Abstract
Pimecrolimus and tacrolimus are calcineurin inhibitors used for the topical treatment of atopic dermatitis. Although structurally similar, they display specific differences including higher lipophilicity and lower skin permeation of pimecrolimus. The aim of the present study was to understand the reason for the differences in skin permeation; in addition, plasma protein binding of the two drugs was analyzed side by side as a basis for comparison of systemic exposure to free drug. Permeation of pimecrolimus and tacrolimus through a silicon membrane was found to be similar; therefore, we assumed that differences in skin permeation could be caused by differences in affinity to skin components. To test this hypothesis, we investigated binding of pimecrolimus and tacrolimus to a preparation of soluble human skin proteins. One binding protein of approximately 15 kDa, probably corresponding to macrophilin12, displayed a similar binding capacity for pimecrolimus and tacrolimus. However, less specific, nonsaturating binding to other proteins was approximately 3-fold higher for pimecrolimus. Because of the high local drug concentration after topical administration, the unspecific, high-capacity binding is probably dominating the permeation through skin. In plasma both drugs bound predominantly to lipoproteins, which may affect disposition differently from albumin binding. The unbound fraction of pimecrolimus in human plasma was approximately 9-fold lower compared with that of tacrolimus (0.4 ± 0.1 versus 3.7 ± 0.8%). In conclusion, these results provide an explanation for the observed lower systemic exposure to pimecrolimus than to tacrolimus after topical application and suggest that differences in systemic exposure to free drug might be even more pronounced.
Pimecrolimus (Elidel) and tacrolimus (Protopic) are calcineurin inhibitors used for the topical treatment of atopic dermatitis (Stuetz et al., 2006). The compounds bind to cytoplasmic proteins of the immunophilin family, in particular to macrophilin12 (Kissinger et al., 1995; Grassberger et al., 1999), which is highly and ubiquitously expressed (Galat, 2003); inhibition of calcineurin occurs in a ternary calcineurin-immunophilin-drug complex. Despite a high degree of structural similarity (Fig. 1), pimecrolimus and tacrolimus display characteristic differences in terms of pharmacological profile (Stuetz et al., 2001; Meingassner et al., 2003; Grassberger et al., 2004; Bavandi et al., 2006; Kalthoff et al., 2007), and physicochemical and pharmacokinetic properties. Regarding physicochemical properties, the higher lipophilicity of pimecrolimus is noteworthy: pimecrolimus features an 8-fold higher octanol-water distribution coefficient than tacrolimus (Billich et al., 2004). Another distinguishing feature between the two agents is their rate of skin permeation: the permeation rate from 1% solutions is approximately 10-fold lower for pimecrolimus than for tacrolimus (Billich et al., 2004); also from the marketed 1% cream the permeation rates of pimecrolimus are approximately 6- and 4.3-fold lower than those from 0.1 and 0.03% tacrolimus ointments, respectively, despite the much higher pimecrolimus concentration in the formulation (Meingassner et al., 2005). The reason for the pronounced difference in skin permeation of the two drugs has so far not been investigated in detail.
Low skin permeation is a favorable property for a topical drug, because it contributes to low systemic exposure levels and thus to a lower risk of systemic side effects. Indeed, topical pimecrolimus is associated with lower systemic drug exposure than tacrolimus (Draelos et al., 2005). In a comparison of systemic exposure levels exposure to both total and unbound drug should be considered. The latter is relevant, because free rather than total drug concentrations may drive wanted or unwanted pharmacological effects. For tacrolimus very different unbound fractions of 1.2 and 27% in plasma have been reported based on different separation techniques (Piekoszewski and Jusko, 1993; Nagase et al., 1994; Zahir et al., 2001); data on pimecrolimus have not been published so far.
Here we report on studies performed to better understand the cause of the difference in skin permeation between pimecrolimus and tacrolimus. In addition, we present data on comparative plasma protein binding to allow for a comparison of systemic exposure to unbound drug, and we identify the major binding partners in plasma for both drugs.
Materials and Methods
Permeation Assay. Permeation was studied using static Franz-type diffusion cells in which silicone elastomer membranes (no. 7-4107, 75 μm thick; Dow-Corning, Coventry, UK) were mounted. The exposed membrane area was 2.54 cm2, and the volume of the receptor chamber was 5.8 ml. Phosphate-buffered saline-ethanol (3:1) was used as receptor phase. All experiments were performed at 32°C in triplicate for 48 h. Pimecrolimus and tacrolimus were applied to the membranes either in solution (propylene glycol-oleyl alcohol, 9:1) at a concentration of 1% (w/v) in a volume of 300 μl or in their marketed formulations [1% pimecrolimus cream (Elidel); Novartis Pharma AG, Basel Switzerland, and 0.1 or 0.03% tacrolimus ointment (Protopic); Astellas Pharma Inc., Tokyo, Japan; applied amount: 300 mg].
Samples of 100 μl were withdrawn from the receptor phase at four to eight time points during the 48-h experiment and replaced by fresh receptor fluid. After addition of an internal standard and dilution with 0.1% formic acid-acetonitrile (50:50), these samples were analyzed directly by high-performance liquid chromatography-MS/MS (see below).
Sample Analysis. Liquid chromatography-MS/MS analysis was carried out with a Hewlett-Packard 1090 M high-performance liquid chromatograph coupled to a Finnigan LCQ mass spectrometer. A Phenomenex Luna C18 column (3 μm, 100 × 2 mm) equipped with a precolumn, was eluted isocratically with a flow rate of 200 μl/min at 60°C. The eluant was 0.1% formic acid-acetonitrile (20:80). The sample injection volume was 10 μl. The effluent was delivered unsplit to the electrospray ionization ion source used in positive mode. Under the chromatographic conditions used, pimecrolimus and tacrolimus yielded a strong sodium adduct. For MS/MS, parent ions were selected at 832.0 m/z with a band width of 4 m/z for pimecolimus and at 826.2 m/z with a band width of 2 m/z for tacrolimus. Collision-induced dissociation was carried out with a collision energy of 28%, yielding a fragment ion at 604.2 m/z for pimecrolimus. A collision energy of 43% was applied to yield a fragment ion at 616.2 m/z for tacrolimus. The quantification of the parent ions was based on the area ratio of the fragment ions to the fragment ion of an internal standard. For calibration, receptor medium was spiked with variable amounts of the analytes, resulting in concentrations of 1 to 1000 ng/ml. Calibration curves were set up both with fresh medium and with medium taken at 48 h from permeation assays with formulations only (i.e., without active compound), to control for possible interference by excipients, which, however, was not observed. The limits of quantification for pimecrolimus and tacrolimus were 10 ng/ml in receptor fluid. Calculation of flux was done as described (Schmook et al., 2001).
Radiolabels and Stock Solutions. Tritium-labeled pimecrolimus (543.3 MBq/mg) and tacrolimus (1035 MBq/mg) (positions of radiolabels are given in Fig. 1) were prepared and supplied by the Isotope Laboratory of Novartis Pharma AG. Ethanolic stock solutions were prepared by serial dilution including unlabeled compound, resulting in final specific activities of 10.9 to 543 and 20.7 to 207 MBq/mg for pimecrolimus and tacrolimus, respectively, and concentrations of 2 to 100 μg/ml (1000 times final assay concentrations).
Biological Matrices. Human plasma (lithium heparin as anticoagulant, plasma pools from three healthy male donors) was delivered frozen from EFS-ALSACE (Strasbourg, France). Fibrin in plasma was removed by centrifugation for 10 min at 10,000g at room temperature and cleared plasma was frozen in aliquots at –20°C and defrosted before use. Human serum albumin (catalog no. A 1887), α1-acid glycoprotein (catalog no. G 9885), and γ-globulins (catalog no. G 4386) were purchased from Sigma-Aldrich (St. Louis, MO), and solutions were prepared in phosphate-buffered saline. Human high-density lipoprotein (HDL) (catalog no. 437641), low-density lipoprotein (LDL) (catalog no. 437644), and very-low-density lipoprotein (VLDL) (catalog no. 437647) were purchased from Calbiochem (San Diego, CA); these lipoproteins were delivered as solutions in 150 mM NaCl and 0.01% EDTA, pH 7.4. The concentrations of the lipoproteins as given by the manufacturer were adjusted to the concentrations we used with phosphate-buffered saline.
The water-soluble protein fraction from human cadaver skin (skin extract) was prepared as follows. Skin specimens were obtained from NDRI, Inc. (Philadelphia, PA). Samples of similar weight from three different male donors were pooled and cut into small pieces (approximately 3 × 3 mm) using sharp scissors. The material was suspended in 10 volumes of ice-cold phosphate-buffered saline and homogenized in a Potter S homogenizer (B. Braun Biotech International GmbH, Melsungen, Germany). The homogenate was first centrifuged at 3500g for 5 min at 4°C. The supernatant was then subjected to a second centrifugation (15,000g for 1 h at 4°C). The protein concentration in the supernatant (= skin extract) was determined using an assay from Bio-Rad (catalog no. 500-0006; Bio-Rad, Hercules, CA) and bovine serum albumin as standard. The extract was frozen, stored at –80°C, and thawed before use. After thawing, the solution was cleared by centrifugation (2000g for 4 min at room temperature) before further use.
Protein Binding. Binding of [3H]pimecrolimus and [3H]tacrolimus to proteins was analyzed by equilibrium gel filtration (Hummel and Dreyer, 1962; Berger and Girault, 2003). Two systems were used. 1) For analysis of total binding two 5-ml HiTrap Desalting columns (GE Healthcare, Little Chalfont, Buckinghamshire, UK) were used in series; here proteins (>5 kDa) were not separated. 2) For separation of binding proteins, a Superose 6HR 10/300 column (GE Healthcare) was used with a separation range of 5 to 5000 kDa. The gel filtration column was equilibrated with phosphate-buffered saline containing the compound under investigation at nominal concentrations of 2, 20, or, 100 ng/ml. The temperature was 37°C for the HiTrap Desalting columns and 22 to 24°C for the Superose 6HR 10/300 column; the flow rate was 0.2 ml/min in both cases. Protein-containing solutions were injected at volumes chosen to ensure that a binding equilibrium was achieved on the column (e.g., 100 μl of plasma, 0.2 mg of skin protein, and 0.2–4 mg of different plasma proteins). The eluate was analyzed for total protein by UV absorption (280 nm) and for total 3H radioactivity in fractions collected by liquid scintillation counting in a Packard TriCarb liquid scintillation counter. Pimecrolimus and tacrolimus concentrations in binding experiments were determined by liquid scintillation counting using the respective specific activities.
Data Analysis. In contrast with the Superose 6HR 10/300 column the HiTrap Desalting column does not separate proteins (>5 kDa), but the total protein runs faster than free [3H]pimecrolimus or [3H]tacrolimus, allowing achievement of binding equilibration on the column. The amount of compound bound (AB) per milliliter of plasma was calculated as (APP – Cfree · VPP)/ Vplasma, with APP being the total amount of compound in the protein peak fractions, Cfree being the actual free compound concentration on the column (average concentration in fractions before the protein peak), VPP being the total volume of the protein peak fractions, and Vplasma being the injected plasma volume in milliliters. The fraction unbound in plasma (fu) was calculated as: 1 – (AB/(AB + Cfree). For purified plasma proteins, unbound fractions in plasma were calculated on the basis of 1) the amount of compound bound per milligram of the protein used, 2) the free compound concentration measured, and 3) a physiologically relevant concentration of the particular plasma protein (Table 3).
Molecular Mass Estimation of Binding Proteins. The Superose 6HR 10/300 column was calibrated using standard proteins [ribonuclease (15.6 kDa), chymotrypsinogen (20.4 kDa), ovalbumin (48.1 kDa), albumin (63.5 kDa), aldolase (171 kDa), catalase (232 kDa), ferritin (391 kDa), and thyroglobulin (725 kDa); GE Healthcare]; elution volumes of standard proteins (mean of two injections) were plotted over the molecular mass (log scale), and a calibration curve was established by linear regression with an R2 value of 0.958. The relative contributions of different separable proteins to the overall binding was roughly estimated by comparison of peak areas of radio signals.
Results
Permeation through a Silicone Membrane. We compared permeation of pimecrolimus and tacrolimus through a silicon elastomer membrane; such membranes are commonly used as artificial barriers for drug release studies from topical formulations. For both agents the permeation rate through the silicon membrane was nearly identical (Table 1) when they were applied as 1% solutions in propylene glycol-oleyl alcohol (9:1). Permeation was also tested for the marketed formulations of the two drugs. When Elidel (1% pimecrolimus cream) was applied to the artificial membrane, permeation of pimecrolimus was 5.7-fold lower than that from the 1% pimecrolimus solution, indicating somewhat lower release from the cream. With Protopic (0.1 and 0.03% tacrolimus ointment) lower permeation compared with Elidel was observed with the membrane. The differences in flux roughly reflected the differences in drug concentrations (Table 1). This finding suggests that the difference in drug concentration causes the difference in flux and confirms that permeation characteristics through the artificial membrane are very similar for pimecrolimus and tacrolimus. In contrast, as reported previously (Billich et al., 2004, Meingassner et al., 2005), the skin permeation of the two agents differs markedly, whether applied as 1% solutions or as commercial cream and ointments, respectively, and is 4 to 10 times lower for pimecrolimus than for tacrolimus (Table 1).
Binding to Soluble Skin Proteins. To shed light on the observed differences in skin permeation of pimecrolimus and tacrolimus, binding to a preparation of soluble skin proteins was analyzed. For both drugs the amount bound per milligram of protein increased apparently linearly with the free drug concentration in the concentration range covered; however, the experimental scatter was relatively large, particularly at high concentrations. This is probably due to the low protein concentration in the skin extracts, resulting in a limited binding capacity and thus a relatively small bound signal above the free concentration at high free concentrations. Some fluctuation in the free concentration on the column can contribute to the scatter. For pimecrolimus the slope of a linear fit was approximately 3-fold higher compared with that for tacrolimus (Fig. 2A), suggesting a generally higher affinity of pimecrolimus in binding to soluble skin proteins.
To identify any major binding proteins in the skin preparation, the proteins were separated on a Superose 6HR 10/300 column equilibrated with pimecrolimus or tacrolimus. For both drugs only one major specific peak was identified. It eluted at a molecular mass of approximately 15 kDa (Fig. 2B). The amounts of compound in this peak were similar for both compounds: 13.5 ± 2.4 and 12.7 ± 3.5 ng/mg of protein (total protein injected) for pimecrolimus and tacrolimus, respectively (n = 3). These values correspond roughly to the y-axis intercept, when total binding is plotted over the free drug levels (Fig. 2A), and could result from saturation of the 15-kDa binding protein at all drug concentrations tested, owing to a dissociation constant below the lowest tested concentration.
Plasma Protein Binding. Using equilibrium gel filtration on HiTrap Desalting columns, we observed that at similar free concentrations of pimecrolimus and tacrolimus approximately 9-fold less of the latter was bound to plasma proteins (Fig. 3A). On the basis of binding data at free concentrations of 0.48 to 93.3 ng/ml (pimecrolimus) and 3.3 to 80.2 ng/ml (tacrolimus), unbound fractions of pimecrolimus and tacrolimus in human plasma were estimated to be 0.4 ± 0.1 and 3.7 ± 0.8%, respectively (Table 2). No major concentration dependence of plasma protein binding was found in the concentration range tested. To identify the major binding proteins for pimecrolimus and tacrolimus, the binding to purified human plasma proteins was analyzed (Table 3). At physiologically relevant concentrations of the different proteins, binding of pimecrolimus was estimated to be highest to lipoproteins, particularly to HDL. For tacrolimus, binding was highest to HDL, followed by binding to VLDL and α1-acid glycoprotein. Binding to α1-acid glycoprotein and γ-globulins was similar for pimecrolimus and tacrolimus. In contrast, binding to human serum albumin and lipoproteins was 5- to 9-fold higher for pimecrolimus (Fig. 4), which probably causes the overall higher binding of pimecrolimus in plasma.
To substantiate these results, binding to human plasma was analyzed on a Superose 6HR 10/300 gel filtration column. When plasma proteins were separated on a column equilibrated with pimecrolimus, the bulk of pimecrolimus eluted at the expected elution volumes of HDL (≥170 kDa, major peak) and LDL (approximately 3500 kDa) (Fig. 3B). The very broad peaks were in line with the variable molecular masses of lipoproteins, which are composed of apoproteins and lipids in somewhat varying ratios. The major protein peak at approximately 70 kDa (Fig. 3B, UV trace), which corresponds mainly to albumin, was not linked to a pimecrolimus peak, confirming the fact that albumin was less relevant for the overall plasma protein binding of pimecrolimus. In contrast to what could be expected from the experiment using purified plasma proteins (Table 3), no peak was found at the expected molecular mass of VLDL. This is probably due to removal of floating VLDL during plasma preparation for injection (removal of fibrin, see Materials and Methods).
Upon separation of human plasma proteins on a Superose 6HR 10/300 column equilibrated with tacrolimus, the highest and relatively slim tacrolimus peak concurred with the second half of the main protein peak (Fig. 3B). This peak probably corresponds to α1-acid glycoprotein -bound tacrolimus. α1-Acid glycoprotein is expected to elute slightly later than serum albumin, because of its lower molecular mass (approximately 43 kDa), but cannot be separated from serum albumin, because the molecular mass difference is insufficient. A broader peak was apparent at the expected elution volume of HDL and a small shoulder at the expected elution volume of LDL. This finding is in line with the experiments performed with isolated plasma proteins (Table 2); again a VLDL peak was missing, which may be attributed to separation of VLDL during plasma preparation.
Discussion
We recently reported that the rate of permeation through human skin is lower for pimecrolimus than for tacrolimus, when we compared 1% solutions and the marketed formulations of the two drugs (Table 1) (Billich et al., 2004; Meingassner et al., 2005). This finding is in line with lower systemic exposure observed in patients treated with pimecrolimus cream compared with those treated with tacrolimus ointment for atopic dermatitis (Draelos et al., 2005) and Netherton syndrome (Oji et al., 2005; Allen et al., 2001). In the present studies we found that permeation through an artificial membrane was similar for the two drugs (Table 1). This result indicates that the lower permeation of pimecrolimus through skin into the receptor fluid is not caused by a significantly slower release of the compound from the formulations compared with that of tacrolimus nor by limited solubility of pimecrolimus in the receptor fluid. Rather it appears that the distribution equilibrium between skin and receptor fluid is more on the side of the skin in the case of pimecrolimus, leading to a markedly reduced permeation through skin compared with tacrolimus. These observations point to stronger binding of pimecrolimus to components of the skin compared with that for tacrolimus.
To test this hypothesis, we investigated binding of pimecrolimus and tacrolimus to a preparation of soluble skin proteins. It is important to note that many major components of the skin, such as, collagen fibers or horny skin, are not covered by these experiments. In the skin protein preparation we identified one specific binding protein with a molecular mass of approximately 15 kDa that displayed a similar binding capacity for pimecrolimus and tacrolimus. Probably this binding peak corresponds to a protein of the tacrolimus binding protein family (FK506-binding proteins, also called macrophilins), which comprises members with molecular masses close to 15 kDa, of which the FK506 binding protein-12 or macrophilin12 is the best characterized and most prevalent (Galat, 2003). At very low free concentrations this low-capacity and high-affinity binding might result in similar amounts bound for pimecrolimus and tacrolimus. However, at higher free drug concentrations the protein preparation had approximately 3-fold more pimecrolimus bound than tacrolimus, suggesting a higher affinity of pimecrolimus in this less specific binding. This result may explain the slower skin permeation of pimecrolimus: Because of the high local drug concentration after topical administration, the unspecific high-capacity binding is probably dominating binding in the upper skin layers. In deeper layers at low total concentrations the specific binding would be similar, resulting in a similar pull for both drugs from the deeper layers. Because of the higher binding capacity for pimecrolimus close to the site of application and a similar pull from deeper layers, the permeation of pimecrolimus would be slower and the concentration gradient in terms of total concentration steeper. Both pimecrolimus and tacrolimus bind to macrophilin12 with high affinity. Reported IC50 values based on different experimental setups are slightly lower for tacrolimus (0.88 ± 0.2 nM) (Weiwad et al., 2006) compared with pimecrolimus (1.8 ± 0.3 nM) (Grassberger et al., 1999). This result suggests that at low total concentrations the free tacrolimus concentration would be lower, and therefore the pull from deeper layers is even somewhat higher for tacrolimus.
In plasma we found that lipoproteins contributed strongly to the overall binding of pimecrolimus and tacrolimus, in line with published data for tacrolimus (Nagase et al., 1994). With ultracentrifugation, an unbound fraction for both drugs of 20 to 30% in human plasma was determined (Piekoszewski et al., 1993; unpublished Novartis internal data). However, although most of the total plasma protein can be separated by ultracentrifugation, lipoproteins can, depending on their characteristic density, either float or sediment very slowly or not at all (Olson, 1998), preventing a complete separation by ultracentrifugation. Therefore, plasma protein binding results based on ultracentrifugation are misleading for highly lipoprotein-bound drugs, because they will underestimate the extent of protein binding. For tacrolimus a 20-fold lower unbound plasma fraction of 1.2% was measured by use of ultrafiltration and equilibrium dialysis (Nagase et al., 1994; Zahir et al., 2001). This value is closer to but approximately 3 times lower than the 3.7% determined in the present study by equilibrium gel filtration. These studies using ultrafiltration and equilibrium dialysis may underestimate, e.g., the concentration of tacrolimus in plasma water due to wall binding or binding to the membrane used for separation as suggested previously (Venkataramanan et al., 1995). With equilibrium gel filtration we experienced significant drug adsorption to the column. However, the actual free concentration could be determined accurately, owing to use of radiolabeled drugs, large volumes available for measurement, and lack of pipetting steps. Gel filtration separates molecules according to size by providing different bed volumes for differently sized molecules. When the system is equilibrated with the drug, the larger and faster moving proteins are always exposed to the same free drug concentration leading to equilibration (Hummel et al., 1962; Berger et al., 2003). Here, in addition to a protein-separating gel filtration column, we used desalting columns, which only separate small molecules (≤1 kDa) from large molecules (≥5 kDa, group separation). This allows short running times and accurate determination of total binding in a complex mixture of proteins, such as plasma. The method is very powerful for determining differences in binding of highly bound drugs, because the amount bound at a defined free concentration can be compared (Fig. 3A), rather than very small free concentrations as in the case of standard techniques. A possible source of error is the separation of low molecular weight plasma components such as fatty acids and bilirubin, which can influence drug binding. On the other hand, major sources of bias such as wall or membrane binding of lipophilic drugs have little impact.
The unbound fraction of pimecrolimus in human plasma was determined here to be 0.4%, i.e., approximately 9-fold lower than that for tacrolimus. Reported clinical exposure levels for both drugs are based on blood concentrations (Draelos et al., 2005). Exposure to free drug can be derived using blood distribution data. In the relevant concentration range, the fraction in plasma is 12% for pimecrolimus versus 2 to 5% for tacrolimus (Nagase et al., 1994; Zollinger et al., 2006). This difference partly compensates for the observed difference in protein binding; therefore, at similar total blood concentrations the free concentration would be 2- to 4-fold lower for pimecrolimus.
For pimecrolimus, binding to plasma lipoproteins was higher compared with binding to albumin and α1-acid glycoprotein, the two main drug-binding plasma proteins (Table 3). Binding of both drugs was highest to HDL, consistent with reported data for tacrolimus (Nagase et al., 1994; Zahir et al., 2001). α1-Acid glycoprotein contributed substantially to the binding of tacrolimus in plasma (Table 3), which agrees with the finding that the unbound fraction of tacrolimus correlates with α1-acid glycoprotein as well as HDL cholesterol levels (Zahir et al., 2004). The binding of pimecrolimus to the different lipoproteins was higher compared with that for tacrolimus, in line with the higher overall plasma protein binding of pimecrolimus and its higher lipophilicity (Fig. 4). To understand the possible impact of this high binding to lipoproteins on the disposition of these drugs, the function and disposition of lipoproteins need to be considered. Plasma lipoproteins mediate lipid transport between tissues; e.g., HDL transports cholesterol to the liver. The liver is the central organ in lipoprotein metabolism and a major fraction of lipoproteins is eventually taken up by the liver via receptor-mediated processes (Ginsberg, 1998; Olson, 1998). Pimecrolimus is cleared mainly by hepatic oxidative metabolism, followed by biliary excretion of metabolites (Zollinger et al., 2006) and tacrolimus is also cleared mainly by liver metabolism (Venkataramanan et al., 1995). Although lipoprotein binding limits the free drug concentration and with it its liver and overall organ uptake (e.g., by passive diffusion), high lipoprotein binding may, on the other hand, enhance drug uptake via lipoprotein-coupled transport into the liver. To clarify how strongly lipoprotein-mediated uptake contributes to the overall liver uptake of pimecrolimus and tacrolimus would need further investigation. However, the effect might be more pronounced for the higher lipoprotein-bound pimecrolimus, potentially contributing to a higher systemic clearance. A direct comparison of the systemic blood clearance of the two drugs is not possible, because for pimecrolimus no intravenous pharmacokinetic study was performed. The blood clearance of tacrolimus is 4 to 6 liters/h (Venkataramanan et al., 1995), which is probably lower than the blood clearance of pimecrolimus (CL/f of 72 liters/h) (Zollinger et al., 2006).
In conclusion, the current study highlights the importance of binding interactions and the interplay between specific high-affinity and unspecific high-capacity binding of topically applied drugs for controlling drug exposure at the target site and in the systemic circulation. The in vitro data presented herein suggest that higher unspecific binding to skin proteins is responsible for the lower skin permeation and the lower systemic exposure upon topical dosing of pimecrolimus compared with tacrolimus. In addition, the side-by-side comparison of plasma protein binding of the two drugs suggests that the difference in exposure to unbound drug is even more pronounced.
Acknowledgments
We thank Ernestine Dobrowolski and Roland Reuschel for excellent technical assistance and Leo Kawai and Felix Waldmeier for valuable scientific advice and comments on the manuscript.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.108.021915.
-
ABBREVIATIONS: AGP, α1-acid glycoprotein; FKBP, FK506-binding proteins; fu, unbound fraction; HDL, high density lipoprotein; LDL, low density lipoprotein; HSA, human serum albumin; LSC, liquid scintillation counting; VLDL, very low density lipoprotein.
- Received April 17, 2008.
- Accepted June 3, 2008.
- The American Society for Pharmacology and Experimental Therapeutics
References
DMD articles become freely available 12 months after publication, and remain freely available for 5 years.Non-open access articles that fall outside this five year window are available only to institutional subscribers and current ASPET members, or through the article purchase feature at the bottom of the page.
|