in glass vials over 480 min at 37°C. The concentration in the supernatant was monitored. The decrease in concentrations of mometasone furoate (MF) indicated the degradation process of the compound. No adsorption was seen for fluticasone propionate (FP) and fluticasone furoate (FF). The columns represent the mean and mean deviation of the mean from triplicate experiments. We determined the affinity of FF to the human lung glucocorticoid receptor by separate analysis of the receptor association and dissociation kinetics. This method is more precise compared to competition assays, especially for high affinity glucocorticoids [1]. For FF we observed a very fast and extensive association with the receptor, with an association rate constant significantly higher than for any other glucocorticoid. In contrast, the dissociation rate constant was almost identical to that of FP. Thus, the difference between FF and FP is mainly based on the more rapid and preferential binding of FF to the receptor. This kinetic behaviour of FF confirms our previous insights into receptor binding characteristics of high-affinity glucocorticoids [1,3]. FP was the first glucocorticoid with receptor binding clearly distinct from other compounds.
In comparison with other glucocorticoids it displayed both a more rapid association and prolonged dissociation from the receptor [1]. The receptor binding kinetics of MF disclosed a high association rate constant while its dissociation rate was almost comparable to FP [3]. Thus, further increase in receptor affinity for FF was related to an increase in the association rate constant which is now also established for this compound.
Interestingly, those glucocorticoids with the highest RRAs do not comply with the previously described linear relationship between lipophilicity and receptor affinity [5]. FF, MF and FP reveal clear differences in their RRA, but their lipophilicity expressed as their relative retention times at a reversed-phase HPLC column is less different than their receptor affinities. However, the correlation between lipophilicity of the active compound and its RRA is still highly significant, though not linear, if the high affinity glucocorticoids FF, FP and MF are included into the analysis. There are glucocorticoids with higher lipophilicity such as BDP and Cicle, but these compounds are pro-drugs with virtually no affinity to the receptor. Both drugs gain activity by ester cleavage in C21 position. Thereby, however, they lose their high lipophilicity.
Since FF is not a pro-drug, its receptor binding affinity and thus activity is associated with the entire molecule. The compound is expected to be stable in the therapeutic target tissue. This is not necessarily seen for all glucocorticoids. We and other research groups recently observed that MF is not stable in lung tissue or plasma and undergoes chemical degradation [3,20,21]. We now elucidated the stability of FF in human lung tissue and found no degradation or metabolism within 24 h at 37°C. The esterase inhibitor dichlorvos was included in one of the incuba-tion mixtures in case of an enzyme-catalyzed hydrolysis of the 17α furoate moiety or of the 17β S-fluoromethyl-carbothioate group. Neither did we determine the resulting metabolites or any other new peaks in the HPLC chromatograms nor did we observe lower FF concentrations in the tissue supernatant in the absence of dichlorvos. In the contrary, we found lower FF concentrations in the presence of the esterase inhibitor, though we do not have a clear explanation for this phenomenon. We conclude that there is no indication of instability or chemical modification of FF in the presence of enzymatically active human lung tissue.
Besides a high receptor binding affinity, a prolonged retention of the glucocorticoid in the lung tissue is a desired property. We compared the tissue binding behaviour of FF with FP and MF. After one hour equilibration of glucocorticoid-saturated lung tissue pieces with human plasma at 37°C, we found highest concentrations of FF and FP compared to MF remaining in the tissue. Obviously, these compounds have the most favourable tissue affinity and it should be expected that the distribution of these glucocorticoids from lung tissue into systemic circulation is slow in vivo. Clinical data confirm this for FP [22].
To conclude, we have characterized the novel glucocorticoid fluticasone furoate. Its relative receptor binding affinity exceeds the RRAs of all other currently clinically used glucocorticoids. Based on the tissue binding experiments a high retention of fluticasone furoate in human lung tissue is expected. These advantageous binding attributes may contribute to a highly efficacious profile for FF as a topical treatment for inflammatory disorders of the respiratory tract.