![]() |
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Astra Zeneca Research & Development Lund, Lund, Sweden
(Received February 21, 2007; Accepted July 9, 2007)
| Abstract |
|---|
|
|
|---|
10-7 M: BUD, CIC, CIC-AM, FP, MF) was added for 20 min. After incubation, one half of the trachea was used for analysis of GC uptake and the other to analyze GC release during 3 h in drug-free medium. GC species in trachea halves were analyzed by radiochromatography. At 20 min, the uptake of BUD was similar to that of CIC/CIC-AM; however, the BUD-ester pool was 9-fold greater (p < 0.01). BUD overall retention in trachea at 3 h was greater than that of other GCs (p < 0.01), and the BUD-ester pool was 3-fold greater than the CIC-AM-ester pool (p < 0.01). Cyclandelate decreased the initial BUD- and CIC-AM-ester pools (p < 0.01), and reduced the overall retention of BUD at 3 h (p < 0.01) but not of CIC-AM. Thus, BUD becomes esterified in the airways more promptly and to a greater extent than CIC-AM, and BUD esterification prolongs BUD airway retention. In contrast, airway retention of CIC-AM and CIC seems to be determined mainly by their lipophilicity, similar to FP and MF, which are not esterified.
Beclomethasone dipropionate (BDP), fluticasone propionate (FP), and the recently introduced mometasone furoate (MF) are examples of highly lipophilic inhaled GCs. After absorption from the airway lumen, BDP is converted to a less lipophilic, active metabolite, beclomethasone monopropionate (BMP), whereas lipophilicity of FP and MF does not change during the passage from the airway lumen to systemic distribution. In contrast to these GCs, inhaled budesonide (BUD) has a moderate lipophilicity, but in airway tissue it becomes esterified with fatty acids forming a highly lipophilic depot of reversible BUD esters that act as a reservoir of active BUD (Miller-Larsson et al., 1998
).
After inhalation, more than 50% of total BUD retained in the rat and human bronchial, nasal, and lung tissue is esterified at the carbon-21-hydroxyl group, primarily as BUD oleate (Miller-Larsson et al., 1998
; Thorsson et al., 1998
; Jendbro et al., 2001
; Petersen et al., 2001
; Maassen van den Brink et al., 2005
). BUD esters are not active; i.e., they do not bind to glucocorticoid receptor (Wieslander et al., 1998
), but they are hydrolyzed slowly, gradually releasing active BUD (Miller-Larsson et al., 1998
; Wieslander et al., 1998
). The formation of BUD esters prolongs the retention and activity of a BUD pulse in rat fibroblast cell line (Wieslander et al., 1998
, 2000
), and esterification of BUD in the airways likely explains the very well documented once daily efficacy of inhaled BUD (Campbell et al., 1991
; Jones et al., 1994
; Banov et al., 2001
; Selroos et al., 2004
). Theoretically, other GCs with a hydroxyl group at carbon-21 can be esterified in a manner similar to that of BUD. These include an endogenous GC, corticosterone, and a synthetic GC, BMP. However, esterification of corticosterone seems not to play a role in corticosterone tissue kinetics and biological activity, probably because corticosterone esters are hydrolyzed rapidly (Hochberg et al., 1991
). Rapid hydrolysis may also explain why BMP esters were not found in rat airways and lungs (Miller-Larsson et al., 1998
).
|
Therefore, in this study, in a rat trachea model, we have compared airway retention of CIC/CIC-AM and CIC-AM esterification with those of BUD, and determined CIC and CIC-AM lipophilicity in relation to BUD. As a reference, and to validate the system, we have used FP and MF, which do not form esters.
| Materials and Methods |
|---|
|
|
|---|
Determination of GC Water Solubility and Lipophilicity. Water solubility. The water solubility of GCs was determined by mixing 1 mg of the GC powder and 5 ml of Milli-Q water. The slurry was sonicated for 5 min and left on a shaker for 17 h. The sample was then filtered through a 0.45-µm Millipore Millex HV filter (Millipore, Billerica, MA). The concentration in the filtrate was determined by high performance liquid chromatography (HPLC) using a Supelcosil LC-18 column (Supelco, Bellefonte, PA), 150 x 4.6 mm. The flow rate was 1.0 ml/min of a mixture of acetonitrile and water, 72:28 for CIC and 47:53 for all other GCs. The GC concentration was calculated using a reference standard for each GC. The limit of quantification at 254 nm was determined to 0.1 µg/ml.
In the analysis of water solubility of MF, there was a second peak in the chromatogram, approximately 30% of the size of the MF, which may indicate a degradation of MF. However, the actual value measured for the MF solubility was 0.03 µg/ml, which was below the lowest standard value. For the purpose of this article, the low limit of quantitation was set to 0.1 µg/ml, and thus a possible stability problem of MF did not affect the result on water solubility reported.
Lipophilicity. The relative lipophilicity of GCs was estimated by the chromatographic capacity factor log k'0. The capacity factor, k', was determined as k'= (tR - t0)/t0, where tR is the retention time for the analyte and t0 is the retention time for a nonretained analyte, e.g., KNO3, in a given liquid chromatography system. The retention time for each analyte in the HPLC system described above for water solubility was determined for at least four different concentrations of acetonitrile in the range of 30 to 70%. For each analyte, the log k' values were calculated and plotted against organic content giving linear relation; the correlation coefficient value (R2) for each analyte was >0.99. From this, the value of log k'0, i.e., k' at 0% organic phase, was extrapolated. For structurally related compounds it has been shown that log k'0 is a good representation of the log P value, i.e., the octanol/water distribution (Tsantili-Kakoulidou et al., 1993
; Rothemund et al., 1994
; Yamagami et al., 2002
).
Study Design. Study design is shown in Fig. 1. Tracheal segments from Brown Norway rats were dissected (approximately 50 mg) and placed in ice-cold standard oxygenated Krebs buffer. A lengthwise incision through the cartilage of the trachea was made and trachea was tied to a glass rod and immersed in 10 ml of medium consisting of standard Krebs buffer with 0.2% glucose and 10% autologous rat plasma (to approximate the protein concentration of airway epithelial lining fluid; Robinson et al., 1989
). Tracheas were preincubated in 37°C oxygenated medium (95% O2/5% CO2) with the esterification inhibitor cyclandelate or vehicle. After 15 min, a 3H-GC (50 µl) was added, and after another 20 min, tracheas were collected, rinsed briefly in fresh buffer, gently and briefly dried on blotting paper to get rid of buffer droplets, and cut lengthwise in half, and the two halves were weighed. One half was frozen at -70°C, awaiting analysis of 3H-GC tissue uptake. The other half was placed in 10 ml of fresh drug-free medium for 3 h to monitor the total radioactivity release from the trachea into medium. Tissue extracts were prepared from both trachea halves to determine 3H-GC tissue concentrations before (initial uptake) and after 3-h release. The total radioactivity of tissue extracts was measured by liquid scintillation counting, and GC species in tissue extracts were analyzed by radiochromatography.
3H-GC Preparation. For all GCs except FP, 50 µl of 3H-GC stock solution in ethanol was added to 10 ml of incubation medium. This resulted in a final concentration of 1.1 x 10-7 to 2.1 x 10-7 M for BUD, CIC, and CIC-AM, and 2.7 to 2.9 x 10-7 M for MF. For FP, 5 µl of 3H-FP ethanol solution together with 45 µl of 99.5% ethanol was used; this resulted in a final concentration of 1.5 to 2.7 x 10-7 M FP. After addition of cyclandelate or its vehicle, the final concentration of ethanol in incubation medium was 1%.
Cyclandelate Preparation. Cyclandelate was suspended in 99.5% ethanol to a concentration of 55.3 mg/ml (2 x 10-1 M) and 50 µl was added to 10 ml of incubation medium giving 10-3 M as the final concentration. This concentration of cyclandelate as well as the incubation time (15 min of preincubation followed by 20 min of incubation with 3H-GC) were titrated in pilot experiments to investigate whether cyclandelate can inhibit GC ester formation without exerting cytotoxic effects on epithelium in trachea. Histological sections of tracheas were examined using scanning electron microscopy and fluorescence analysis of DNA to investigate epithelial integrity and reveal any damage to epithelial cell nuclei. At the cyclandelate concentration and incubation time applied, normal appearance of epithelium was observed, i.e., normal density of epithelial cells and evenly stained cell nuclei without chromatin fragmentation. The histological analysis was performed by Dr. Jonas Erjefält at the Department of Experimental Medical Science at Lund University (Lund, Sweden).
Release of GCs from Trachea. During the 3-h release period, 200-µl samples were withdrawn from incubation medium (10 ml) at various time points and the volume was replaced with fresh medium. The radioactivity of withdrawn samples was analyzed by liquid scintillation counting (200 µlin10 ml of Ultima Gold scintillation cocktail; Packard, Groningen, Netherlands) in a Tri-Carb Spectrometer 2200CA (Packard). The accumulated release of radioactivity was calculated for each time point and expressed as percentage of total radioactivity initially present in trachea (calculated as a sum of released and remaining radioactivity in trachea at 3 h).
Analysis of GC Species in Tracheal Extracts. Trachea extracts were prepared by microwave-assisted extraction in 2 ml of 99.5% ethanol in a microwave oven (Mars X or MES 1000; CEM Corporation, Matthews, NC) for 30 min at 90°C with a power of 13 to 17%. The efficiency of extraction measured with spiked control samples was nearly complete (>90%). The extracts were filtered and stored (at -20°C) before analysis. Total radioactivity was measured in 50-µl extracts by liquid scintillation counting in 10 ml of Ultima Gold scintillation cocktail (Packard) in a Tri-Carb Spectrometer 2200CA (Packard).
For analysis of GC species, 500 µl of extract was injected into the liquid chromatography system combined with on-line scintillation detection in a Radiomatic Flo-One beta A500 radiochromatography detector (Packard). The reverse-phase column (Supelcosil LC-18-DB, 3.3 cm x 4.6 mm, 3 µm; Supelco Inc.) was used with a three-phase ethanol gradient: phase A containing 5% ethanol and 0.1% acetic acid, phase B containing 95% ethanol and 0.1% acetic acid, and phase C containing 0.1% acetic acid. The gradient was as follows: 0 to 7.6 min, 65% A/35% B; 7.6 to 15.2 min, 15% A/85% B; 15.2 to 20 min 100% B. The flow rate of phases A and B was 0.35 ml/min for the first 3 min, and that of phase C was 0.65 ml/min for the first 3 min and thereafter 1 ml/min of phases A and B. The recovery of the total radioactivity in this system versus liquid scintillation counting was 90 to 106%.
GC species were identified and confirmed in pilot experiments applying liquid chromatography and mass spectroscopy techniques using an HP1100 LC system (Agilent Technologies, Waldbronn, Germany) in line with a QTOF mass spectrometer (Waters, Manchester, UK). The identification was based on comparison of chromatographic retention times of reference compounds and mass spectrometry data.
Data Analysis. The tissue concentration (Q) of GC species in the samples, expressed as pmol/g, was calculated from the following equation: Q = DPM x DF/(C x SA x W), where DPM is the mean number of disintegrations per minute (dpm) measured in the sample, DF is dilution factor (= 4; total extract volume/LC injection volume), C = 2.22 x 106 dpm/µCi, SA is the specific radioactivity of GC applied (µCi/pmol), and W is the weight (grams) of tissue samples. Because the 3H-GC concentrations in incubation medium differed somewhat between GCs (see 3H-GC Preparation), Q values were normalized to the concentration of 1.0 x 10-7 M and expressed as pmol/g. When two or more species of a GC were detected, the sum of their Q values was calculated, and is referred to here as the total tissue concentration of a GC in trachea.
Statistical Analysis. Data are presented as arithmetic means ± S.E.M. Data were analyzed by one-way analysis of variance with significance level down to 1% using Astute software 1.5 (DDU Software, Leed, UK). Differences were considered significant at p < 0.05.
| Results |
|---|
|
|
|---|
|
Tissue Uptake and Retention of GCs. Radiochromatographic analysis of trachea extracts revealed fractions of active GCs, GC esters, and parent GC (for CIC). In experiments with BUD or CIC-AM, unmodified GCs and GC esters were detected (Fig. 2, A-D). In experiments with CIC, three fractions were detected: parent CIC, CIC-AM, and CIC-AM esters (Fig. 2, E and F). Thus for BUD, CIC-AM, and CIC, the total tissue concentration in trachea is represented by a sum of individual GC species. In experiments with FP and MF, only unmodified GCs were detected (not shown).
|
|
Analysis of esterified species in trachea revealed that a much higher percentage of BUD was in the esterified form than for the CIC/CIC-AM (Fig. 4). After the 20-min incubation, BUD esters made up 46% of the total BUD content in trachea, whereas CIC-AM esters made up 6 to 7% of the total CIC content (p < 0.01 versus BUD) after both CIC or CIC-AM incubation. After the subsequent 3-h release period in GC-free medium, BUD esters made up 92% of the total BUD as compared with 52% or 67% of CIC-AM esters after incubation with CIC or CIC-AM, respectively (p < 0.01 for both versus BUD). In the absolute amounts, the BUD-ester pool was 9-fold greater than the CIC-AM-ester pool after 20 min of incubation (p < 0.01), and 3-fold greater after the 3-h release period (p < 0.01). These results show that BUD was esterified in trachea more promptly and to a greater extent than CIC-AM.
|
Role of GC Esterification for Tissue Uptake and Retention of GCs. The relationship between esterification and retention of BUD, CIC, and CIC-AM was evaluated by radiochromatographic analysis of tracheas incubated with the esterification inhibitor cyclandelate for 15 min before and during the 20-min incubation with GCs (examples of typical radiochromatograms are shown in Fig. 5). In the experiments involving CIC incubation, cyclandelate significantly inhibited hydrolysis of CIC to CIC-AM (not shown), and this effect prevented any conclusions regarding the consequences of ester inhibition on CIC kinetics in trachea.
|
Incubation with cyclandelate significantly decreased the initial BUD-ester pool in trachea to 31% of the vehicle-treated value (p < 0.01) and increased the pool of BUD 1.8-fold (p < 0.01), but it did not affect the total BUD uptake at 20 min (1.1-fold increase, p = 0.2) (Fig. 6). Similarly, incubation of CIC-AM with cyclandelate significantly decreased CIC-AM esters to 6% of the vehicle-treated value (p < 0.01) and increased the pool of CIC-AM 2-fold (p < 0.01), but it also significantly increased 1.9-fold (p < 0.01) the total uptake of CIC-AM at 20 min (Fig. 6). The uptake of MF was not affected by cyclandelate, whereas the uptake of FP was reduced by 19%, but this difference was not statistically significant (p = 0.1; data not shown).
|
|
| Discussion |
|---|
|
|
|---|
The uptake of BUD into rat tracheal tissue was roughly comparable to the uptake of either CIC or CIC-AM. The absolute total tissue concentration of these GCs reached in trachea after 20 min of incubation was at the level of
10-7 M (10-7 mol/kg). Similar tissue concentrations in rat trachea, and a 10-fold lower level in lung tissue, were earlier obtained in vivo after inhalation of BUD and FP (Miller-Larsson et al., 1998
) and recently after inhalation of CIC (Watz et al., 2006
). The
10-8 M concentration in lung tissue corresponds to the levels obtained in humans 1 to 4 h after inhalation of high clinical doses (1-2 mg) of GCs (Van den Bosch et al., 1993
; Esmailpour et al., 1997
; Thorsson et al., 1998
; Maassen van den Brink et al., 2005
; Watz et al., 2006
).
The experiments on in vitro uptake of GCs were conducted with 10% autologous rat serum in the incubation medium. This concentration was chosen to approximate the protein concentration present in airway lining fluid. The protein amount in airway lining fluid is difficult to measure, and it is known to increase in inflammation and certain airway diseases, but we decided to use a conservatively low value for these experiments (Robinson et al., 1989
). This is an important technical consideration, because the tissue uptake of lipophilic molecules with high protein binding may be impaired by binding to proteins in culture medium, or it may be increased by the use of low protein concentrations. CIC and CIC-AM bind to plasma proteins at
99% (Rohatagi et al., 2005
) compared with 88% binding of BUD (Ryrfeldt et al., 1982
). Accordingly, Jerre et al. (2006
) have shown that uptake of CIC into cultured human airway epithelial cells depends strongly on protein concentration in culture medium. At physiologic protein concentration in airway lining fluid, represented by 10% serum in culture medium, CIC was poorly taken up into airway epithelial cells, and its uptake over the 2-h incubation was 3-fold lower than that of BUD. In contrast, Nave et al. (2005b
) have reported higher uptake of CIC than of BUD into human alveolar type II epithelial cell line. This apparent discrepancy between the results of Jerre et al. (2006
) and those of Nave et al. (2005b
) are likely due to the very low protein concentration (0.1% bovine serum albumin) in the incubation medium in the Nave et al. (2005b
) study. In an in vivo study in which BUD or CIC was instilled intratracheally into rats, the uptake of BUD into trachea and lung tissue was higher than that of CIC (Gullstrand et al., 2005
). It is conceivable that besides the initial uptake, the tissue retention of GCs with high protein binding can also be increased by using too low of a protein concentration in the incubation medium in in vitro experiments.
Despite roughly equal uptake of BUD and CIC/CIC-AM into rat trachea in the present study, there were two major differences in tracheal tissue kinetics between BUD and CIC/CIC-AM. The first one was that a much larger proportion of BUD than of CIC-AM was esterified within the 20 min and remained esterified after the 3-h release period. Similarly, in an above-mentioned in vivo study, in which rats were instilled intratracheally with BUD or CIC, BUD retained in trachea tissue was esterified to a significantly greater extent than CIC-AM (and in contrast to BUD esters, CIC-AM esters were not detectable in lung tissue; Gullstrand et al., 2005
). In the present study, to investigate CIC-AM esterification per se, independently of the efficiency of the transformation of CIC into CIC-AM, we have incubated trachea with CIC or CIC-AM. In both cases the extent of CIC-AM esterification was very similar, showing that CIC was efficiently converted to CIC-AM in this trachea in vitro system.
The second major difference was detected by use of an inhibitor of esterification, cyclandelate. Esterification of BUD at carbon-21 was earlier shown to be dependent on CoA and ATP and probably catalyzed by the ATP- and CoA-dependent microsomal acyl-CoA:cholesterol acyl transferase (ACAT) (Tunek et al., 1997
). It was earlier shown that cyclandelate, an inhibitor of ACAT, decreased the amount of BUD esters and shortened the retention and activity of BUD pulse in a rat fibroblast cell line (Wieslander et al., 2000
). In the present study we have confirmed and extended these findings demonstrating that reduction of the initial BUD-ester pool by cyclandelate significantly decreased retention of BUD in rat tracheal tissue at 3 h. Cyclandelate treatment also nearly completely blocked the initial CIC-AM-ester pool; however, in contrast to BUD, it did not decrease the retention of CIC-AM in rat trachea at 3 h.
It was not expected that cyclandelate should affect tissue kinetics of FP or MF in trachea, and it had no effect on either their uptake or retention, which validates the system. The retention of FP and MF in trachea was similar to that of CIC and CIC-AM (Fig. 7). FP, MF, CIC, and CIC-AM are all highly lipophilic GCs (Table 1), and high lipophilicity was shown to increase the airway retention of inhaled GCs (Brattsand, 1997
). The very high lipophilicity of BUD esters (Table 1), rapidly formed in the airways, retards absorption of BUD into blood circulation and explains the prolonged airway retention of inhaled BUD. The major CIC-AM ester, oleate, was reported to be 5 times more lipophilic than BUD oleate (Nave et al., 2004
). However, a slow rate of net formation of CIC-AM esters probably explains the finding that the blockage of the initial small pool of CIC-AM esters did not prolong CIC-AM retention in trachea in the present study. It seems that after inhalation of CIC, CIC-AM retention is determined by parent CIC and CIC-AM lipophilicity, similar to retention of inhaled FP and MF.
Surprisingly, the presence of cyclandelate significantly increased nearly 2-fold the initial total uptake of CIC-AM but not that of BUD or other GCs tested. This outcome occurred despite the fact that CIC-AM esters initially made up only 6 to 7% of the total CIC-AM in tracheal tissue, whereas BUD esters made up nearly 50% of the total BUD. We do not have an explanation for the increased total uptake of CIC-AM by cyclandelate. Furthermore, we have expected that the blockage of BUD esterification would decrease BUD uptake since this was shown in the rat fibroblast cell line (Wieslander et al., 2000
). The finding that the BUD uptake into tracheal tissue was not dependent on BUD esterification and was roughly equal to that of the 2 to 4 times greater lipophilic MF, CIC-AM, and CIC, but lower than the 3-fold greater lipophilic FP, suggests that, besides lipophilicity, there are other factors that influence GC uptake into airway tissue. For example, evidence is growing that cellular concentrations of GCs are regulated not only by GC passive diffusion, dependent on small size and high lipophilicity of GC molecules, but also via efflux transporters, such as P-glycoprotein. The efficiency of P-glycoprotein-mediated efflux of various GCs seems to depend on subtle differences in GC structure (Yates et al., 2003
). Recent data in colon carcinoma cells suggest that BDP, MF, and CIC-AM, but not FP, are substrates to P-glycoprotein (Cooray et al., 2006
), whereas data on BUD are controversial (Dilger et al., 2004
; Cooray et al., 2006
). However, it is not yet known whether GC transporters operate in airway and lung cells.
The observed difference between BUD and CIC-AM rates of net ester formation are likely to result from equilibrium differences between ester synthesis and lipase/esterase-catalyzed hydrolysis. It cannot be excluded that these differences are due to different enzymes regulating esterification and/or hydrolysis of these two GCs, although ACAT seems to be involved in the esterification of both BUD and CIC-AM since cyclandelate (ACAT inhibitor) reduced the initial formation of both BUD and CIC-AM esters. Furthermore, the hydrolysis of GC esters is a relatively stereoselective process (Hochberg et al., 1991
), and hydrolysis of BUD esters may be decelerated by the steric hindrance of the acetal group at 16
- and 17
-carbons. Contrariwise, a rapid hydrolysis of enzymatically accessible corticosterone esters (Hochberg et al., 1991
) and cortisol esters (Dr. A. Tunek, AstraZeneca, personal communication) may explain why few esters of corticosterone and cortisol are detected, and why biological activity of these endogenous GCs is not dependent on ester formation.
Rapid formation of BUD esters in the airways and their slow hydrolysis contribute not only to prolonged airway retention and activity (resulting in once daily efficacy; Campbell et al., 1991
; Jones et al., 1994
; Banov et al., 2001
; Selroos et al., 2004
), but probably also to high airway selectivity (Edsbäcker and Jendbro, 1998
; Miller-Larsson et al., 2000
; Jendbro and Johansson, 2002
). This is because 1) esterification of inhaled BUD in the airway tissue retards BUD absorption into blood circulation, and 2) BUD esterification occurs to a much lower extent in skeletal muscles (10-15%) than in the airways (more than 50%) and does not occur in plasma (Miller-Larsson et al., 1999
; Jendbro et al., 2001
). It is important to note that the results of pharmacokinetic and pharmacodynamic modeling suggest that the faster the formation of esters in the airways and the slower their hydrolysis, the greater the airway selectivity of inhaled GCs (Jendbro and Johansson, 2002
). Indeed, we have earlier shown that after 4 days of repeated intratracheal instillation of BUD into rats, the total tissue concentration of BUD over 24 h was 20- to 40-fold greater in trachea than in skeletal muscles and plasma, whereas respective ratios for CIC were 2-fold lower than those for BUD (Gullstrand et al., 2005
). Thus, although inhaled CIC is reported to have a high airway selectivity, according to the results of this study, this is unlikely to be due to airway esterification of CIC-AM.
In summary, this study gives the evidence that prolonged retention of BUD in the airways is due to the rapid and extensive formation of a highly lipophilic and reversible BUD-ester depot. In contrast, CIC-AM was esterified only slowly and to a lower degree than BUD, which may explain the lower retention of CIC-AM in the tracheal tissue and the finding that CIC-AM retention was not dependent on CIC-AM esterification in this rat trachea model. These results suggest that unlike inhaled BUD, retention of CIC-AM after inhalation of CIC is mainly determined by parent CIC and CIC-AM lipophilicity, similar to inhaled FP and MF, which do not form esters.
| Acknowledgments |
|---|
| Footnotes |
|---|
ABBREVIATIONS: GC, glucocorticosteroid; BUD, budesonide; CIC, ciclesonide; CIC-AM, ciclesonide-active metabolite; FP, fluticasone propionate; MF, mometasone furoate; BDP, beclomethasone dipropionate; BMP, beclomethasone monopropionate; HPLC, high-performance liquid chromatography; ACAT, acyl-CoA:cholesterol acyl transferase; AUC, area under the retaining curve.
Address correspondence to: Anna Miller-Larsson, AstraZeneca R&D Lund, 221 87 Lund, Sweden. E-mail: anna.miller-larsson{at}astrazeneca.com
| References |
|---|
|
|
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||